CN115422918A - Narrative capability evaluation method and device for narrative object - Google Patents
Narrative capability evaluation method and device for narrative object Download PDFInfo
- Publication number
- CN115422918A CN115422918A CN202210970190.4A CN202210970190A CN115422918A CN 115422918 A CN115422918 A CN 115422918A CN 202210970190 A CN202210970190 A CN 202210970190A CN 115422918 A CN115422918 A CN 115422918A
- Authority
- CN
- China
- Prior art keywords
- narrative
- score
- target media
- preset
- media work
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
- G06F40/216—Parsing using statistical methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Probability & Statistics with Applications (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the invention provides a narrative capability evaluation method and a narrative capability evaluation device of a narrative object, wherein the method comprises the following steps: obtaining a plurality of target media works of a narrative object and a first score of each target media work on a media presentation platform; determining a second score of each target media work in each preset narrative dimension related to the subject matter of the target media work based on the score model corresponding to each preset narrative dimension; mapping the first score of each target media work to a normalized score interval corresponding to a media display platform to obtain a third score of the target media work; determining a fourth score of each target media work according to the second score of each target media work in each preset narrative dimension and the third score of the target media work; determining a narrative ability score for the narrative object on each material based on the fourth score for each target media work for that material. By adopting the technical scheme, the accuracy of the narrative capability evaluation of the narrative object can be improved.
Description
Technical Field
The invention relates to the technical field of big data analysis, in particular to a narrative capability evaluation method and device for narrative objects.
Background
The author of a media work is a narrative object of the media work, the narrative capabilities of which have a decisive effect on the quality of the media work. The assessment of the narrative capabilities of narrative objects plays an important role in the early and late assessments of media work items.
Currently, the evaluation of narrative abilities of narrative objects is primarily based on the quantity and quality of media works that the narrative objects have authored. But the number and quality of media pieces that a narrative object has created presents a sample tilt problem. This would make the evaluation of the narrative capabilities of the narrative objects less accurate.
Disclosure of Invention
The embodiment of the invention aims to provide a narrative ability evaluation method and device for a narrative object so as to improve the accuracy of the narrative ability evaluation of the narrative object. The specific technical scheme is as follows:
in a first aspect of the embodiments of the present invention there is provided first a method of narrative capability assessment of a narrative object, comprising:
obtaining a plurality of target media pieces of a narrative object and a first score of each target media piece on a media presentation platform;
determining a second score of each target media work in each preset narrative dimension related to the subject matter of the target media work based on the score model corresponding to each preset narrative dimension;
mapping the first score of each target media work to a normalized score interval corresponding to the media display platform to obtain a third score of the target media work;
determining a fourth score of each target media work according to the second score of each target media work in each preset narrative dimension and the third score of the target media work;
determining a narrative ability score for said narrative object on each subject based on a fourth score for each target media work for that subject.
In one possible implementation, the method further includes:
acquiring subject associated information of each target media work;
and determining the subject of each target media work according to the subject correlation information of each target media work.
In one possible implementation, the story associated information includes user portrait data and narrative maneuver information;
the step of determining the subject to which the target media work belongs according to the subject associated information of each target media work comprises the following steps:
converting user portrait data and narrative technique information of each target media work into a feature vector of the target media work;
and determining a vector class to which the characteristic vector of each target media work belongs by adopting a preset clustering algorithm, wherein the subject corresponding to the vector class is the subject to which the target media work belongs.
In one possible implementation manner, the theme related information comprises a work label and scenario brief introduction information;
the step of determining the subject matter to which the target media works belong according to the subject matter associated information of each target media work comprises the following steps:
extracting keywords of each subject from the product label and the plot brief introduction information of each target media product;
determining the score and the value of a keyword corresponding to each target media work in each subject as the subject score of the subject corresponding to the target media work;
and determining the subject associated with the highest subject score corresponding to each target media work as the subject to which the target media work belongs.
In one possible implementation manner, the step of determining a second score of each target media work in each preset narrative dimension associated with the subject matter of the target media work based on the scoring model corresponding to each preset narrative dimension includes:
extracting a scoring parameter of each preset narrative dimension associated with the subject matter of each target media work from scoring associated information associated with each target media work;
and inputting the scoring parameters of each target media work in each preset narrative dimension into a scoring model corresponding to the preset narrative dimension to obtain a second score of the target media work in the preset narrative dimension.
In one possible implementation manner, the scoring related information includes image materials and text materials;
the scoring model of each preset narrative dimension related to the image material is a convolutional neural network model;
the scoring model of each preset narrative dimension associated with the text data is a NLP (neural-Linear Programming, natural language processing) model.
In one possible implementation, the method further includes: aiming at each preset narrative dimension, training to obtain a scoring model of the preset narrative dimension by adopting the following steps:
acquiring training data, wherein the training data is grade associated information of the marking grade with the preset narrative dimension;
extracting sample scoring parameters for the preset narrative dimension from the training data;
inputting the sample scoring parameters of the preset narrative dimension into a scoring model corresponding to the preset narrative dimension to obtain a prediction score of the preset narrative dimension;
determining the model loss of the preset narrative dimension according to the prediction score and the marking score of the preset narrative dimension;
if the score model corresponding to the preset narrative dimension is determined to be converged according to the model loss of the preset narrative dimension, finishing the score model training corresponding to the preset narrative dimension;
if the score model corresponding to the preset narrative dimension is determined not to be converged according to the model loss of the preset narrative dimension, adjusting parameters of the score model corresponding to the preset narrative dimension, and re-executing the step of inputting the sample score parameters of the preset narrative dimension into the score model corresponding to the preset narrative dimension to obtain a prediction score of the preset narrative dimension.
In one possible implementation, the step of determining a fourth score for each target media work based on the second score for the target media work in each preset narrative dimension and the third score for the target media work comprises:
according to the weight coefficient of each preset narrative dimension and the weight coefficient of the media display platform, carrying out weighting processing on a second score of each target media work in each preset narrative dimension and a third score of the target media work to obtain a fourth score of the target media work; or
Selecting the largest score from the second score of the target media composition in each preset narrative dimension and the third score of the target media composition as the fourth score of the target media composition.
In one possible implementation, the determining a narrative ability score for the narrative object on each material according to a fourth score for each target media work for that material comprises:
for each subject, carrying out average processing on the fourth score of each target media work of the subject to obtain narrative ability scores of the narrative objects on the subject; or
For each topic, selecting the largest fourth score from the fourth scores of each target media work of the topic as the narrative ability score of the narrative object on the topic.
In one possible implementation, the method further includes:
determining a target subject matter of a media work to be created;
obtaining narrative ability scores for a plurality of candidate narrative objects on the target material;
and determining the narrative object of the media work to be authored from the candidate narrative objects with the narrative capability scores larger than the preset score threshold value.
In a second aspect of an embodiment of the present invention, there is also provided an apparatus for narrative capability assessment of narrative objects, comprising:
the system comprises a first acquisition module, a second acquisition module and a third acquisition module, wherein the first acquisition module is used for acquiring a plurality of target media works of a narrative object and a first score of each target media work on a media display platform;
the first determination module is used for determining a second score of each target media work in each preset narrative dimension related to the subject matter of the target media work based on the score model corresponding to each preset narrative dimension;
the mapping module is used for mapping the first score of each target media work to the normalized score interval corresponding to the media display platform to obtain a third score of the target media work;
a second determining module, configured to determine a fourth score of each target media work according to a second score of each target media work in each preset narrative dimension and a third score of the target media work;
and a third determining module for determining a narrative ability score of the narrative object on each subject according to a fourth score of each target media work of the subject.
In one possible implementation, the apparatus further includes:
the second acquisition module is used for acquiring the subject associated information of each target media work;
and the fourth determining module is used for determining the subject to which the target media work belongs according to the subject related information of each target media work.
