CN115936003A - Software function point duplicate checking method, device, equipment and medium based on neural network - Google Patents
Software function point duplicate checking method, device, equipment and medium based on neural network Download PDFInfo
- Publication number
- CN115936003A CN115936003A CN202211522902.2A CN202211522902A CN115936003A CN 115936003 A CN115936003 A CN 115936003A CN 202211522902 A CN202211522902 A CN 202211522902A CN 115936003 A CN115936003 A CN 115936003A
- Authority
- CN
- China
- Prior art keywords
- function point
- point
- checked
- function
- functional
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Machine Translation (AREA)
Abstract
The application discloses a software function point duplicate checking method, device, equipment and medium based on a neural network, wherein the method comprises the following steps: acquiring a software function point list based on a function point method; training the model by using the software function point detail table to obtain a trained model and a function point detail table with marking information, wherein the function point detail table with marking information is used as a historical function point characteristic index; inputting the list of the software function points to be checked into the trained model, and outputting the characteristic indexes of the function points to be checked of each function point in the list of the software function points to be checked; calculating the similarity between the historical function point characteristic index and the function point characteristic index to be checked; and determining similar function points in the software function point list to be checked according to the similarity and the set threshold value, and outputting the software function point list with repeated labeling information. The method and the device improve the efficiency and the accuracy of duplicate checking, and reduce the project repeated construction cost.
Description
Technical Field
The present application relates to the field of computer technologies, and in particular, to a method, an apparatus, a device, and a medium for checking duplicate data of a software function based on a neural network.
Background
The functional point method (IFPUG-FPA) is a software scale measurement method which is based on user requirements, has specific algorithm support, is objectively independent of technology implementation, can be estimated before actual development, has the characteristics of strong scientificity, authority, operability and the like, and is a method adopted by national standards and industrial standards. The function point counting process refers to a process of quantifying function user requirements and evaluating non-function point user requirements to count application system software functions.
Considering that more and more software projects exist, repeated or similar projects exist in the same field, the same or similar function points exist in the same project and the like, unreasonable function items can be found through repeated checking of the function points, and repeated construction is effectively avoided. However, a large number of software function points of projects are accumulated, manual duplicate checking is difficult to perform only, the labor cost and the waiting cost are inevitably huge, and meanwhile, the function point data needs to be identified and judged manually one by one, so that the factors such as subjectivity, physiological fatigue and the like are difficult to overcome, and the duplicate checking quality is difficult to ensure.
Therefore, how to effectively reduce the time and cost for checking the duplicate of the function point by the user and improve the duplicate checking accuracy rate is a problem that the technical personnel needs to solve urgently.
Disclosure of Invention
The application provides a software function point duplicate checking method based on a neural network on one hand, and aims to solve the technical problems of high time and cost, and low efficiency and accuracy in the process of function point duplicate checking in the prior art.
The technical scheme adopted by the application is as follows:
a software function point duplicate checking method based on a neural network comprises the following steps:
acquiring a software function point list based on a function point method, wherein the software function point list comprises function point hierarchical structure information and function point category information;
training a model by using a software function point detail table to obtain a trained model and a function point detail table with label information, wherein the function point detail table with label information is used as a historical function point characteristic index, and the historical function point characteristic index comprises a historical function point grade, a historical function point word vector and a historical function point type;
inputting a detailed table of software function points to be checked into the trained model, and outputting feature indexes of the function points to be checked of each function point in the detailed table of the software function points to be checked, wherein the feature indexes of the function points to be checked comprise the grade of the function points to be checked, word vectors of the function points to be checked and the types of the function points to be checked;
calculating the similarity between the historical function point characteristic index and the function point characteristic index to be checked;
and determining similar function points in the software function point detail table to be checked according to the similarity and the set threshold value, and outputting the software function point detail table with repeated labeling information to obtain a software function point check result.
Preferably, the training of the model by using the software function point list to obtain the trained model and the function point list with the labeling information, wherein the function point list with the labeling information is used as a characteristic index of a historical function point, specifically comprises the following steps:
reading the hierarchical structure in the software function point list through a function point hierarchical model to form a function point text, describing the hierarchical structure of the function point text through natural language, adding the hierarchical structure into the function point text, and completing the text data of the function point;
acquiring the supplemented functional point text data and massive Chinese data to train a word vector model, so that the word vector model can understand the real semantics of the functional point text;
and training the model by taking the supplemented functional point text data as a training corpus and the functional point categories as training labels to obtain a functional point classification model for directly extracting key semantic features of the functional point text data, and calculating the functional point category probability based on the key semantic features.