In one possible implementation, the story associated information includes user portrait data and narrative maneuver information; the fourth determining module is specifically configured to:
converting user portrait data and narrative manipulation information of each target media work into a feature vector of the target media work;
and determining a vector class to which the characteristic vector of each target media work belongs by adopting a preset clustering algorithm, wherein the subject corresponding to the vector class is the subject to which the target media work belongs.
In one possible implementation manner, the subject related information includes a work label and scenario brief introduction information; the fourth determining module is specifically configured to:
extracting keywords of each subject from the product label and the plot brief introduction information of each target media product;
determining the score and the value of a keyword corresponding to each target media work in each subject as the subject score of the subject corresponding to the target media work;
and determining the subject matter associated with the highest subject matter score corresponding to each target media work as the subject matter to which the target media work belongs.
In a possible implementation manner, the first determining module is specifically configured to:
extracting a scoring parameter of each preset narrative dimension associated with the subject matter of each target media work from scoring associated information associated with each target media work;
and inputting the scoring parameters of each target media work in each preset narrative dimension into a scoring model corresponding to the preset narrative dimension to obtain a second score of the target media work in the preset narrative dimension.
In one possible implementation manner, the scoring related information includes image materials and text materials; the scoring model of each preset narrative dimension related to the image material is a convolutional neural network model; and the scoring model of each preset narrative dimension related to the text data is an NLP model.
In one possible implementation, the apparatus further includes: the training module is used for training a scoring model of each preset narrative dimension, and comprises the following steps:
the third acquisition submodule is used for acquiring training data, wherein the training data is grade associated information of the marking grade with the preset narrative dimension;
the extraction submodule is used for extracting a sample scoring parameter of the preset narrative dimension from the training data;
the input submodule is used for inputting the sample scoring parameters of the preset narrative dimension into a scoring model corresponding to the preset narrative dimension to obtain the prediction score of the preset narrative dimension;
the training submodule is used for determining the model loss of the preset narrative dimension according to the prediction score and the marking score of the preset narrative dimension; if the score model corresponding to the preset narrative dimension is determined to be convergent according to the model loss of the preset narrative dimension, finishing the score model training corresponding to the preset narrative dimension; if the scoring model corresponding to the preset narrative dimension is determined not to be converged according to the model loss of the preset narrative dimension, adjusting parameters of the scoring model corresponding to the preset narrative dimension, and re-executing the step of inputting the sample scoring parameters of the preset narrative dimension into the scoring model corresponding to the preset narrative dimension to obtain the predicted score of the preset narrative dimension.
In a possible implementation manner, the second determining module is specifically configured to:
according to the weight coefficient of each preset narrative dimension and the weight coefficient of the media display platform, carrying out weighting processing on a second score of each target media work in each preset narrative dimension and a third score of the target media work to obtain a fourth score of the target media work; or
Selecting the largest score from the second score of the target media composition in each preset narrative dimension and the third score of the target media composition as the fourth score of the target media composition.
In a possible implementation manner, the third determining module is specifically configured to:
for each subject, carrying out average processing on the fourth score of each target media work of the subject to obtain narrative ability scores of the narrative objects on the subject; or
For each topic, selecting the largest fourth score from the fourth scores of each target media work of the topic as the narrative ability score of the narrative object on the topic.
In one possible implementation, the apparatus further includes:
the fifth determining module is used for determining a target subject matter of the media work to be created;
a fourth obtaining module for obtaining narrative ability scores of a plurality of candidate narrative objects on the target subject matter;
and the sixth determining module is used for determining the narrative object of the media work to be authored from the candidate narrative objects with the narrative capability scores larger than the preset score threshold value.
In a third aspect of the present invention, there is also provided an electronic device, including: the system comprises a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory complete mutual communication through the communication bus; a memory for storing a computer program; a processor for implementing the method of assessing the narrative abilities of a narrative object of the first aspect when executing a program stored in the memory.
In a fourth aspect of embodiments of the present invention there is also provided a computer readable storage medium having stored therein a computer program which, when executed by a processor, implements the method of narrative capability assessment of narrative objects of the first aspect.
In the technical solution provided in the embodiment of the present invention, on one hand, a normalized score interval corresponding to the media display platform, that is, scores of media works of various topics displayed on the media display platform are normalized to the same interval, and a first score of a target media work on the media display platform is mapped to the normalized score interval corresponding to the media display platform to obtain a third score of the target media work, so that a sample tilt problem caused by an influence of a user group of the media work display platform is reduced, that is, the score reasonableness of the media work display platform is improved; and evaluating the narrative ability of the narrative object by using the obtained more reasonable third score, thereby improving the accuracy of the narrative ability evaluation of the narrative object.
On the other hand, the narrative ability of the narrative object is subdivided into narrative abilities on various subjects, namely, the narrative abilities of the narrative objects on various subjects are evaluated by combining the second scores and the third scores of each preset narrative dimension related to the subjects of a plurality of target media works, so that the sample tilt problem caused by the quantity and quality differences of the media works which are created by the narrative objects on different subjects is reduced, and the accuracy of the narrative ability evaluation of the narrative objects is further improved.
Of course, not all of the advantages described above need to be achieved at the same time in the practice of any one product or method of the invention.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below.
FIG. 1 is a first flowchart of a method for evaluating narrative abilities of narrative objects according to one embodiment of the present invention.
FIG. 2 is a second flowchart of a method for evaluating narrative abilities of narrative objects according to an embodiment of the present invention.
FIG. 3 is a third flowchart illustrating a method for evaluating narrative abilities of narrative objects according to an embodiment of the present invention.
FIG. 4 is a fourth flowchart illustrating a method for evaluating narrative abilities of narrative objects according to an embodiment of the present invention.
FIG. 5 is a fifth flowchart illustrating a method for evaluating narrative abilities of narrative objects according to an embodiment of the present invention.
Fig. 6 is a flowchart illustrating a method for training a score model with preset narrative dimensions according to an embodiment of the present invention.
FIG. 7 is a sixth flowchart illustrating a method for evaluating narrative abilities of narrative objects according to an embodiment of the present invention.
FIG. 8 is a seventh flowchart illustrating a method for evaluating narrative abilities of narrative objects according to one embodiment of the present invention.
FIG. 9 is a flow chart illustrating a method for determining narrative objects according to one embodiment of the present invention.
FIG. 10 is a schematic structural diagram of an apparatus for evaluating narrative abilities of narrative objects according to an embodiment of the present invention.
Fig. 11 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be described below with reference to the drawings in the embodiments of the present invention.
The media works are in a medium form facing to the public, a story is told through a certain narrative method and is watched by a user, and the psychological needs of a target user are met, so that the virtuous circle of consumption supply and demand is achieved. Media works include, but are not limited to, voice works, image works, written works, multimedia works (e.g., movie and television works), and the like. The voice works can include, but are not limited to, audio novels, radio broadcasting works and the like, the image works can include, but are not limited to, comics and other media works formed by one or more images and the like, the literal works can include, but are not limited to, network novels, network texts and the like, and the multimedia works can include, but are not limited to, television shows, movies, art programs and the like.
The narrative object is the author of the media work. The creator may be an independent individual, such as a director and a theatrical production of a multimedia work. The author may also be an authoring team, and the embodiment of the present invention is not limited in this respect.
The media display platform is a platform for displaying media works and can be video playing software, novel reading software and the like. For example, a user watches a movie on a certain video playing software, the user can score the movie on the video playing software, and the video playing software can obtain the score of the movie on the video playing software based on the scores of all the users watching the movie on the movie.