Preferably, the acquiring the supplemented functional point text data and massive chinese data to train a word vector model, so that the word vector model can understand the true semantics of the functional point text, specifically comprising the steps of:
constructing a training corpus: constructing a self-training corpus for a word vector model from a text which comprises supplemented functional point text data and massive Chinese data and is completely free of marking information;
constructing a word vector model: constructing a word vector model based on CBOW, wherein the word vector model deduces missing real words by using text context to obtain more accurate text semantics and converts the words into vectors representing the text semantics, and the word vector distance represented by the words with more similar semantics is closer;
training a word vector model: training the word vector model by using the prepared self-training corpus to enable the word vector model to have semantic comprehension capability;
and (3) storing a word vector model: and storing the trained word vector model for directly calculating the word vector of the functional point.
Preferably, the supplemented functional point text data is used as a training corpus, the functional point categories are used as training labels, the model is trained, a functional point classification model is obtained and used for directly extracting key semantic features of the functional point text data, and functional point category probability calculation is performed based on the key semantic features, and the method specifically comprises the following steps:
constructing a training corpus: the supplemented functional point text data is used as a training corpus, and the functional point category is used as a label to ensure the accuracy of the training data;
constructing a functional point classification model: the functional point classification model adopts a convolutional neural network model and comprises an input layer, a convolutional layer, a pooling layer, a full-link layer and an output layer;
training a functional point classification model: training a function point classification model after a preprocessing process is carried out on the prepared function point text data with class information, adjusting model parameters through a back propagation technology, and ensuring that the model is trained to an optimal state through adjusting hyper-parameters for multiple times, so as to classify the function point text data accurately finally;
and outputting a classification result: and storing the trained function point classification model for classifying the software function points.
Preferably, the step of inputting the detailed table of the software function points to be checked into the trained model and outputting the feature index of the function point to be checked of each function point in the detailed table of the software function points to be checked specifically includes the steps of:
reading the hierarchical structure in the software function point list to be checked and duplicated by the function point hierarchical model to form a function point text, describing the hierarchical structure of the function point text by natural language, adding the hierarchical structure into the function point text, complementing the text data of the function point, and outputting and storing the level of the function point to be checked and duplicated;
the functional point word vector model deduces missing real words by using the functional point text data context to obtain more accurate text semantics, converts the words into vectors representing the text semantics, and outputs and stores functional point word vectors to be searched;
the functional point classification model identifies the hidden relation between texts by extracting the key semantic features of the functional point text data, correctly classifies the functional point texts, and outputs and stores each type of the functional points to be checked in the functional point detail table of the software to be checked.
Preferably, the calculating the similarity between the historical function point feature index and the function point feature index to be checked specifically includes the steps of:
respectively calculating the level similarity of the function points to be checked and duplicated, the word vector similarity of the function points to be checked and duplicated and the type similarity of the function points to be checked and duplicated;
calculating the similarity of the function points to be checked and found according to the level similarity of the function points to be checked and found, the word vector similarity of the function points to be checked and found and the type similarity of the function points to be checked and found in a weighting manner:
W＝α*X+β*Y+θ*Z
in the formula:
alpha, beta and theta are weights occupied by corresponding indexes;
w is the similarity of the functional points to be checked;
x is the similarity of the level of the functional points to be checked;
y is the similarity of the functional point word vectors to be checked;
and Z is the similarity of the types of the functional points to be checked.
Preferably, the step of calculating the level similarity of the functional points to be found and found, the word vector similarity of the functional points to be found and the type similarity of the functional points to be found and found respectively includes the following steps:
when calculating the level similarity of the functional points to be checked and repeated, determining the level similarity of the functional points to be checked and repeated according to the level depth of the functional points to be checked and repeated in the functional module by adopting a step-by-step comparison principle, wherein the level similarity of the functional points to be checked and repeated is positively correlated with the level depth of the functional points to be checked and repeated in the functional module;
when the similarity of the word vectors of the functional points to be checked and found is calculated, the similarity between the historical functional point word vectors and the cosine similarity of the word vectors of the functional points to be checked and found is obtained by converting the text data of the functional points into the word vectors;
and when calculating the similarity of the types of the functional points to be checked, dividing the types of the functional points output by the model into EI, EQ, EO, ELF and ILF, and setting the similarity between different functional points according to the different types of the functional points.
The application also provides a software function point duplicate checking device based on the neural network, which comprises:
the software function point list acquisition module is used for acquiring a software function point list based on a function point method, wherein the software function point list comprises function point hierarchical structure information and function point category information;
the historical function point feature index calculation module is used for training the model by using the software function point detail table to obtain a trained model and a function point detail table with label information, wherein the function point detail table with label information is used as a historical function point feature index, and the historical function point feature index comprises a historical function point grade, a historical function point word vector and a historical function point type;
the function point feature index calculation module to be checked is used for inputting the software function point detailed table to be checked into the trained model and outputting the function point feature index to be checked of each function point in the software function point detailed table to be checked, wherein the function point feature index to be checked comprises the function point grade to be checked, the function point word vector to be checked and the function point type to be checked;
the similarity calculation module is used for calculating the similarity between the historical function point characteristic indexes and the function point characteristic indexes to be checked;
and the function point duplicate checking result output module is used for determining similar function points in the software function point detail table to be checked according to the similarity and the set threshold value, outputting the software function point detail table with repeated marking information and obtaining a software function point duplicate checking result.