Currently, the evaluation of narrative abilities of narrative objects is primarily based on the quantity and quality of media works that the narrative object has created. However, the number and quality of media works that a narrative object has created presents a sample tilt problem that makes the assessment of the narrative capabilities of the narrative object less accurate. The sample tilt problem is mainly caused by the following two reasons:
1. media works of different themes are affected by a user group of the media work display platform, so that the problem of sample inclination exists in the quantity and quality of media works created by the acquired narrative objects. For example: on a certain media work display platform, young users generally have higher movie scores for comedy topics and generally have lower movie scores for historical topics. This results in a virtually high media work score for comedy material created by narrative objects and a virtually low media work score for historical material created by narrative objects on the same media work presentation platform, making the assessment of narrative abilities of narrative objects less accurate.
2. Sample tilt problems caused by the difference in the quantity and quality of media pieces on different subjects for narrative objects. For example: narrative objects take more media pieces of comedy material and have a higher score, while narrative objects take less media pieces of suspense material and have a lower score. However, the evaluation of narrative abilities of narrative objects at present is based on all media works which have been created by the narrative objects, and the subject types of the media works which have been created by the narrative objects are not subdivided, so that only the overall evaluation of the narrative objects on all subjects can be obtained, but the evaluation of the narrative abilities of the narrative objects on different subjects is not clear, and the evaluation of the narrative abilities of the narrative objects is not accurate enough.
In order to improve the accuracy of narrative capability assessment of narrative objects, the embodiment of the invention provides a narrative capability assessment method of narrative objects, as shown in figure 1, the method comprises the following steps:
step S11, a plurality of target media works of the narrative object and a first score of each target media work on the media display platform are obtained.
And S12, determining a second score of each target media work in each preset narrative dimension related to the subject matter of the target media work based on the score model corresponding to each preset narrative dimension.
And S13, mapping the first score of each target media work to a normalized score interval corresponding to the media display platform to obtain a third score of the target media work.
Step S14, determining a fourth score of each target media work according to the second score of each target media work in each preset narrative dimension and the third score of the target media work;
step S15, according to the fourth score of each target media work of each subject, determining narrative ability score of narrative object on the subject.
Therefore, by adopting the technical scheme provided by the embodiment of the invention, on one hand, a normalized scoring interval corresponding to the media display platform is obtained, namely, the scores of the media works of various themes displayed on the media display platform are normalized to the same interval, the first score of the target media work on the media display platform is mapped to the normalized scoring interval corresponding to the media display platform, so that the third score of the target media work is obtained, the problem of sample inclination caused by the influence of a user group of the media work display platform is reduced, and the scoring rationality of the media work display platform is improved; and evaluating the narrative ability of the narrative object by using the obtained more reasonable third score, thereby improving the accuracy of the narrative ability evaluation of the narrative object.
On the other hand, the narrative ability of the narrative object is subdivided into the narrative ability on each subject matter, namely the narrative ability of the narrative object on each subject matter is evaluated by combining the second score and the third score of each preset narrative dimension associated with the subject matter of a plurality of target media works, so that the sample inclination problem caused by the quantity and quality difference of the media works which are already authored by the narrative object on different subject matters is reduced, and the accuracy of the narrative ability evaluation of the narrative object is further improved.
For convenience of description and understanding, the following description is made by taking an electronic device as an execution subject and does not have a limiting effect.
In step S11 above, the target media piece is a media piece that has been authored by a narrative object. Upon evaluating a narrative capability for a narrative object, an electronic device obtains a plurality of target media pieces of the narrative object and obtains a first score for each target media piece on a media presentation platform. If one target media work is displayed on a plurality of media display platforms, the electronic equipment acquires a first score of the target media work for each media display platform, and acquires a plurality of first scores.
In the step S12, the electronic device presets one or more narrative dimensions, i.e. preset narrative dimensions. The preset narrative dimensions can be set according to the main selling points of the subject matter to which the media work belongs.
For example, a female subject material may be associated with two preset narrative dimensions, that is, a set discussion preference of a person and a good chemical reaction of a person interaction, based on the fact that the female subject material attracts a user group mainly including a female with an emotional line of the interaction between the persons as a narrative main line and further, the set discussion preference of the person and the good chemical reaction of the person interaction are mainly sold. Of course, other preset narrative dimensions may be associated with the female subject matter.
For another example, the suspense-type material may be associated with two preset narrative dimensions of a case and a manufacturing suspense, based on the suspense-type material attracting a user group mainly of young men and further mainly selling the tension stimulus and the manufacturing suspense of the story. Of course, other pre-set narrative dimensions may be associated with the suspense material.
The preset narrative dimensions may also include a uniform track dimension and an action dimension, among others.
The primary selling point, i.e., a specific narrative technique, may also be referred to as a default narrative dimension. The higher the degree of completion (i.e., the second score) of the preset narrative dimension of the subject matter association, the more satisfying the psychological needs of the user of the media work and the more enjoyable the user of the media work.
A material may be associated with a plurality of predefined narrative dimensions, each of which corresponds to a scoring model. The input of the scoring model may be a scoring parameter obtained from image material, text material and voice material of the media work, and the output is a score (i.e., a second score) of the media work in a corresponding narrative dimension, and the scoring parameter will be described in detail later.
After obtaining target media pieces that have been authored by a narrative object, the electronic device may determine the story to which each target media piece belongs, each story being associated with one or more preset narrative dimensions. And for each target media work, the electronic equipment determines a second score of each preset narrative dimension related to the subject matter of the target media work based on the score model corresponding to each preset narrative dimension.
In the embodiment of the invention, the subject to which each target media work belongs can be pre-calibrated, and the target media works can be classified in a designated manner to determine the subject to which the target media works belong. Hereinafter, two embodiments for determining the subject to which the target media piece belongs will be given in detail, and will not be described herein again.
In the step S13, a normalized scoring interval of the media display platform is preset in the electronic device, and the first score of each topic displayed on the media display platform can be mapped to the normalized scoring interval. After the first scores of the media works authored by the narrative objects on the media display platform are obtained, the electronic equipment maps the first scores of the target media works to the normalization score intervals corresponding to the media display platform aiming at each target media work, and then the third scores of the target media works are obtained.
On the same media display platform, due to the fact that the user groups have different degrees of love on the media works of different themes, scores of the media works of different themes are distributed in different scoring intervals. In the embodiment of the invention, the electronic equipment normalizes the scoring interval of each subject on the media display platform to an interval, namely the normalized scoring interval corresponding to the media display platform, and maps the first score of the target media work on the media display platform to the normalized scoring interval corresponding to the media display platform, so that the scoring moisture of the media display platform on the preference deviation of the subject to the target media work can be removed, and the real score of the target media work on the media display platform can be objectively reflected.
In the embodiment of the invention, the mode of normalizing the scoring interval of each subject to the specified normalized scoring interval can be set according to actual requirements. For example, score normalization may be implemented by adding or subtracting scores, or by percentage.
For example, on the same media display platform, the user's preference degree for the media works of comedy subjects is greater than that for the media works of suspense subjects, and the preference degree for the media works of comedy subjects is greater than that of the media works of history subjects. For example, scores of media works of comedy themes are concentrated in a score interval of 7-9 points, scores of media works of suspense themes are concentrated in a score interval of 5-7 points, scores of media works of historical themes are concentrated in a score interval of 3-5 points, and the electronic equipment can normalize the score interval of each theme on the media display platform to a score interval of 6-8 points in an adding and subtracting score mode. That is, the score of the media works of comedy themes is subtracted by 1, the score of the media works of suspense themes is added by 1, and the score of the media works of historical themes is added by 3. At this time, the subject of the target media work is comedy subject, the first score on the media display platform is 8 points, the first score is mapped to the normalized score interval corresponding to the media display platform, and a third score of the target media work is obtained, wherein the third score is 7 points (the 8 points minus 1 point).