The present application also provides an electronic device including a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor executes the program to implement the steps of the neural network-based software function duplication checking method.
In another aspect, the present application further provides a storage medium, where the storage medium includes a stored program, and when the program runs, the apparatus on which the storage medium is located is controlled to execute the steps of the neural network-based software function duplication checking method.
Compared with the prior art, the method has the following beneficial effects:
the method and the device have the advantages that the manual mode is replaced by the machine mode to check the functional points, the efficiency and the accuracy of checking the duplicate are improved, and project repeated construction is reduced. Meanwhile, the method is inspired by big data clustering technology, based on the fact that the function point texts are converted into word vectors with context semantic features, and a fast indexing algorithm for comparing massive function point texts is supported, the similarity between the function point texts can be rapidly compared through the method, the calculation amount required by duplicate checking is reduced to 1/100 of the original amount, the duplicate checking of 10 ten thousand historical requirements is guaranteed to be completed within 0.1 second, the duplicate checking efficiency is greatly improved, and the duplicate checking of massive data can be rapidly completed within a short time. According to the method, the similarity of the function points is calculated by adopting a plurality of dimension indexes for comparison, and then the similarity of the function points is calculated in a weighting mode, so that the similarity of the text of the function points is calculated by adopting the cosine similarity of single dimension word vectors, the hierarchy and type comparison of the function points is increased, and the accuracy of duplicate checking is effectively improved.
In addition to the objects, features and advantages described above, other objects, features and advantages will be apparent from the present application. The present application will now be described in further detail with reference to the accompanying drawings.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate embodiments of the application and, together with the description, serve to explain the application and are not intended to limit the application. In the drawings:
fig. 1 is a schematic flow chart of a software function point duplicate checking method based on a neural network according to a preferred embodiment of the present application.
Fig. 2 is a detailed table diagram of software functional points in the preferred embodiment of the present application.
Fig. 3 is a flow chart illustrating the sub-steps of step S2 according to the preferred embodiment of the present application.
Fig. 4 is a flow chart illustrating the sub-steps of step S22 according to the preferred embodiment of the present application.
Fig. 5 is a flow chart illustrating the sub-steps of step S23 according to the preferred embodiment of the present application.
Fig. 6 is a flow chart illustrating the sub-steps of step S3 according to the preferred embodiment of the present application.
Fig. 7 is a flow chart illustrating the sub-steps of step S4 according to the preferred embodiment of the present application.
Fig. 8 is a flow chart illustrating the sub-steps of step S41 according to the preferred embodiment of the present application.
Fig. 9 is a schematic diagram of a similarity calculation process of the to-be-reviewed function point level according to the preferred embodiment of the present application.
Fig. 10 is a block diagram of a software function duplication checking device based on a neural network according to a preferred embodiment of the present application.
Fig. 11 is a schematic block diagram of an electronic device entity of the preferred embodiment of the present application.
Fig. 12 is an internal structural view of a computer device according to a preferred embodiment of the present application.
Detailed Description
It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict. The present application will be described in detail below with reference to the embodiments with reference to the attached drawings.
Referring to fig. 1, a preferred embodiment of the present application provides a software function duplication checking method based on a neural network, including the steps of:
s1, acquiring a software function point detail table based on a function point method, wherein the software function point detail table comprises function point hierarchical structure information and function point category information, and as shown in figure 2, the software function point detail table is generated in the process of evaluating project workload by using the function point method for an enterprise and is formed through precipitation and accumulation. The software function point list clearly records information such as the hierarchical structure, the function point type and the like of the function point, the data is subjected to multiple evaluations and audits, the quality of the data is checked layer by layer, and the data quality and the reliability can be fully ensured;
s2, training the model by using the software function point detail table to obtain a trained model and a function point detail table with label information, wherein the function point detail table with label information is used as a historical function point characteristic index, and the historical function point characteristic index comprises a historical function point grade, a historical function point word vector and a historical function point type;
s3, inputting a detailed list of the software function points to be checked into the trained model, and outputting feature indexes of the function points to be checked of each function point in the detailed list of the software function points to be checked, wherein the feature indexes of the function points to be checked comprise the grade of the function points to be checked, the word vectors of the function points to be checked and the types of the function points to be checked;
s4, calculating the similarity between the historical function point characteristic indexes and the function point characteristic indexes to be checked;
and S5, determining similar function points in the software function point list to be checked according to the similarity and the set threshold value, and outputting the software function point list with repeated marking information to obtain a software function point check result.