In the embodiment of the present invention, step S12 may be executed first, and then step S13 may be executed; step S13 may be executed first, and then step S12 may be executed; step S12 and step S13 may be performed simultaneously. The execution sequence of step S12 and step S13 is not particularly limited in the embodiment of the present invention.
After the second score of each target media work in each preset narrative dimension and the third score of the target media work are determined, the step S14 is executed, that is, for each target media work, the fourth score of the target media work is determined according to the second score of the target media work in each preset narrative dimension and the third score of the target media work.
A plurality of target media pieces may be included under each subject, and after the fourth score of each target media piece under each subject is determined, the step S15 is executed, that is, the narrative ability score of the narrative object on different subjects is determined according to the fourth score of each target media piece under each subject.
Based on the embodiment shown in FIG. 1, the embodiment of the present invention further provides a narrative ability evaluation method of narrative objects, as shown in FIG. 2, the method may include the following steps S21-S27. Wherein step S21 is the same as step S11 described above, and steps S24 to S27 are the same as steps S12 to S15 described above.
And step S22, obtaining the subject related information of each target media work.
Wherein, the subject related information is the information needed for determining the subject of the media work. For example, the material-related information may include user portrait data and narrative technique information, as well as a work tag and scenario profile information.
The user portrait data can be obtained by analyzing the historical viewing records of the user, and the user portrait data can include the age distribution interval, the viewing duration, the occupation, the geographic position and the like of the user. The narrative technique information is a key element constituting the media work, such as the preset narrative dimension. Different media works have different narrative technique information.
And when the subject of the target media work is determined, the electronic equipment acquires the subject related information of the target media work.
And step S23, determining the subject of each target media work according to the subject related information of each target media work.
Since different media presentation platforms have different classifications for the subject matter of the target media work, the narrative ability of the narrative object on each subject matter cannot be accurately evaluated according to the classifications of the subject matter of the target media work by the different media presentation platforms. In the embodiment of the invention, the subject to which the target media work belongs is determined in a unified manner, namely according to the subject associated information of the target media work, so that the problem that different media display platforms have different classifications for the subject to which the target media work belongs can be effectively avoided, and the accuracy of evaluating the narrative ability of the narrative object on each subject is further improved.
In one embodiment of the present invention, the story associated information may include user portrait data and narrative maneuver information. Based on this, the embodiment of the present invention further provides a narrative capability assessment method for narrative objects, as shown in fig. 3, the method may include the following steps: steps S31-S38, wherein steps S31-S32 are the same as steps S21-S22 described above, and steps S35-S38 are the same as steps S24-S27 described above. Steps S33-S34 are one possible implementation of step S23.
Step S33, converting the user portrait data and narrative technique information of each target media work into the feature vector of the target media work.
And step S34, determining a vector class to which the feature vector of each target media work belongs by adopting a preset clustering algorithm, wherein the subject corresponding to the vector class is the subject to which the target media work belongs.
In the technical scheme provided by the embodiment of the invention, the subject to which each target media work belongs is determined by adopting a clustering algorithm on the basis of user portrait data from a user and narrative technique information from the media works. Therefore, the way of determining the subject establishes a link between the narrative skill of the media work and the psychological needs of the user, which can more accurately determine the subject to which the target media work belongs, thereby improving the accuracy of evaluating the narrative ability of the narrative object.
In step S33, the electronic device converts the user portrait data and narrative technique information of each target media work into a feature vector of the target media work.
In the embodiment of the invention, the electronic equipment can preset a conversion rule, and according to the conversion rule, the electronic equipment converts the user portrait data and narrative manipulation information of each target media work into the feature vector of the target media work. The conversion rule can be set according to actual requirements.
In one example, the transformation rule may be: extracting data of preset dimensions from user portrait data and narrative manipulation information, converting the extracted data of each preset dimension into characteristic values according to the corresponding relation between the pre-stored data of each preset dimension and the characteristic values, and forming characteristic vectors by the characteristic values.
For example, data 1 in a predetermined dimension 1 and data 2 in a predetermined dimension 2 are extracted from user portrait data and narrative technique information. The electronic device stores a feature value 1 corresponding to data 1 of a preset dimension 1 in advance, and a feature value 2 corresponding to data 2 of a preset dimension 2 in advance, so that the electronic device can determine that the feature vector is { feature value 1, feature value 2}.
In another example, the transformation rule may be: inputting user image data into a corresponding neural network to obtain a first characteristic sequence, inputting narrative manipulation information into another corresponding neural network to obtain a second characteristic sequence, and forming a characteristic vector by the first characteristic sequence and the second characteristic sequence.
In the step S34, the preset clustering algorithm may be a K-Means clustering algorithm (K-Means clustering algorithm), a Mean-Shift algorithm (Mean Shift algorithm), or the like, which is not specifically limited in the embodiment of the present invention.
The vector class is obtained by clustering a plurality of sample media works by adopting a preset clustering algorithm. For example, the electronic device obtains a plurality of training samples, wherein the training samples are subject associated information of sample media works; respectively converting a plurality of training samples into feature vectors, adopting a preset clustering algorithm, selecting a preset number of feature vectors as anchor points, and clustering the feature vectors of the plurality of training samples to obtain a preset number of classes, wherein each class can be understood as a set of training samples of which the feature vectors are similar to user portrait data and narrative technique information of the anchor points; secondly, taking vectors with the distance between the vectors and the anchor point being larger than the preset distance in one class as new anchor points, and clustering the feature vectors of the training samples again to obtain a plurality of classes; and finally forming a plurality of stable classes as vector classes respectively. One vector class can be called as a basic disc, and one basic disc corresponds to one subject, namely, narrative skill information of each basic disc is different. In the embodiment of the invention, the finally obtained vector class can be regarded as a media work psychological demand model which establishes a link between narrative skills of media works and psychological demands of users.
Under the condition of determining a plurality of vector classes, aiming at each target media work, the electronic equipment adopts a preset clustering algorithm to perform clustering processing on the target media work to obtain a vector class to which the feature vector of the target media work belongs, and according to the vector class to which each target media work belongs, the subject to which the target media work belongs can be determined, namely the subject corresponding to the vector class can be used as the subject to which the target media work belongs.
In one embodiment of the invention, the story associated information includes a work label and scenario profile information. Based on this, the embodiment of the present invention further provides a narrative capability assessment method for narrative objects, as shown in fig. 4, the method may include the following steps: steps S41 to S49, wherein steps S41 to S42 are the same as steps S21 to S22 described above, and steps S46 to S49 are the same as steps S24 to S27 described above. Steps S43-S45 are one possible implementation of step S23.
Step S43, extracting keywords of each subject from the product label and the plot brief introduction information of each target media product.
Step S44, aiming at each target media work, determining the score and the value of the keyword corresponding to the target media work in each subject as the subject score of the subject corresponding to the target media work.
Step S45, aiming at each target media work, determining the subject matter associated with the highest subject matter score corresponding to the target media work as the subject matter to which the target media work belongs.
In the technical scheme provided by the embodiment of the invention, the subject score of a target media work in each subject is determined to determine the subject to which the target media work belongs, namely, a cross numerical system is used for scoring the work label of each target media work and the subject key words extracted from the plot brief introduction information, scores are accumulated on each subject, and the subject with the highest score is determined as the subject to which the target media work belongs. Therefore, in the embodiment of the invention, the media works are subjected to cross verification among different subjects, the degree of meeting the psychological needs of users of the works is quantized, and the subjects of the target media works can be more accurately determined by considering the subject scores of the target media works in each subject, so that the accuracy of evaluating the narrative ability of narrative objects is improved.
In step S43, the electronic device may preset the keywords of each subject and the scores of the keywords in different subjects, or may also understand the weights of the keywords in different subjects, the keywords are words closely related to the subjects, the keywords of different subjects may be the same or different, and the scores of the keywords of different subjects may be the same or different. However, the keywords and the scores of the keywords are not completely the same for different subjects.