In the embodiment, a machine mode replaces a manual mode to check the function points, so that the efficiency and the accuracy of checking the function points are improved, and the repeated construction of projects is reduced. Meanwhile, the embodiment is inspired by big data clustering technology, based on the fact that the function point texts are converted into word vectors with context semantic features, and a fast indexing algorithm for comparing massive function point texts is supported, the similarity between the function point texts can be rapidly compared through the method, the calculation amount required by duplicate checking is reduced to 1/100 of the original amount, the duplicate checking of 10 ten thousand historical requirements can be completed within 0.1 second, the duplicate checking efficiency is greatly improved, and the duplicate checking of massive data can be rapidly completed within a short time. In the embodiment, the functional point similarity is calculated by comparing a plurality of dimensional indexes and calculating the functional point similarity in a weighting mode, and compared with the method of calculating the functional point text similarity by using single dimensional word vector cosine similarity, the functional point hierarchy and type comparison is increased, and the accuracy of duplicate checking is effectively improved.
Preferably, as shown in fig. 3, the training of the model by using the software function point detail table to obtain the trained model and the function point detail table with the labeling information, where the function point detail table with the labeling information is used as the characteristic index of the historical function point, specifically includes the steps of:
s21, reading the hierarchical structure (including first level, second level, third level, fourth level and the like) in the software function point list through a function point hierarchical model to form a function point text, describing the hierarchical structure of the function point text through natural language, adding the hierarchical structure into the function point text, and completing the text data of the function point, such as: the original function point is described as 'login function point single sign on', the completed text is 'the first-level function is the login function, and the second-level function is the single sign on'. The information completion of the function point text can help a subsequent model to understand the hierarchical information of the function point text, and more reasonable judgment is made;
taking the function point with the sequence number of 1 in fig. 2 as an example, the function point text read by the machine is "face basic information management face local management face basic management user group creation", and the completed text is: the method comprises the following steps of functional module face basic information management, wherein a first-level function is face local management, a second-level function is face basic management, a third-level function is face basic management, and a fourth-level function is user group creation.
And S22, acquiring the supplemented functional point text data and massive Chinese data to train a word vector model, so that the word vector model can understand the real semantics of the functional point text. The word vector model can map words to a high-dimensional space, and the distances between words which are thought to be close in the high-dimensional space are close to each other, so that a computer simulates the process of reading and recognizing characters by people, and the computer has the capability of semantic understanding. In order to facilitate subsequent training, the function point texts which are graded in the previous step are classified according to the following steps of 8:2, dividing the training data and the test data into training data and test data at random in proportion;
and S23, training the model by taking the supplemented functional point text data as training corpora and the functional point categories as training labels to obtain a functional point classification model for directly extracting key semantic features of the functional point text data, and calculating the functional point category probability based on the key semantic features.
Preferably, as shown in fig. 4, the acquiring the supplemented functional point text data and massive chinese data to train a word vector model, so that the word vector model can understand the real semantics of the functional point text, specifically includes the steps of:
s221, constructing a training corpus: constructing a self-training corpus for a word vector model from a text which comprises supplemented functional point text data and massive Chinese data and is completely free of marking information;
s222, constructing a word vector model: constructing a word vector model based on CBOW, wherein the word vector model deduces missing real words by using text context to obtain more accurate text semantics and converts the words into vectors representing the text semantics, and the word vector distance represented by the words with more similar semantics is closer;
s223, training a word vector model: the word vector model is trained by using the prepared self-training corpus, so that the word vector model has semantic understanding capability, a large amount of training corpus is required to be supplied for the word vector model to train, parameters of the model are continuously adjusted by a back propagation technology, and the model is trained to an optimal state by adjusting the hyper-parameters for multiple times, so that the word vector model can have the semantic understanding capability, and the distance relation of two functional point word vectors on a two-dimensional equal-height plane can be viewed;
s224, storing a word vector model: and storing the trained word vector model for directly calculating the word vector of the functional point.