For example, the keywords of comedy themes include police, fun, love, and the keywords of police gangster themes include police, gangster, love. In comedy themes, the score of the keyword "police" is 50 points, the score of the keyword "fun" is 90 points, the score of the keyword "love" is 40 points, in police themes, the score of the keyword "police" is 95 points, the score of the keyword "gangster" is 90 points, and the score of the keyword "love" is 40 points.
For a target media work, for each subject, extracting keywords of the subject from the work label and the plot brief introduction information of the target media work. The electronic device may preset a plurality of keywords of a theme, and when extracting the keywords of the theme from the product tag and the scenario profile information of the target media work, all the preset keywords of the theme may be extracted, some preset keywords of the theme may be extracted, or none preset keywords of the theme may be extracted.
In step S44, for a target media work, for each subject, the electronic device accumulates the keywords belonging to the subject to obtain a score and a value of the keyword corresponding to the target media work in the subject, where the score and the value are the subject score of the subject corresponding to the target media work.
In step S45, for each target media piece, the electronic device selects the highest subject score from the subject scores corresponding to the target media piece, and takes the subject associated with the highest subject score as the subject to which the target media piece belongs.
For example, keywords "spy", "quest", "inference", "snaking" and "snaking" are extracted from a target media production. In the questionable subject, 9 points are scored for "spy", "exploration" 3 points, 7 points are scored for "reasoning", and 0 point is scored for "touting abdomen"; in the adventure subjects, 0 is scored for "detective", 9 is scored for "adventure", 0 is scored for "reasoning", 0 is scored for "puzzling abdomen"; in comedy themes, "spy" scores 0, "quest" scores 0, "snaking abdomen" scores 8, and "reasoning" is not included. Therefore, the keyword score sum value corresponding to the target media work in the suspense subject matter is 19 points (9 +7+3 points), the keyword score sum value corresponding to the target media work in the adventure subject matter is 9 points, the keyword score sum value corresponding to the target media work in the comedy subject matter is 8 points, the highest keyword score sum value is 19 points, namely the highest subject matter score is 19 points, 19 points correspond to the suspense subject matter, and then the subject matter to which the target media work belongs can be determined to be the suspense subject matter.
Based on the embodiment shown in FIG. 1, the embodiment of the present invention further provides a narrative ability assessment method of narrative objects, as shown in FIG. 5, the method may include the following steps S51-S56. Wherein step S51 is the same as step S11, and steps S54-S56 are the same as steps S13-S15. Steps S52-S53 are one possible implementation of step S12.
Step S52, extracting the score parameter of each preset narrative dimension related to the subject matter of each target media work from the score related information related to each target media work.
And S53, inputting the scoring parameters of each target media work in each preset narrative dimension into a scoring model corresponding to the preset narrative dimension to obtain a second score of the target media work in the preset narrative dimension.
According to the technical scheme provided by the embodiment of the invention, the narrative objects are scored on each preset narrative dimension on different subject matters, and then the narrative capacity requirements of the narrative objects on different subject matters are limited, so that the accuracy of the narrative capacity evaluation of the narrative objects is further improved.
In step S52, each target media work is associated with rating association information, which may include, but is not limited to, image material, text material, voice material, and the like. The image material may be a video segment, a work picture, etc. of the media work, the text material may include a work introduction of the media work, a user's evaluation of the media work, etc., and the voice material may include a voice segment of the media work, etc.
The preset narrative dimensions can be understood as the primary selling points in the material. The scoring parameter of the default narrative dimension may be understood to be a parameter that embodies the primary selling point. For example, a movie of the suspense subject matter takes the tension stimulus of the story and the suspense manufacturing as main selling points, namely two preset narrative dimensions including the tension stimulus of the story and the suspense manufacturing are included, and a parameter capable of showing the tension stimulus degree of the story and a scoring parameter of the preset narrative dimension, such as action fluency, plot compactness and the like, of the tension stimulus and the preset narrative dimension, and a parameter capable of showing the suspense manufacturing and a scoring parameter of the preset narrative dimension, such as plot reversal times and the like, of the suspense manufacturing are manufactured. There may be one or more scoring parameters for a preset narrative dimension.
And aiming at each target media work, the electronic equipment extracts the scoring parameter of each preset narrative dimension associated with the subject matter of the target media work from the scoring associated information of the target media work. At this point, one or more scoring parameters are extracted in each preset narrative dimension for a target media work.
In step S53, a plurality of preset narrative dimensions may be associated with a subject, each corresponding to a scoring model. Multiple subjects may be associated with the same default narrative dimension, and the scoring model associated with the default narrative dimension may be shared by multiple subjects. The scoring model can be determined according to different types of scoring associated information needing to be processed, and the scoring (namely, the completion degree) of the media works in corresponding narrative dimensions can be directionally quantized through the scoring model according to the scoring associated information of the media works. For example, when the scoring correlation information is an image material, the scoring model may be a convolutional neural network model, that is, the scoring model of each preset narrative dimension related to the image material is a convolutional neural network model; when the scoring correlation information is the text data, the scoring model can be an NLP model, that is, the scoring model of each preset narrative dimension related to the text data is the NLP model.
The electronic equipment extracts a scoring parameter of a target media work in each preset narrative dimension associated with the subject matter of the target media work, inputs the scoring parameter of the preset narrative dimension into a scoring model corresponding to the preset narrative dimension aiming at each preset narrative dimension, and outputs a second score of the target media work in the preset narrative dimension after the scoring model processes the input scoring parameter.
The network model may also be other network models, such as a deep neural network model, and may be specifically selected and determined according to actual requirements, which is not specifically limited herein.
Based on the embodiment shown in fig. 5, the embodiment of the invention further provides a method for training a score model of a preset narrative dimension, as shown in fig. 6, for each preset narrative dimension, the electronic device may refer to the process shown in fig. 6 to train and obtain the score model of the preset narrative dimension. The training method of the scoring model may include the following steps.
Step S61, acquiring training data, wherein the training data is grade associated information of a marking grade with a preset narrative dimension;
the training data may include scoring association information for a plurality of media works, and the specific number may be set according to actual requirements. For example, when the scoring model is required to have higher scoring precision, a larger amount of scoring associated information of the media works may be used; when the device performance is poor, a smaller amount of scoring correlation information for the media production may be employed.
Step S62, extracting the sample scoring parameter of the preset narrative dimension from the training data.
And S63, inputting the sample scoring parameters of the preset narrative dimension into a scoring model corresponding to the preset narrative dimension to obtain a prediction score of the preset narrative dimension.
And step S64, determining the model loss of the preset narrative dimension according to the prediction score and the marking score of the preset narrative dimension.
In the embodiment of the present invention, the electronic device may preset a loss threshold. After obtaining the model loss, if the model loss is less than a loss threshold, it may be determined that the scoring model converges, otherwise, it may be determined that the scoring model does not converge.
In the embodiment of the invention, the electronic device can also preset a loss threshold and an iteration threshold. After obtaining the model loss, if the model loss is less than a loss threshold, determining that the scoring model converges; if the training iteration times are larger than the equal iteration threshold, the convergence of the scoring model can be determined; otherwise, it is determined that the scoring model does not converge.
The loss threshold and the iteration threshold can be set according to actual requirements.
Step S65, if the score model corresponding to the preset narrative dimension is determined to be converged according to the model loss of the preset narrative dimension, finishing the training of the score model corresponding to the preset narrative dimension;
and S66, if the scoring model corresponding to the preset narrative dimension is determined not to be converged according to the model loss of the preset narrative dimension, adjusting the parameters of the scoring model corresponding to the preset narrative dimension, and executing the step S63 again.
By adopting the technical scheme provided by the embodiment of the invention, the scoring model with the preset narrative dimensions is trained by utilizing the training data, so that the scoring model can fully learn the change rule of the training data, the trained scoring model can be used for more accurately predicting the second score of the target media work in each preset narrative dimension, and the narrative capability requirements of the narrative object on different subjects are more accurately limited, thereby further improving the accuracy of the narrative capability evaluation of the narrative object.
The description of the above steps S61-S66 is relatively simple, and refer to the related description of fig. 1-5.
Based on the embodiment shown in FIG. 1, the embodiment of the present invention further provides a narrative ability assessment method of narrative objects, as shown in FIG. 7, the method may comprise the following steps S71-S75, wherein the steps S71-S73 are the same as the steps S11-S13, and the step S75 is the same as the step S15. Step S74 is one possible implementation of step S14.
S74, according to the weight coefficient of each preset narrative dimension and the weight coefficient of the media display platform, carrying out weighting processing on the second score of each target media work in each preset narrative dimension and the third score of the target media work to obtain a fourth score of the target media work.
In the embodiment of the invention, the electronic equipment records the weight coefficient of each preset narrative dimension and the weight coefficient of the media display platform in advance. For each target media work, the electronic device may perform weighting processing on the second score of the target media work in each preset narrative dimension and the third score of the target media work by using the weight coefficient of each preset narrative dimension and the weight coefficient of the media presentation platform, so as to obtain a fourth score of the target media work. The weighting coefficient of each preset narrative dimension and the weighting coefficient of the media display platform can be manually distributed or obtained through neural network training, and the embodiment of the invention is not particularly limited to this.
For example, the electronic device has recorded therein in advance: a weighting coefficient 1 of a narrative dimension 1, a weighting coefficient 2 of a narrative dimension 2 and a weighting coefficient 3 of a media display platform are preset. The second score of the target media work a in the preset narrative dimension 1 is score 1, the second score in the preset narrative dimension 2 is score 2, and the third score of the target media work is score 3. Then, the electronic device may determine that the fourth score of the target media work a is: score 1 weight coefficient 1+ score 2 weight coefficient 2+ score 3 weight coefficient 3.
Optionally, the sum of the weighting coefficients of the plurality of preset narrative dimensions and the weighting coefficient of the media presentation platform is 1, which facilitates the calculation of the fourth score of the media work by the electronic device.
In the technical scheme provided by the embodiment of the invention, the fourth score of the target media work is obtained by weighting the second score of the target media work in each preset narrative dimension and the third score of the target media work, so that the influence of the inclination of the sample to the score of the target media work is further removed, the score of the target media work is more objectively fair, and the accuracy of narrative capability evaluation of narrative objects is further improved.
In the embodiment of the present invention, the electronic device may also determine the fourth score of the target media production in other manners. For example, the largest score from the second score of the target media piece in each preset narrative dimension and the third score of the target media piece is selected as the fourth score of the target media piece. For example, the target media work a has a second score of 1 in the preset narrative dimension 1, a second score of 2 in the preset narrative dimension 2, and a third score of 3. Wherein, the score 1> the score 2> the score 3, the electronic device may determine that the score 1 is a fourth score of the target media work a.
Based on the embodiment shown in FIG. 1, the embodiment of the present invention further provides a narrative ability evaluation method of narrative objects, as shown in FIG. 8, the method may include the following steps S81-S85, wherein the steps S81-S84 are the same as the steps S11-S14, and the step S85 is an implementable mode of the step S15.
And S85, averaging the fourth scores of the target media works of the subject to obtain narrative ability scores of narrative objects on the subject.
In the embodiment of the invention, one or more target media works of one subject can be provided. After a fourth score of a target media work which is created on each subject by a narrative object is determined, the fourth score of each target media work of each subject is averaged to obtain a narrative ability score of the narrative object on each subject.
For example, the target media composition of the subject b includes target media compositions 1-3, wherein the fourth score of the target media composition 1 is score 1', the fourth score of the target media composition 2 is score 2', and the fourth score of the target media composition 3 is score 3'. The electronic device may then determine that the narrative capability score of the narrative object on the material b is: (score 1' + score 2' + score 3 ')/3.
According to the technical scheme provided by the embodiment of the invention, the narrative ability scores of narrative objects on all the subjects are obtained by combining the second scores and the third scores of each preset narrative dimension associated with the subjects of a plurality of target media works, so that the sample inclination problem caused by the quantity and quality differences of the media works authored by the narrative objects on different subjects is reduced, and the accuracy of narrative ability evaluation of the narrative objects is further improved.
In the embodiment of the invention, the electronic equipment can also adopt other modes to determine the narrative ability scores of narrative objects on various matters. For example, for each material, the largest fourth score is selected from the fourth scores for each target media work for that material as the narrative ability score for the narrative object on that material.
For example, the target media pieces of the material c include target media pieces 1-3, where the fourth score of the target media piece 1 is score 1, the fourth score of the target media piece 2 is score 2, the fourth score of the target media piece 3 is score 3, and score 1> score 2> score 3, then the electronic device may determine that the narrative ability score of the narrative object on the material c is: and a score of 1.
Based on the embodiments shown in fig. 1 to 8, the embodiment of the present invention further provides a method for determining narrative objects, as shown in fig. 9, the method may include the following steps S91 to S93.
S91, determining a target subject of the media work to be created.
When a media work (i.e. a media work to be created) needs to be created, the electronic device can obtain the subject of the media work to be created, i.e. the target subject. The target subject matter may be input by the user into the electronic device, or may be the target subject matter of the media work to be created determined by the method shown in fig. 2, fig. 3 or fig. 4, which is not limited herein.
S92, obtaining narrative ability scores of a plurality of candidate narrative objects on the target subject matter.
For the determination of narrative ability scores for each candidate narrative object on the target material, reference is made to the associated description in sections 1-8.
And S93, determining the narrative object of the media work to be created from the candidate narrative objects with the narrative capability scores larger than the preset score threshold value.
After obtaining narrative ability scores of a plurality of candidate narrative objects on a target material, the electronic equipment screens the candidate narrative objects with the narrative ability scores larger than a preset score threshold value on the target material from the candidate narrative objects, and determines narrative objects of media works to be created from the screened candidate narrative objects.
For example, the electronic device may output a plurality of screened candidate narrative objects in order of high to low narrative capability scores. In this way, the user may select a narrative object of a media work to be created based on the output candidate narrative objects.
As another example, the electronic device may select the candidate narrative object with the highest narrative capability score as the narrative object for the media work to be created.
In the embodiment of the invention, the electronic equipment can also determine the narrative object of the media work to be created from the screened candidate narrative objects by combining the information such as the production cycle duration of the media work, the number of the media works of the created target subject and the like. For example, the electronic equipment determines the candidate narrative object with the shortest media work production cycle time length from the screened candidate narrative objects as the narrative object of the media work to be created.
In the technical scheme provided by the embodiment of the invention, the narrative object of the media work to be created is determined based on the target subject matter of the media work to be created and by combining the narrative capability score of the candidate narrative object on the target subject matter. Under the condition of improving the accuracy of narrative capability evaluation of narrative objects on target subject matters, suitable narrative objects can be recommended for different types of media works to be created, so that higher-quality media works can be obtained, and the commercial value of the media works to be created is further improved.
Based on the same inventive concept, the narrative ability evaluation method for narrative objects according to the above embodiments of the present invention, accordingly, an embodiment of the present invention provides a narrative ability evaluation device for narrative objects, whose schematic structural diagram is shown in fig. 10, comprising:
a first obtaining module 101 for obtaining a plurality of target media works of a narrative object and a first score of each target media work on a media display platform;
the first determination module 102 is configured to determine, based on a scoring model corresponding to each preset narrative dimension, a second score of each target media work in each preset narrative dimension associated with a subject matter of the target media work;
the mapping module 103 is configured to map the first score of each target media piece to a normalized score interval corresponding to the media display platform to obtain a third score of the target media piece;
a second determining module 104, configured to determine a fourth score of each target media work according to the second score of the target media work in each preset narrative dimension and the third score of the target media work;
a third determination module 105 for determining a narrative ability score for a narrative object on each material based on a fourth score for each target media work for that material.