Preferably, as shown in fig. 5, the method for performing function point classification based on the complemented function point text data as a training corpus and the complemented function point category as a training label trains a model to obtain a function point classification model for directly extracting key semantic features of the function point text data, and performs function point category probability calculation based on the key semantic features specifically includes the steps of:
s231, constructing a training corpus: the supplemented functional point text data is used as a training corpus, and the functional point category is used as a label to ensure the accuracy of the training data;
s232, constructing a functional point classification model: the functional point classification model adopts a convolutional neural network model and comprises an input layer, a convolutional layer, a pooling layer, a full-link layer and an output layer, and the convolutional operation is performed on the convolutional layer at certain intervals by sliding a window of a filter at all positions once, so that the output of the convolutional operation can be obtained; extracting primary features by using local word sequence information in a pooling layer, combining the primary features into high-level features, and omitting the step of feature engineering in the traditional machine learning through convolution and pooling operations; the full-connection layer inputs the features extracted by the convolution and pooling layers into a classifier for classification, and the output layer adopts a Softmax classification model to classify the functional point text data into 5 categories which respectively correspond to 'EI', 'EQ', 'EO', 'EIL' and 'ILF' of the functional point text;
s233, training a functional point classification model: training a function point classification model after a preprocessing process is carried out on the prepared function point text data with class information, adjusting model parameters through a back propagation technology, and ensuring that the model is trained to an optimal state through adjusting hyper-parameters for multiple times, so as to classify the function point text data accurately finally;
s234, outputting a classification result: and storing the trained function point classification model for classifying the software function points.
Preferably, as shown in fig. 6, the inputting the detailed table of software function points to be checked into the trained model, and outputting the feature index of the function point to be checked of each function point in the detailed table of software function points to be checked specifically includes the steps of:
s31, reading a hierarchical structure in a detailed list of the function points of the software to be checked and duplicated by a function point hierarchical model to form a function point text, describing the hierarchical structure of the function point text through natural language, adding the hierarchical structure into the function point text, supplementing the text data of the function points, and outputting and storing the level of the function points to be checked and duplicated;
s32, deducing missing real words by using the context of the functional point text data through a functional point word vector model, acquiring more accurate text semantics, converting the words into vectors representing the text semantics, and outputting and storing the functional point word vectors to be searched;
and S33, recognizing the hidden relation among the texts by the functional point classification model through extracting the key semantic features of the functional point text data, correctly classifying the functional point texts, and outputting and storing the type of each functional point to be checked in the functional point detailed list of the software to be checked.
Preferably, as shown in fig. 7, the calculating the similarity between the historical function point feature index and the function point feature index to be repeatedly checked specifically includes the steps of:
s41, calculating the level similarity, word vector similarity and type similarity of the functional points to be checked and duplicated respectively;
specifically, as shown in fig. 8, the calculating the level similarity of the function points to be checked and duplicated, the word vector similarity of the function points to be checked and duplicated, and the type similarity of the function points to be checked and duplicated respectively includes the following steps:
s411, as shown in fig. 9, when calculating the level similarity of the function points to be checked and repeated, determining the level similarity of the function points to be checked and repeated according to the level depth of the function points to be checked and repeated in the function module by using a step-by-step comparison principle, where the level similarity of the function points to be checked and the level depth of the function points to be checked and repeated in the function module are positively correlated, that is, the level similarity (X) of the function points is shown in table 1:
table 1:
in this embodiment, according to the historical function point text level and the level lookup table 1 of the function point to be checked, if the type of the function point to be checked is similar to the historical function point text level in the second level, the similarity X of the function point to be checked in the level is 60%.
S412, when calculating the similarity of the word vectors of the functional points to be found, calculating the cosine similarity of the word vectors A and B of the historical functional points by converting the text data of the functional points into the word vectors to obtain the similarity between the word vectors A and B of the functional points to be found:
in the formula: n represents the dimensional degree of the space vector, and i represents the ith dimensional space;
in this embodiment, the above formula is used to calculate according to the word vectors of the historical function points and the word vectors of the function points to be found, and if the word vectors of the historical function points are a (1, 3, 2) and the word vectors of the function points to be found are B (2, 4, 5), the similarity Y of the word vectors of the function points to be found is:
s413, when calculating the similarity of the types of the function points to be checked, dividing the types of the function points output by the model into five types, i, EQ, EO, ELF, and ILF, and setting the type similarity (Z) between different function points according to the difference of the types of the function points, as shown in table 2:
table 2:
in this embodiment, according to the lookup table 2 for the types of the history function points and the types of the function points to be checked, if the type of the history function point is EI and the type of the function point to be checked is EO, the similarity Z of the types of the function points to be checked is 60%;
s42, calculating similarity of the functional points to be checked according to the level similarity of the functional points to be checked, the word vector similarity of the functional points to be checked and the type similarity of the functional points to be checked and checked in a weighting mode:
W＝α*X+β*Y+θ*Z＝30％*60％+40％*95.63％+30％*60％＝74.25％，
in the formula:
and alpha, beta and theta are weights occupied by corresponding indexes and are respectively set as: 30%, 40% and 30%;
w is the similarity of the functional points to be checked;
x is the similarity of the level of the functional points to be checked;
y is the similarity of the functional point word vectors to be checked;
and Z is the similarity of the types of the functional points to be checked.
As shown in fig. 10, another aspect of the present application further provides a software function duplication checking apparatus based on a neural network, including:
the software function point list acquisition module is used for acquiring a software function point list based on a function point method, wherein the software function point list comprises function point hierarchical structure information and function point category information;
the historical function point feature index calculation module is used for training the model by using the software function point detail table to obtain a trained model and a function point detail table with label information, wherein the function point detail table with label information is used as a historical function point feature index, and the historical function point feature index comprises a historical function point grade, a historical function point word vector and a historical function point type;
the function point feature index calculation module to be checked is used for inputting the software function point detail table to be checked into the trained model and outputting the function point feature index to be checked of each function point in the software function point detail table to be checked, wherein the function point feature index to be checked comprises the function point grade to be checked, the function point word vector to be checked and the function point type to be checked;
the similarity calculation module is used for calculating the similarity between the historical function point characteristic indexes and the function point characteristic indexes to be checked;
and the function point duplicate checking result output module is used for determining similar function points in the software function point detail table to be checked according to the similarity and the set threshold value, outputting the software function point detail table with repeated marking information and obtaining a software function point duplicate checking result.
As shown in fig. 11, the preferred embodiment of the present application further provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and when the processor executes the computer program, the steps of the neural network-based software function click repetition method in the above embodiments are implemented.
As shown in fig. 12, the preferred embodiment of the present application also provides a computer device, which may be a terminal or a biopsy server, and the internal structure thereof may be as shown in fig. 10. The computer device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communicating with other external computer devices through network connection. The computer program is executed by a processor to realize the steps of the software function point duplicate checking method based on the neural network.
Those skilled in the art will appreciate that the architecture shown in fig. 12 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
The preferred embodiment of the present application further provides a storage medium, where the storage medium includes a stored program, and when the program runs, the apparatus where the storage medium is located is controlled to execute the steps of the software function point duplicate checking method based on the neural network in the foregoing embodiment.
It should be noted that the steps illustrated in the flowcharts of the figures may be performed in a computer system such as a set of computer-executable instructions and that, although a logical order is illustrated in the flowcharts, in some cases, the steps illustrated or described may be performed in an order different than presented herein.
If the functions of the method of the present embodiment are implemented in the form of software functional units and sold or used as independent products, the functions may be stored in one or more storage media readable by a computing device. Based on such understanding, part of the contribution to the prior art of the embodiments of the present application or part of the technical solution may be embodied in the form of a software product stored in a storage medium and including several instructions for causing a computing device (which may be a personal computer, a server, a mobile computing device or a network device) to execute all or part of the steps of the method described in the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein. The scheme in the embodiment of the application can be implemented by adopting various computer languages, such as object-oriented programming language Java and transliterated scripting language JavaScript.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While the preferred embodiments of the present application have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all alterations and modifications as fall within the scope of the application.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present application without departing from the spirit and scope of the application. Thus, if such modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is intended to include such modifications and variations as well.
Claims (10)
1. A software function point duplicate checking method based on a neural network is characterized by comprising the following steps:
acquiring a software function point list based on a function point method, wherein the software function point list comprises function point hierarchical structure information and function point category information;
training a model by using a software function point detail table to obtain a trained model and a function point detail table with label information, wherein the function point detail table with label information is used as a historical function point characteristic index, and the historical function point characteristic index comprises a historical function point grade, a historical function point word vector and a historical function point type;
inputting a detailed table of software function points to be checked into the trained model, and outputting feature indexes of the function points to be checked of each function point in the detailed table of the software function points to be checked, wherein the feature indexes of the function points to be checked comprise the grade of the function points to be checked, word vectors of the function points to be checked and the types of the function points to be checked;
calculating the similarity between the historical function point characteristic index and the function point characteristic index to be checked;
and determining similar function points in the software function point detail table to be checked according to the similarity and the set threshold value, and outputting the software function point detail table with repeated labeling information to obtain a software function point check result.
2. The method for checking the duplication of software function points based on the neural network as claimed in claim 1, wherein the method for training the model by using the software function point detail table to obtain the trained model and the function point detail table with the labeling information, the function point detail table with the labeling information is used as the characteristic index of the historical function point, and specifically comprises the following steps:
reading a hierarchical structure in the software function point list through a function point hierarchical model to form a function point text, describing the hierarchical structure of the function point text through natural language, adding the function point text into the function point text, and completing function point text data;
acquiring the supplemented functional point text data and massive Chinese data to train a word vector model, so that the word vector model can understand the real semantics of the functional point text;
and training the model by taking the supplemented functional point text data as a training corpus and the functional point categories as training labels to obtain a functional point classification model for directly extracting key semantic features of the functional point text data, and calculating the functional point category probability based on the key semantic features.
3. The software function point duplication checking method based on the neural network as claimed in claim 2, wherein the step of acquiring the complemented function point text data and massive Chinese data to train a word vector model so that the word vector model can understand the real semantics of the function point text specifically comprises the steps of:
constructing training corpora: constructing a self-training corpus for a word vector model from a text which comprises supplemented functional point text data and massive Chinese data and is completely free of marking information;
constructing a word vector model: constructing a word vector model based on CBOW, wherein the word vector model deduces missing real words by using text context to obtain more accurate text semantics and converts the words into vectors representing the text semantics, and the word vector distance represented by the words with more similar semantics is closer;
training a word vector model: training the word vector model by using the prepared self-training corpus to enable the word vector model to have semantic comprehension capability;
and (3) storing a word vector model: and storing the trained word vector model for directly calculating the word vector of the functional point.
4. The software function point duplication checking method based on the neural network as claimed in claim 2, wherein the supplemented function point text data is used as a training corpus and function point categories are used as training labels, the model is trained to obtain a function point classification model for directly extracting key semantic features of the function point text data, and the function point category probability calculation is performed based on the key semantic features, specifically comprising the steps of:
constructing a training corpus: the supplemented functional point text data is used as a training corpus, and the functional point category is used as a label to ensure the accuracy of the training data;
constructing a functional point classification model: the functional point classification model adopts a convolutional neural network model and comprises an input layer, a convolutional layer, a pooling layer, a full-link layer and an output layer;
training a functional point classification model: training a function point classification model after a preprocessing process is carried out on the prepared function point text data with class information, adjusting model parameters through a back propagation technology, and ensuring that the model is trained to an optimal state through adjusting hyper-parameters for multiple times, so as to classify the function point text data accurately finally;
outputting a classification result: and storing the trained function point classification model for classifying the software function points.
5. The software function point duplicate checking method based on the neural network as claimed in claim 1, wherein the software function point list to be duplicated is input into the trained model, and the feature index of the function point to be duplicated of each function point in the software function point list to be duplicated is output, specifically comprising the steps of:
reading the hierarchical structure in the software function point list to be checked and duplicated by the function point hierarchical model to form a function point text, describing the hierarchical structure of the function point text by natural language, adding the hierarchical structure into the function point text, complementing the text data of the function point, and outputting and storing the level of the function point to be checked and duplicated;
the functional point word vector model deduces missing real words by using the context of the functional point text data to obtain more accurate text semantics, and outputs and stores functional point word vectors to be searched after the words are converted into vectors representing the text semantics;
the functional point classification model identifies the hidden relation between texts by extracting the key semantic features of the functional point text data, correctly classifies the functional point texts, and outputs and stores each type of the functional points to be checked in the functional point detail table of the software to be checked.
6. The method for duplicate checking of software function points based on a neural network as claimed in claim 1, wherein the calculating the similarity between the characteristic index of the historical function point and the characteristic index of the function point to be checked specifically comprises the steps of:
respectively calculating the level similarity of the function points to be checked and duplicated, the word vector similarity of the function points to be checked and duplicated and the type similarity of the function points to be checked and duplicated;
calculating the similarity of the functional points to be checked according to the level similarity of the functional points to be checked, the word vector similarity of the functional points to be checked and the type similarity of the functional points to be checked and checked in a weighting manner:
W＝α*X+β*Y+θ*Z
in the formula:
alpha, beta and theta are weights occupied by corresponding indexes;
w is the similarity of the functional points to be checked;
x is the similarity of the level of the functional points to be checked;
y is the similarity of the functional point word vectors to be checked;
and Z is the similarity of the types of the functional points to be checked.
7. The software function point duplicate checking method based on the neural network as claimed in claim 6, wherein the step of calculating the level similarity of the function points to be duplicated, the word vector similarity of the function points to be duplicated, and the type similarity of the function points to be duplicated respectively comprises the following steps:
when calculating the level similarity of the functional points to be checked and duplicated, determining the level similarity of the functional points to be checked and duplicated according to the level depth of the functional points to be checked and duplicated in the functional module by adopting a step-by-step comparison principle, wherein the level similarity of the functional points to be checked and duplicated is positively correlated with the level depth of the functional points to be checked and duplicated in the functional module;
when the similarity of the word vectors of the functional points to be found is calculated, the similarity between the word vectors of the historical functional points and the word vectors of the functional points to be found is obtained by calculating the cosine similarity of the word vectors of the historical functional points and the word vectors of the functional points to be found;
and when calculating the similarity of the types of the functional points to be checked, dividing the types of the functional points output by the model into EI, EQ, EO, ELF and ILF, and setting the similarity between different functional points according to the different types of the functional points.
8. A software function point duplicate checking device based on a neural network is characterized by comprising:
the software function point list acquisition module is used for acquiring a software function point list based on a function point method, wherein the software function point list comprises function point hierarchical structure information and function point category information;
the historical function point feature index calculation module is used for training the model by using the software function point detail table to obtain a trained model and a function point detail table with label information, wherein the function point detail table with label information is used as a historical function point feature index, and the historical function point feature index comprises a historical function point grade, a historical function point word vector and a historical function point type;
the function point feature index calculation module to be checked is used for inputting the software function point detailed table to be checked into the trained model and outputting the function point feature index to be checked of each function point in the software function point detailed table to be checked, wherein the function point feature index to be checked comprises the function point grade to be checked, the function point word vector to be checked and the function point type to be checked;
the similarity calculation module is used for calculating the similarity between the historical function point characteristic indexes and the function point characteristic indexes to be checked;
and the function point duplicate checking result output module is used for determining similar function points in the software function point detailed table to be checked according to the similarity and the set threshold value, and outputting the software function point detailed table with repeated marking information to obtain a software function point duplicate checking result.
9. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor,
the processor, when executing the program, implements the steps of the neural network based software function point duplication checking method of any one of claims 1 to 7.
10. A storage medium including a stored program, characterized in that,
controlling a device on which the storage medium is located to perform the steps of the neural network-based software function point duplication checking method according to any one of claims 1 to 7 when the program is run.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211522902.2A CN115936003A (en) | 2022-11-30 | 2022-11-30 | Software function point duplicate checking method, device, equipment and medium based on neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211522902.2A CN115936003A (en) | 2022-11-30 | 2022-11-30 | Software function point duplicate checking method, device, equipment and medium based on neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115936003A true CN115936003A (en) | 2023-04-07 |
Family
ID=86650067
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211522902.2A Pending CN115936003A (en) | 2022-11-30 | 2022-11-30 | Software function point duplicate checking method, device, equipment and medium based on neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115936003A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117251445A (en) * | 2023-10-11 | 2023-12-19 | 杭州今元标矩科技有限公司 | Deep learning-based CRM data screening method, system and medium |
-
2022
- 2022-11-30 CN CN202211522902.2A patent/CN115936003A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117251445A (en) * | 2023-10-11 | 2023-12-19 | 杭州今元标矩科技有限公司 | Deep learning-based CRM data screening method, system and medium |
| CN117251445B (en) * | 2023-10-11 | 2024-06-04 | 杭州今元标矩科技有限公司 | Deep learning-based CRM data screening method, system and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110084239B (en) | Method for reducing overfitting of network training during off-line handwritten mathematical formula recognition | |
| CN111046670B (en) | Entity and relationship combined extraction method based on drug case legal documents | |
| CN112434535B (en) | Element extraction method, device, equipment and storage medium based on multiple models | |
| CN113191148B (en) | Rail transit entity identification method based on semi-supervised learning and clustering | |
| CN113673254B (en) | Knowledge distillation position detection method based on similarity maintenance | |
| CN109902271A (en) | Text data mask method, device, terminal and medium based on transfer learning | |
| CN111274817A (en) | Intelligent software cost measurement method based on natural language processing technology | |
| CN115357719B (en) | Power audit text classification method and device based on improved BERT model | |
| CN113138920B (en) | Software defect report allocation method and device based on knowledge graph and semantic role labeling | |
| US20240078389A1 (en) | Generative language model for few-shot aspect-based sentiment analysis | |
| CN116205482A (en) | Important personnel risk level assessment method and related equipment | |
| CN109933783A (en) | A kind of essence of a contract method of non-performing asset operation field | |
| Garrido-Munoz et al. | A holistic approach for image-to-graph: application to optical music recognition | |
| CN115936003A (en) | Software function point duplicate checking method, device, equipment and medium based on neural network | |
| CN118332138B (en) | Navigation notification text processing method, computer program product and terminal | |
| CN117975464A (en) | Identification method and system of electric secondary drawing text information based on U-Net | |
| CN117421226A (en) | Defect report reconstruction method and system based on large language model | |
| CN114692615B (en) | Small sample intention recognition method for small languages | |
| CN117390454A (en) | Data labeling method and system based on multi-domain self-adaptive data closed loop | |
| CN116562294A (en) | Bridge text small sample named entity recognition method based on prompt learning | |
| CN117272228A (en) | Data fusion method, system, equipment and medium based on heterogeneous graph embedding | |
| CN115964486A (en) | Small sample intention recognition method based on data enhancement | |
| CN111046934B (en) | SWIFT message soft clause recognition method and device | |
| CN116089605A (en) | Text emotion analysis method based on transfer learning and improved word bag model | |
| CN115422349A (en) | Hierarchical text classification method based on pre-training generation model |
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 |