Optionally, the narrative ability assessment apparatus of said narrative object may further comprise:
the second acquisition module is used for acquiring the subject associated information of each target media work;
and the fourth determining module is used for determining the subject to which the target media work belongs according to the subject related information of each target media work.
Optionally, the story associated information includes user portrait data and narrative maneuver information;
the fourth determining module may be specifically configured to: converting user portrait data and narrative manipulation information of each target media work into a feature vector of the target media work;
and determining a vector class to which the characteristic vector of each target media work belongs by adopting a preset clustering algorithm, wherein the subject corresponding to the vector class is the subject to which the target media work belongs.
Optionally, the subject matter associated information comprises a work label and scenario brief introduction information;
the fourth determining module may be specifically configured to: extracting keywords of each subject from the product label and the plot brief introduction information of each target media product;
determining the score and the value of a keyword corresponding to each target media work in each subject as the subject score of the subject corresponding to the target media work;
and determining the subject associated with the highest subject score corresponding to each target media work as the subject to which the target media work belongs.
Optionally, the first determining module 102 may be specifically configured to:
extracting a scoring parameter of each preset narrative dimension associated with the subject matter of each target media work from scoring associated information associated with each target media work;
and inputting the scoring parameters of each target media work in each preset narrative dimension into a scoring model corresponding to the preset narrative dimension to obtain a second score of the target media work in the preset narrative dimension.
Optionally, the scoring related information includes image material and text material; the scoring model of each preset narrative dimension related to the image material is a convolutional neural network model; the scoring model of each preset narrative dimension associated with the text material is an NLP model.
Optionally, the narrative ability assessment apparatus of said narrative object may further comprise:
the third acquisition module is used for acquiring training data, wherein the training data is grade associated information of the marking grade with the preset narrative dimension;
the extraction module is used for extracting sample scoring parameters of the preset narrative dimension from the training data;
the input module is used for inputting the sample scoring parameters of the preset narrative dimension into a scoring model corresponding to the preset narrative dimension to obtain the prediction score of the preset narrative dimension;
the training module is used for determining the model loss of the preset narrative dimension according to the prediction score and the marking score of the preset narrative dimension; if the score model corresponding to the preset narrative dimension is determined to be convergent according to the model loss of the preset narrative dimension, finishing the score model training corresponding to the preset narrative dimension; if the score model corresponding to the preset narrative dimension is determined not to be converged according to the model loss of the preset narrative dimension, adjusting parameters of the score model corresponding to the preset narrative dimension, and re-executing the step of inputting the sample score parameters of the preset narrative dimension into the score model corresponding to the preset narrative dimension to obtain a prediction score of the preset narrative dimension.
Optionally, the second determining module 104 may specifically be configured to:
according to the weight coefficient of each preset narrative dimension and the weight coefficient of the media display platform, carrying out weighting processing on a second score of each target media work in each preset narrative dimension and a third score of the target media work to obtain a fourth score of the target media work; or
Selecting the largest score from the second score of the target media composition in each preset narrative dimension and the third score of the target media composition as the fourth score of the target media composition.
Optionally, the third determining module 105 may be specifically configured to:
for each subject, carrying out average processing on the fourth score of each target media work of the subject to obtain narrative ability scores of narrative objects on the subject; or
For each topic, selecting the largest fourth score from the fourth scores of each target media work of the topic as the narrative ability score of the narrative object on the topic.
Optionally, the narrative ability assessment apparatus of said narrative object may further comprise:
the fifth determining module is used for determining a target subject of the media work to be created;
a fourth obtaining module for obtaining narrative ability scores of a plurality of candidate narrative objects on the target subject matter;
a sixth determining module for determining a narrative object of a media work to be authored from candidate narrative objects having a narrative ability score greater than a preset score threshold.
An embodiment of the present invention further provides an electronic device, as shown in fig. 11, which includes a processor 111, a communication interface 112, a memory 113, and a communication bus 114, where the processor 111, the communication interface 112, and the memory 113 complete mutual communication through the communication bus 114,
a memory 113 for storing a computer program;
the processor 111 is configured to execute at least the following steps when executing the program stored in the memory 113:
obtaining a plurality of target media pieces of a narrative object and a first score of each target media piece on a media presentation platform;
determining a second score for each target media work in each preset narrative dimension associated with the subject matter of the target media work;
mapping the first score of each target media work to a normalized score interval corresponding to a media display platform to obtain a third score of the target media work;
determining a fourth score of each target media work according to the second score of each target media work in each preset narrative dimension and the third score of the target media work;
determining a narrative ability score for the narrative object on each material based on the fourth score for each target media work for that material.
The communication bus mentioned in the above terminal may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The communication bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is shown, but this does not mean that there is only one bus or one type of bus.
The communication interface is used for communication between the terminal and other equipment.
The Memory may include a Random Access Memory (RAM) or a non-volatile Memory (non-volatile Memory), such as at least one disk Memory. Optionally, the memory may also be at least one memory device located remotely from the processor.
The Processor may be a general-purpose Processor, including a Central Processing Unit (CPU), a Network Processor (NP), and the like; the device can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, or a discrete hardware component.
In yet another embodiment of the present invention, there is further provided a computer readable storage medium having a computer program stored therein, the computer program when executed by a processor implementing the method of assessing the narrative capabilities of any one of the narrative objects of the preceding embodiments.
In a further embodiment provided by the present invention there is also provided a computer program product comprising instructions which, when run on a computer, cause the computer to perform the method of narrative capability assessment of a narrative object of any of the above embodiments.
In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. The procedures or functions according to the embodiments of the invention are all or partially generated when the computer program instructions are loaded and executed on a computer. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire (e.g., coaxial cable, fiber optic, digital Subscriber Line (DSL)) or wirelessly (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that incorporates one or more of the available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid State Disk (SSD)), among others.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising a … …" does not exclude the presence of another identical element in a process, method, article, or apparatus that comprises the element.
All the embodiments in the present specification are described in a related manner, and the same and similar parts among the embodiments may be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the apparatus embodiment, since it is substantially similar to the method embodiment, the description is relatively simple, and reference may be made to the partial description of the method embodiment for relevant points.
The above are only preferred embodiments of the present invention, and are not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention shall fall within the protection scope of the present invention.
Claims (13)
1. A method of assessing the narrative capabilities of a narrative object, comprising:
obtaining a plurality of target media pieces of a narrative object and a first score of each target media piece on a media presentation platform;
determining a second score of each target media work in each preset narrative dimension related to the subject matter of the target media work based on a score model corresponding to each preset narrative dimension;
mapping the first score of each target media work to a normalized score interval corresponding to the media display platform to obtain a third score of the target media work;
determining a fourth score of each target media work according to the second score of each target media work in each preset narrative dimension and the third score of the target media work;
determining a narrative ability score for said narrative object on each subject based on a fourth score for each target media work for that subject.
2. The method of claim 1, further comprising:
acquiring subject associated information of each target media work;
and determining the subject of each target media work according to the subject correlation information of each target media work.
3. The method of claim 2, wherein said story associated information includes user portrait data and narrative maneuver information;
the step of determining the subject of each target media work according to the subject associated information of each target media work comprises the following steps:
converting user portrait data and narrative manipulation information of each target media work into a feature vector of the target media work;
and determining a vector class to which the characteristic vector of each target media work belongs by adopting a preset clustering algorithm, wherein the subject corresponding to the vector class is the subject to which the target media work belongs.
4. The method of claim 2, wherein the story associated information comprises a work label and scenario profile information;
the step of determining the subject to which the target media work belongs according to the subject associated information of each target media work comprises the following steps:
extracting keywords of each subject from the product label and the plot brief introduction information of each target media product;
determining the score and the value of a keyword corresponding to each target media work in each subject as the subject score of the subject corresponding to the target media work;
and determining the subject associated with the highest subject score corresponding to each target media work as the subject to which the target media work belongs.
5. The method of claim 1, wherein said step of determining a second score for each target media work in each preset narrative dimension associated with the subject matter of the target media work based on a scoring model corresponding to each preset narrative dimension comprises:
extracting a scoring parameter of each preset narrative dimension associated with the subject matter of each target media work from scoring associated information associated with each target media work;
and inputting the scoring parameters of each target media work in each preset narrative dimension into a scoring model corresponding to the preset narrative dimension to obtain a second score of the target media work in the preset narrative dimension.
6. The method of claim 5, wherein the scoring correlation information includes image material and text material;
the scoring model of each preset narrative dimension related to the image material is a convolutional neural network model;
and the scoring model of each preset narrative dimension related to the text data is a Natural Language Processing (NLP) model.
7. The method of claim 5 or 6, further comprising: aiming at each preset narrative dimension, training to obtain a scoring model of the preset narrative dimension by adopting the following steps:
acquiring training data, wherein the training data is grade associated information of the marking grade with the preset narrative dimension;
extracting sample scoring parameters for the preset narrative dimension from the training data;
inputting the sample scoring parameters of the preset narrative dimension into a scoring model corresponding to the preset narrative dimension to obtain the prediction score of the preset narrative dimension;
determining the model loss of the preset narrative dimension according to the prediction score and the marking score of the preset narrative dimension;
if the score model corresponding to the preset narrative dimension is determined to be convergent according to the model loss of the preset narrative dimension, finishing the score model training corresponding to the preset narrative dimension;
if the scoring model corresponding to the preset narrative dimension is determined not to be converged according to the model loss of the preset narrative dimension, adjusting parameters of the scoring model corresponding to the preset narrative dimension, and re-executing the step of inputting the sample scoring parameters of the preset narrative dimension into the scoring model corresponding to the preset narrative dimension to obtain the predicted score of the preset narrative dimension.
8. The method of claim 1, wherein said step of determining a fourth score for each target media work based on a second score for the target media work in each of the predefined narrative dimensions, and a third score for the target media work, comprises:
according to the weight coefficient of each preset narrative dimension and the weight coefficient of the media display platform, carrying out weighting processing on a second score of each target media work in each preset narrative dimension and a third score of the target media work to obtain a fourth score of the target media work; or
Selecting the largest score from the second score for the target media work in each preset narrative dimension and the third score for the target media work as the fourth score for the target media work.
9. The method of claim 1, wherein said determining a narrative ability score for said narrative object on each subject matter according to a fourth score for each target media work for that subject matter comprises:
for each subject, carrying out average processing on the fourth score of each target media work of the subject to obtain narrative ability scores of the narrative objects on the subject; or
For each topic, selecting the largest fourth score from the fourth scores of each target media work of the topic as the narrative ability score of the narrative object on the topic.
10. The method of claim 1, further comprising:
determining a target subject matter of a media work to be created;
obtaining narrative ability scores for a plurality of candidate narrative objects on the target material;
and determining the narrative object of the media work to be authored from the candidate narrative objects with the narrative capability scores larger than the preset score threshold value.
11. A narrative capability assessment apparatus for narrative objects, comprising:
the system comprises a first acquisition module, a second acquisition module and a third acquisition module, wherein the first acquisition module is used for acquiring a plurality of target media works of a narrative object and a first score of each target media work on a media display platform;
the first determination module is used for determining a second score of each target media work in each preset narrative dimension related to the subject matter of the target media work based on the score model corresponding to each preset narrative dimension;
the mapping module is used for mapping the first score of each target media work to the normalized score interval corresponding to the media display platform to obtain a third score of the target media work;
a second determining module, configured to determine a fourth score of each target media work according to a second score of each target media work in each preset narrative dimension and a third score of the target media work;
a third determination module for determining a narrative ability score for said narrative object on each subject matter according to a fourth score for each target media work for that subject matter.
12. An electronic device is characterized by comprising a processor, a communication interface, a memory and a communication bus, wherein the processor and the communication interface are used for realizing mutual communication by the memory through the communication bus;
a memory for storing a computer program;
a processor for implementing the method steps of any of claims 1-10 when executing a program stored in the memory.
13. A computer-readable storage medium, characterized in that a computer program is stored in the computer-readable storage medium, which computer program, when being executed by a processor, carries out the method steps of any one of the claims 1-10.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210970190.4A CN115422918A (en) | 2022-08-12 | 2022-08-12 | Narrative capability evaluation method and device for narrative object |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210970190.4A CN115422918A (en) | 2022-08-12 | 2022-08-12 | Narrative capability evaluation method and device for narrative object |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115422918A true CN115422918A (en) | 2022-12-02 |
Family
ID=84199268
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210970190.4A Pending CN115422918A (en) | 2022-08-12 | 2022-08-12 | Narrative capability evaluation method and device for narrative object |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115422918A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117275319A (en) * | 2023-11-20 | 2023-12-22 | 首都医科大学附属北京儿童医院 | Device for training language emphasis ability |
-
2022
- 2022-08-12 CN CN202210970190.4A patent/CN115422918A/en active Pending
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117275319A (en) * | 2023-11-20 | 2023-12-22 | 首都医科大学附属北京儿童医院 | Device for training language emphasis ability |
CN117275319B (en) * | 2023-11-20 | 2024-01-26 | 首都医科大学附属北京儿童医院 | Device for training language emphasis ability |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110162703B (en) | Content recommendation method, training device, content recommendation equipment and storage medium | |
CN110267119B (en) | Video precision and chroma evaluation method and related equipment | |
Ghadiyaram et al. | Massive online crowdsourced study of subjective and objective picture quality | |
CN110781321B (en) | Multimedia content recommendation method and device | |
CN111444357B (en) | Content information determination method, device, computer equipment and storage medium | |
CN110012356A (en) | Video recommendation method, device and equipment and computer storage medium | |
US9430532B2 (en) | Media content rankings for discovery of novel content | |
US20150243279A1 (en) | Systems and methods for recommending responses | |
US11881010B2 (en) | Machine learning for video analysis and feedback | |
US20170046748A1 (en) | Method and system for personifying a brand | |
JP2011248829A (en) | Evaluation prediction device, evaluation prediction method and program | |
CN113962965A (en) | Image quality evaluation method, device, equipment and storage medium | |
CN114218488A (en) | Information recommendation method and device based on multi-modal feature fusion and processor | |
CN111931073B (en) | Content pushing method and device, electronic equipment and computer readable medium | |
KR20220135150A (en) | Method and device for providing customized cultural and art contents curation and for recommending mate playing the contents with user | |
CN115422918A (en) | Narrative capability evaluation method and device for narrative object | |
CN115618024A (en) | Multimedia recommendation method and device and electronic equipment | |
Potter | 24 A General Framework for Media Psychology Scholarship | |
CN117633371B (en) | Recommendation method, device and readable storage medium based on multi-attention mechanism | |
KR102524461B1 (en) | Method for providing unified interface for product reviews and purchase | |
US11232325B2 (en) | Data analysis system, method for controlling data analysis system, and recording medium | |
CN116980665A (en) | Video processing method, device, computer equipment, medium and product | |
US10970728B2 (en) | System and method for collecting personality information | |
CN113298283A (en) | Content object prediction method and device and content object recommendation method | |
KR102474282B1 (en) | Method and apparatus for providing product purchase service based on review video |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |