CN116720646A

CN116720646A - Division method, division device and electronic equipment

Info

Publication number: CN116720646A
Application number: CN202210191464.XA
Authority: CN
Inventors: 吴嫒博; 刘萌; 孙朝旭; 蔡晓凤; 卢鑫鑫; 叶礼伟; 李果; 杨晖; 覃伟枫; 滕达
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2022-02-28
Filing date: 2022-02-28
Publication date: 2023-09-08

Abstract

The application provides a learning area dividing method, a learning area dividing device and electronic equipment, wherein the method relates to the technical field of cloud education of cloud technology, and comprises the following steps: acquiring a learning level prediction model obtained through learning; the method comprises the steps of obtaining a feature vector of each cell by carrying out feature construction on original data of each cell in m cells; m is an integer greater than 0; inputting the feature vector of each cell into the degree prediction model to obtain the degree demand number of each cell; acquiring the number of the degree supplies of each of n schools; n is an integer greater than 0; the number of the academic demands of each cell and the number of the academic supplies of each school are input into a division model to obtain division results for dividing the m cells into the n schools. The method provided by the application not only can reduce the entrance difficulty of schools, but also can ensure the balanced development of education.

Description

Division method, division device and electronic equipment

Technical Field

The embodiment of the application relates to the technical field of cloud education of cloud technology, in particular to a learning area dividing method, a learning area dividing device and electronic equipment.

Background

The traditional method for dividing the learning area mainly adopts an operation planning linear programming method, namely a learning area dividing model is built according to information such as schools, residential points, road networks and the like, so that the division of the learning area is realized. In general, each school district may include one to several public schools, a primary school district may be provided with only one primary school, a large school district may be provided with several primary schools or middle schools, a plurality of school schools are provided, and students choose to read schools on the basis of nearby entrance of adjacent areas. However, the fertility policy is considered to bring about an increase in fertility rate, and thus a change in the size of the school-age population; in addition, with the development of town, more follow-up children enter a first city and a second city along with parents, so that some school districts cannot accommodate more and more children of proper age.

In addition, when the school is read by adopting the principle of nearby entrance, only the distance between the resident point and the school is considered, and the problems that part of schools are difficult to enter and the schools need to be read by selecting are solved. In addition, it is difficult to ensure education balance development by the conventional division method. For example, there is a possibility that the academic feed of schools may be unbalanced.

Disclosure of Invention

The application provides a learning area dividing method, a learning area dividing device and electronic equipment, which can not only reduce the difficulty of learning in schools, but also ensure the balanced development of education.

In a first aspect, the present application provides a method for distinguishing between regions, comprising:

acquiring a learning level prediction model obtained through learning;

the method comprises the steps of obtaining a feature vector of each cell by carrying out feature construction on original data of each cell in m cells; m is an integer greater than 0;

inputting the feature vector of each cell into the degree prediction model to obtain the degree demand number of each cell;

acquiring the number of the degree supplies of each of n schools; n is an integer greater than 0;

the number of the academic demands of each cell and the number of the academic supplies of each school are input into a division model to obtain division results for dividing the m cells into the n schools.

In a second aspect, the present application provides a device for dividing a school zone, comprising:

a first acquisition unit configured to acquire a degree prediction model obtained by learning;

the construction unit is used for obtaining the feature vector of each cell by carrying out feature construction on the original data of each cell in the m cells; m is an integer greater than 0;

The prediction unit is used for inputting the characteristic vector of each cell into the degree prediction model to obtain the degree demand number of each cell;

a second acquisition unit configured to acquire a degree supply number for each of the n schools; n is an integer greater than 0;

the dividing unit is used for inputting the number of the degree demands of each cell and the number of the degree supplies of each school into the division model to obtain division results for dividing the m cells into the n schools.

In a third aspect, the present application provides an electronic device, comprising:

a processor adapted to implement computer instructions; the method comprises the steps of,

a computer readable storage medium storing computer instructions adapted to be loaded by a processor and to perform the method of the first aspect described above.

In a fourth aspect, embodiments of the present application provide a computer readable storage medium storing computer instructions that, when read and executed by a processor of a computer device, cause the computer device to perform the method of the first aspect described above.

In a fifth aspect, embodiments of the present application provide a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions, so that the computer device performs the method of the first aspect described above.

According to the embodiment of the application, the degree demand number of each district is obtained through the degree prediction model, and then the degree division result for dividing the m districts into the n schools is obtained by utilizing the degree division model based on the degree demand number of each district and the degree supply number of each school. On the other hand, the division result of the school is designed to be used for representing the division result of the m cells into the n schools, so that the problem of selecting the just-read schools can be avoided, and the balanced development of education is further ensured.

In addition, the scheme of the application can also help education departments and regional managers to plan long-term education policies and manage short-term resource investment, and further can reduce the work difficulty of the education departments and regional managers.

Drawings

Fig. 1 is an example of a system framework provided by an embodiment of the present application.

Fig. 2 is a schematic flow chart of a method for dividing a study area according to an embodiment of the present application.

FIG. 3 is another schematic flow chart of a method for distinguishing between regions according to an embodiment of the present application.

FIG. 4 is a schematic diagram of a result of the division of the study according to the embodiment of the application.

Fig. 5 is a schematic block diagram of a degree prediction model provided by an embodiment of the present application.

Fig. 6 is a schematic block diagram of a device for dividing a school district according to an embodiment of the present application.

Fig. 7 is a schematic block diagram of an electronic device provided by an embodiment of the present application.

Detailed Description

The following description of the embodiments of the present application will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.

The scheme provided by the application can relate to the technical field of artificial intelligence (Artificial Intelligence, AI).

The AI is a theory, a method, a technology and an application system which simulate, extend and extend human intelligence by using a digital computer or a machine controlled by the digital computer, sense environment, acquire knowledge and acquire an optimal result by using the knowledge. In other words, artificial intelligence is an integrated technology of computer science that attempts to understand the essence of intelligence and to produce a new intelligent machine that can react in a similar way to human intelligence. Artificial intelligence, i.e. research on design principles and implementation methods of various intelligent machines, enables the machines to have functions of sensing, reasoning and decision.

It should be appreciated that artificial intelligence techniques are a comprehensive discipline involving a wide range of fields, both hardware-level and software-level techniques. Artificial intelligence infrastructure technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation/interaction systems, mechatronics, and the like. The artificial intelligence software technology mainly comprises a computer vision technology, a voice processing technology, a natural language processing technology, machine learning/deep learning and other directions.

With research and advancement of artificial intelligence technology, research and application of artificial intelligence technology is being developed in various fields, such as common smart home, smart wearable devices, virtual assistants, smart speakers, smart marketing, unmanned, automatic driving, unmanned aerial vehicles, robots, smart medical treatment, smart customer service, etc., and it is believed that with the development of technology, artificial intelligence technology will be applied in more fields and with increasing importance value.

The embodiment of the application also relates to Machine Learning (ML) in the artificial intelligence technology, wherein ML is a multi-domain interdisciplinary, and relates to a plurality of disciplines such as probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and the like. It is specially studied how a computer simulates or implements learning behavior of a human to acquire new knowledge or skills, and reorganizes existing knowledge structures to continuously improve own performance. Machine learning is the core of artificial intelligence, a fundamental approach to letting computers have intelligence, which is applied throughout various areas of artificial intelligence. Machine learning and deep learning typically include techniques such as artificial neural networks, confidence networks, reinforcement learning, transfer learning, induction learning, teaching learning, and the like.

The embodiment of the application also relates to the technical field of cloud education of cloud technology.

Wherein, cloud education (Cloud Computing Education is abbreviated as CCEDU) refers to education platform services based on cloud computing business model application. On the cloud platform, all education institutions, training institutions, recruitment service institutions, propaganda institutions, industry associations, management institutions, industry media, legal structures and the like are integrated into a resource pool in a concentrated cloud mode, all resources are mutually displayed and interacted, the purposes are achieved according to needs, and therefore education cost is reduced, and efficiency is improved.

Fig. 1 is an example of a system framework 100 provided by an embodiment of the present application.

As shown in FIG. 1, the system framework 100 may be an application system, and embodiments of the present application are not limited to a particular type of application. The system frame 100 includes: terminal 131, terminal 132, and server cluster 110. Terminals 131 and 132 may each be connected to server cluster 110 through wireless or wired network 120.

The terminals 131 and 132 may be at least one of a smart phone, a game console, a desktop computer, a tablet computer, an electronic book reader, an MP4 player, and a laptop portable computer. Terminals 131 and 132 have applications installed and running. The application may be a map application, for example, a map application with a degree dividing function, or a degree dividing application that may call data in the map application to divide a region of a cell. The terminals 131 and 132 may be terminals used by the users 141 and 142, respectively, and user accounts are registered in applications running in the terminals 131 and 132.

The server cluster 110 includes at least one of a server, a plurality of servers, a cloud computing platform, and a virtualization center, and may also be a node cluster of a blockchain network. Server cluster 110 is used to provide background services for applications, such as applications on terminals 131 and 132. Optionally, server cluster 110 performs primary computing work and terminals 131 and 132 perform secondary computing work; alternatively, server cluster 110 performs secondary computing, and terminals 131 and 132 perform primary computing; alternatively, a distributed computing architecture is used for collaborative computing between terminals 131 and 132 and server cluster 110.

Alternatively, taking the example that the system framework 100 is a web browsing system, the server cluster 110 includes: an access server 112, a web server 111, and a data server 113. Access server 112 may be one or more, access server 112 may be deployed nearby in different cities, and access server 112 is configured to receive service requests from terminals 131 and 132 and forward the service requests to the corresponding servers for processing. The web server 111 is a server for providing a web page, in which embedded point codes are integrated, to the terminals 131 and 132; the data server 113 is for receiving data reported by the terminals 131 and 132.

Of course, terminals 131, 132 and even server cluster 110 may be used to divide the learning area of a cell.

In view of the above, the application provides a learning area dividing method, a learning area dividing device and electronic equipment, which can not only reduce the difficulty of learning in schools, but also ensure the balanced development of education.

Fig. 2 is a schematic flow chart of a method 200 for dividing a study area according to an embodiment of the present application. The method 200 may be performed by any electronic device having data processing capabilities. For example, the electronic device may be implemented as a server. The server may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing cloud services, a cloud database, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, basic cloud computing services such as big data and an artificial intelligent platform, and the server may be directly or indirectly connected through a wired or wireless communication mode. For convenience of description, a prediction method provided by the present application will be described below by taking an apparatus for identifying score of a video frame as an example.

As shown in fig. 2, the method 200 may include some or all of the following:

S210, acquiring a degree prediction model obtained through learning;

s220, carrying out feature construction on the original data of each cell in the m cells to obtain a feature vector of each cell; m is an integer greater than 0;

s230, inputting the feature vector of each cell into the degree prediction model to obtain the degree demand number of each cell;

s240, acquiring the number of degree supplies of each of the n schools; n is an integer greater than 0;

s250, inputting the number of the degree demands of each district and the number of the degree supplies of each school into a division model to obtain division results for dividing the m districts into the n schools.

In this embodiment, the number of academic demands of each cell is obtained through the academic forecast model, and then the academic division result for dividing the m cells into the n schools is obtained by using the academic division model based on the number of academic demands of each cell and the number of academic supplies of each school. On the other hand, the division result of the school is designed to be used for representing the division result of the m cells into the n schools, so that the problem of selecting the just-read schools can be avoided, and the balanced development of education is further ensured.

Illustratively, the number of academic demands may be an age population, wherein the age population may refer to: a population which has a resident residence and reaches a legal entrance age in a certain area is a population which reaches the legal entrance age of the obligation education (6-11 is a primary school age population, and 12-14 is a junior middle school age population).

For example, the number of academic demands per cell may be a population that reaches the legal entrance age for primary obligation education. For another example, the number of academic demands per cell may be a population that reaches the legal entrance age for middle school obligation education. For another example, the number of academic demands per cell may be a population that reaches the legal entrance age for primary and middle school obligations education.

FIG. 3 is another schematic flow chart of a method 300 for zoning according to an embodiment of the present application.

As shown in FIG. 3, the method 300 of discriminating a study may be performed by a plurality of functional modules. For example, the division method 300 may be performed by the degree providing module 310, the degree prediction module 320, the data analysis module 330, and the division module 340. Alternatively, the degree prediction model 320 may be an LSTM network. Alternatively, the chemical division module 340 may be a linear programming model based on dynamic programming improvement.

The academic feed module 310 may be configured to obtain the number of academic feeds of each of the n schools, and send the number of academic feeds of each of the n schools to the division model 340; the degree prediction model 320 may be used to perform feature construction on the raw data of each of the m cells to obtain a feature vector of each cell; and transmits the number of the degree demands of each cell obtained based on the feature vector of each cell to the division model 340. The data analysis module 330 may be used to generate or obtain data for use by the zoning model 340 and send it to the zoning module 340. Based on this, the division module 340 acquires the number of degree supplies for each school transmitted from the degree supply module 310, the number of degree demands for each cell transmitted from the degree prediction module 320, and the data used by the division module 340 transmitted from the data analysis module 330, and then divides the m cells into the n schools based on the received data.

Illustratively, the data generated by the data analysis module 330 may be data used by an objective function in the chemical compartmentalization model 340. Optionally, the data generated by the data analysis module 330 may also include data used by the constraints of the chemical compartmentalization model 340.

Illustratively, the data generated by the data analysis module 330 may include:

1. the location information of each school and the location information of each cell. Alternatively, the location information may be latitude and longitude information. Alternatively, the data analysis module 330 may calculate the distance of each cell with respect to each school based on the location information of each school and the location information of each cell. The distance between each cell and each school may be information such as a straight line distance, a walking distance, a driving distance, etc. between each cell and each school. The distance of each cell from each school may be used to calculate a function value of the objective function of the school partition model 340.

Of course, in other alternative embodiments, the distance of each cell relative to each school may be converted into the time of each cell relative to each school, and the time of each cell relative to each school may also be used to calculate the function value of the objective function of the school partition model 340.

2. The grade of each school. Alternatively, the data analysis module 330 may determine the grade of each school in accordance with educational resources. Optionally, the educational resources may include information such as a teacher resource, a student resource, historical learning data, and the like. Alternatively, the data analysis module 330 may determine the grade of each school based on experience or research results. Alternatively, the data analysis module 330 may determine the grade of each school with respect to each cell, for example, the data analysis module 330 may determine the grade of each school with respect to each cell based on the distance between each cell and each school, or may determine the grade of each school in other manners, which will not be described in detail herein.

After receiving the data generated by the data analysis module 330, the chemical division model 340 may also calculate a function value of the objective function based on the data generated by the data analysis module 330 and apply the function value to the constraint condition. For example, if the data generated by the data analysis module 330 includes the location information of each school and the location information of each cell, or if the data generated by the data analysis module 330 includes the distance between each cell and each school, the objective function used by the learning division model 340 may be: a linear function is constructed based on the distance of each cell relative to each school. Further, if the data generated by the data analysis module 330 includes the level of each school, the constraints used by the learning compartment model 340 may include: based on the constraints that can be built based on the rank of each school.

Of course, in other alternative embodiments, the data generated by the data analysis module 330 may also include other data and analysis results of the other data. For example, the data generated by the data analysis module 330 may also include educational resource configuration information for each school, including but not limited to, class number, teacher to student ratio, resident convenience, etc.

In addition, the present application is not particularly limited to the constraint conditions used for the chemical division model 340. Constraints used by the compartmentalization model 340 may be constructed, for example, based on problems in compartmentalization. For example, problems with the division of school zones include, but are not limited to: imbalance of the duty ratio of external caretakers and children, imbalance of the spatial distribution of schools, imbalance of high-quality educational resources and demands, and the like.

As shown in fig. 4, based on the scheme of the application, the dynamic planning and global optimization method is adopted to divide the school zone of each cell, so that not only can the student resources of the school be balanced, but also the entrance difficulty of the school can be reduced, unbalance of the supply of the school can be avoided, and the balanced development of education can be ensured. In addition, as the obtained division result of the school can divide one district into one school, the problem of selecting just reading the school can be avoided, and the balanced development of education is further ensured.

Of course, if each school is sponsored with primary and middle school classes, the scheme of the application can be adopted to divide primary and middle school areas of each cell simultaneously. If only primary school classes or middle school classes are issued by a part of schools, primary school areas and middle school areas of each cell can be respectively divided by adopting the scheme of the application, and the application is not specifically described.

In some embodiments, the S250 may include:

determining whether a j-th school of the n schools is an alternative school of an i-th cell of the m cells; determining at least one school zone division mode of the ith cell based on whether the jth school is an alternative school to the ith cell; calculating at least one loss value of the ith cell in the at least one learning division mode by using an objective function in the learning prediction model; the at least one loss value is used for evaluating the loss of the cells with divided ranks in the m cells when the ith cell is divided according to the at least one division mode; and determining an alternative school used by the learning division mode with the minimum loss value in the at least one loss value as the school divided for the ith cell.

The dynamic planning of the dynamic planning is a branch of operation planning and is a process of optimizing a solving decision process, in this embodiment, based on at least one loss value of the ith cell in the at least one learning region division mode, the learning region division mode with the minimum loss is solved by using the thought of the dynamic planning, so that not only the learning region division of the ith cell can be realized, but also the learning region division difficulty can be reduced, and the division efficiency can be improved.

In some embodiments, the parameter x is determined _ij The method comprises the steps of carrying out a first treatment on the surface of the If the jth school is the candidate school of the ith cell, the x _ij The value of (2) is a first numerical value; if the jth school is not the candidate school for the ith cell, the x _ij The value of (2) is a second value; the first value is greater than the second value; based on the x _ij And the distance d to school of the ith cell relative to the jth school _ij A loss value of a current division pattern of the at least one degree division pattern is determined.

That is, it can be based on the x _ij And d is the _ij Is determined as an objective function of the degree-partitioning model. For example, it can be based on this x _ij And d is the _ij A functional expression of the product of (2) is determined as an objective function of the degree-partitioning model.

Illustratively, the first value is 1 and the second value is 0.

Illustratively, the parameter x _ij May be a 0-1 variable, defined as follows:

where i e {0,1,2, …, m-1} represents the number of m cells, j e {0,1,2, …, n-1} represents the number of n schools.

In some embodiments, based on the x _ij D is at least one of _ij And determining a loss value for the current learn zone division mode from: the distance to school of the first i-1 cells in the n cells, the average distance to school of the first i-1 cells, the relaxation factor of the school in which the first i-1 cells are located, and the relaxation factor of the j-th school; wherein the relaxation factor may be used to represent the number of students that the school needs to accommodate elastically.

Illustratively, the relaxation factor of each school is determined from at least one of the following information for each school: the difference between the number of classes set at the time of each school's start and the number of classes existing in each school, the number of remaining classrooms in each school, and the number of functional classrooms that can be modified into a class-taking classroom. Illustratively, the difference between the number of classes set at the time of each school' S operation and the number of classes existing in each school may be denoted as S2, the remaining number of classrooms in each school may be denoted as S4, and the number of functional classrooms in each school which may be modified into a class room may be denoted as S5. In one implementation, s2+s4+s5 is inversely proportional to the value of the relaxation factor. In another implementation, s2+s6+s5 is inversely proportional to the value of the relaxation factor; and S6, determining according to the ratio of the number of the remaining class numbers to the number of the non-class numbers of each school. For example, S6 is a ratio of the number of remaining classes to the number of non-class classes in each school or an integer obtained by rounding up or down the ratio.

Illustratively, when the each school is a new school or a school with a remaining classroom, the relaxation factor of the each school is determined according to the following information of the each school: the difference between the number of classes set at the time of each school's start and the number of classes existing in each school, the number of remaining classrooms in each school, and the number of functional classrooms that can be modified into a class-taking classroom. When the schools are old schools or schools without residual classrooms, the relaxation factor of each school is determined according to the following information of each school: the difference between the number of classes set at the time of each school's start and the number of classes existing in each school and the number of functional classrooms that can be modified into class classrooms.

For example, for a newly-opened school or a school that is not in work to six grades (such as a school just in two years, then only students of one grade and two grades), there may be two indicators, one being the number of remaining classrooms S3, and one being the ratio of the number of remaining classrooms to the number of non-opened grades, denoted S6, such as a total of 30 classes, only 5 classes of one grade, 4 classes of two grades, then s3=30-5-4=21 classes, s6= (30-4-5)/(6-2), rounded to 5 classes.

In this embodiment, by introducing the relaxation factor of the school where the previous i-1 cell is located and the relaxation factor of the jth school, the degree supply of the school where the previous i-1 cell is located and the jth cell can be dynamically adjusted, so that the rationality and the balance of the degree supply of the school where the previous i-1 cell is located and the jth cell are facilitated, that is, the difficulty of entrance can be reduced, and the balanced development of education can be ensured.

In addition, as the relaxation factors of the schools where the previous i-1 cells are located and the relaxation factors of the jth schools can reflect the dynamic change of the education resources, the change of the education resources can be considered in real time in the dividing process of the education areas, and the method is further beneficial to the guiding proposal provided for long-term education planning and short-term investment schemes in practical application.

In some embodiments, based on the x _ij D is at least one of _ij And the distance to the first i-1 cells, determining a first parameter; the first parameter is used for evaluating the total correction distance of the cells with divided degree in the m cells when the ith cell performs division of the division according to the current division mode of the division; based on the x _ij D is at least one of _ij Determining a second parameter from the calibrated distances of the first i-1 cells and the average distance; the second parameter is used for evaluating the ith cell to learn according to the current learning partition mode the difference between the distance to school of the cells of the divided degree in dividing the m cells; determining a third parameter based on the relaxation factor of the school in which the first i-1 cell is located and the relaxation factor of the j-th school; the third parameter is used to evaluate the ith smallThe total number of students which need to be elastically accommodated by the school to which the cell of the divided degree in the m cells is divided when the region is divided according to the current division mode of the region; and determining a loss value of the current division mode based on the first parameter, the second parameter and the third parameter.

In this embodiment, by introducing the objective function designed based on the first parameter, the second parameter and the third parameter, the learning region division model is an improved linear programming model, so that not only can the learning region division of the ith cell be realized on the basis of the nearby entrance, but also the education balance of the divided learning region and the school where the learning region is located can be ensured.

Specifically, on one hand, by calculating the total to school distance of the cells with divided degree among the m cells, and optimizing it to the shortest (i.e., near-entrance); on the other hand, by calculating the variance of the distances of the differences between the distances of the cells divided into the degree among the m cells, and optimizing it to the minimum; on the other hand, by calculating the total number of students to be elastically accommodated in the school to which the divided-degree cell is divided in the m cells and optimizing the total number to be minimized, namely, the fewer the number of students to be elastically accommodated in the school is, the better the number of students to be elastically accommodated in the school is, not only can the division of the learning area of the ith cell be realized on the basis of the nearby entrance, but also the educational balance of the cells of the divided learning area, namely, the balance of the entrance distance, can be ensured, and the educational balance of the school in which the cells of the divided learning area are positioned, namely, the balance of the number of students to be elastically accommodated in the school can be ensured.

Illustratively, the objective function of the compartmentalization model may be designed as a function of:

wherein the first part indicates that the total distance to school of the cells with the divided degree is the shortest (i.e. near entrance), the second part indicates that the variance of the distance to school of the cells with the divided degree is the smallest, and the third part indicates that the total number of students to be elastically accommodated by the schools to which the cells with the divided degree are divided is the smallest, i.e. the fewer students to be elastically accommodated by the schools are the better.

Specifically, x _ij Indicating whether the jth school is an alternative school to the ith cell; for example, if the jth school is an alternative school to the ith cell, the x _ij The value of (2) is 1; if the jth school is not the candidate school for the ith cell, the x _ij The value of (2) is 0; d, d _ij The distance from the ith cell to the jth school is represented (may be a straight line distance, a walking distance, a driving distance, or the like), μ represents the average of the distances from all the cells to the schools, that is, the average distance, δ represents the relaxation factor of the schools to which the cells of the m cells are divided (that is, the number of students to which the schools of the m cells are divided may be elastically accommodated), and the schools to which the cells of the m cells are divided may be the sum of the relaxation factor of the schools in which the previous i-1 cells are located and the relaxation factor of the jth school. C represents the coefficient of δ.

In this embodiment, the improved linear programming model in the learning region division provides a more scientific learning region planning scheme by applying the dynamic programming and global optimization method. Specifically, the following aspects are embodied:

1. in the process of obtaining the division result of the school, on one hand, the division model of the school is beneficial to analyzing the supply and demand relationship of the school through introducing the demand number of the school of each district and the supply number of the school of each school, so as to obtain the division result of the school meeting the balance of the supply and demand, and further, not only the student resources of the school can be balanced, but also the entrance difficulty of the school can be reduced, unbalance of the supply of the school can be avoided, and further, the balanced development of the education is ensured. On the other hand, the division result of the school is designed to be used for representing the division result of the m cells into the n schools, so that the problem of selecting the just-read schools can be avoided, and the balanced development of education is further ensured.

2. By introducing the relaxation factor of the school in which the previous i-1 cell is located and the relaxation factor of the jth school, the school level supply quantity of the school in which the previous i-1 cell is located and the jth cell can be dynamically adjusted, and further, the rationality and the balance of the school level supply quantity of the previous i-1 cell and the jth cell are facilitated, namely, the entrance difficulty can be reduced, and the balanced development of education can be ensured.

3. A greater number of optimization objectives are employed.

In conclusion, the scheme provided by the application can divide the school zone more accurately, so that not only can student resources of the school be balanced, but also the entrance difficulty of the school is reduced, unbalance of the supply of the school can be avoided, and the balanced development of education is ensured.

In some embodiments, it is determined whether the jth school is an alternative school to the ith cell based on constraints of the objective function.

Illustratively, this x can be based on _ij The determined constraint is determined as a constraint of the objective function.

Illustratively, based on the x _ij The determined constraints include, but are not limited to: based on the x _ij Constraint of convenience of each cell determined and based on x _ij And determining the equilibrium constraint condition of each school. For example, based on the x _ij Constraints on the determined convenience of each cell include, but are not limited to: based on the x _ij The constraint condition of the time of employment and/or distance of employment of each cell is determined based on the x _ij The determined equilibrium constraints for each school include, but are not limited to: based on the x _ij The constraint condition of the employment coverage rate of each school is determined and is based on the x _ij Determining constraint conditions of the employment rate of each school based on the x _ij Constraints on the facility usage of each school are determined.

In other words, the x can be based on the multiple angles such as the time of employment, the coverage rate of employment, the utilization rate of school facilities, the convenience of residents, the fairness of school layout and the balance of education development _ij Constraint conditions of the objective function are determined, and then the supply-demand relationship of student resources and the supply-demand relationship of education resources are further defined, so that more accurate division of learning areas is realized, each student is guaranteed to be capable of receiving good education, and education balance is guaranteed.

In some embodiments, the first constraint of the objective function is: the number of schools into which the i-th cell is divided is less than or equal to a preset threshold.

The preset threshold is, for example, 1, that is, the i-th cell can only be divided into one school. For example, the first constraint may be implemented as the following relationship:

where i ε {0,1,2, …, m-1}.

Wherein x is _ij Indicating whether the jth school is an alternative school to the ith cell; for example, if the jth school is an alternative school to the ith cell, the x _ij The value of (2) is 1; if the jth school is not the candidate school for the ith cell, the x _ij The value of (2) is 0.

In this embodiment, when the first constraint condition is introduced and the preset threshold is further set to 1, this is equivalent to enabling the school age population in the same district to read in one school as much as possible, avoiding dividing the same district into a plurality of schools, that is, avoiding the problem of selecting the school to read in place, and further ensuring the balanced development of education.

In some embodiments, the second constraint of the objective function is: the sum of the number of the degree demands of all the cells divided into the jth school and the number of the degree demands of the ith cell is less than or equal to the sum of the number of the degree supplies of the jth school and the relaxation factor threshold of the jth school.

Illustratively, the second constraint may be implemented as the following relationship:

where j ε {0,1,2, …, n-1}.

Wherein x is _ij Indicating whether the jth school is an alternative school to the ith cell; for example, if the jth school is an alternative school to the ith cell, the x _ij The value of (2) is 1; if the jth school is not the candidate school for the ith cell, the x _ij The value of (2) is 0.s is(s) _i Representing the number of school age population of the ith cell, namely the number of school level demands of the ith cell, D _j The number of degrees that the jth school can provide, i.e., the number of degrees offered by the jth school, is represented. Delta _j The relaxation factor of the jth school (i.e., the number of students that the jth school can flexibly accommodate) is represented.

In this embodiment, by introducing the second constraint condition, the supply-demand relationship between the number of degree demands of the cell divided into the jth school and the number of degree supplies of the jth school can be constrained, and by introducing the relaxation factor of the jth school, the number of degree supplies of the jth school can be made to have a certain elasticity, so that not only the possibility of the jth school as an alternative school of the ith cell can be increased, but also the difficulty of entering the study can be reduced.

In some embodiments, the third constraint of the objective function is: the sum of the number of the degree demands of all the cells divided into the jth school and the number of the degree demands of the ith cell is smaller than or equal to the threshold value of the degree supply number of the jth school and the degree rate of the jth school; wherein the employment threshold is determined based on the school equality of the j-th school.

Illustratively, this third constraint may be implemented as the following relationship:

where K ε {0,1,2, …, K-1}.

Wherein x is _ij Indicating whether the jth school is an alternative school to the ith cell; for example, if the jth school is an alternative school to the ith cell, the x _ij The value of (2) is 1; if the jth school is not the candidate school for the ith cell, the x _ij The value of (2) is 0.s is(s) _i Representing the number of school age population of the ith cell, namely the number of school level demands of the ith cell, D _j The number of degrees that the jth school can provide, i.e., the number of degrees offered by the jth school, is represented. r is (r) _k The score of the number of degree supplies of the school with the rank K is represented, and K represents the classification of the school into K ranks.

In this embodiment, the third constraint condition is introduced, so that the ten-wire can constrain the employment rates of schools with different grades, and further, the balance of educational resources can be ensured.

Of course, in other alternative embodiments, constraints may be made from other angles based on the ith cell or the jth school. For example, it may be constrained from a minimum learning time or the like to determine whether the jth school may be considered as an alternative school for the ith cell or the like, which will not be specifically described in the embodiments of the present application.

In some embodiments, the S240 may include:

determining the number of class supplies for each school based on: the six-grade existing class number of each school, the difference between the class number set when each school is started and the existing class number of each school, the number of remaining classrooms of each school, and the number of functional classrooms convertible into class classrooms; the number of degree supplies for each school is determined based on the number of class supplies for each school.

By way of example, the number of class supplies that can be provided in the next school year of each school may be calculated by analyzing the spatial layout of each school, including the number of students, the number of staff members, the number of class stages of the school, the number of functional classrooms (convertible into classrooms) of the school, etc., in each level of each school, and further determining the number of degree supplies in the next school year of each school based on the number of class supplies that can be provided in the next school year of each school. For example, the product of the number of class supplies available for the next school year of each school and the number of students that each class can accommodate may be determined as the number of level supplies available for the next school year of each school.

Illustratively, the six-grade existing class number of each school may be denoted as S1, the difference between the class number set at the time of starting each school and the existing class number of each school may be denoted as S2, the remaining class number of each school may be denoted as S4, and the number of functional classrooms that each school may be modified as a class may be denoted as S5. Based on this, s1+s2+s3+s4+s5 may be taken as the number of class supplies available for the next school year for each school; of course, s1+s2+s3+s6+s5 may be the number of class supplies available in the next school year for each school; and S6, determining according to the ratio of the number of the remaining class numbers to the number of the non-class numbers of each school. For example, S6 is a ratio of the number of remaining classes to the number of non-class classes in each school or an integer obtained by rounding up or down the ratio.

Through the statistics, the number of class supplies available in the next school year of each school can be obtained, and the number of people (such as 45) in each class can be multiplied to obtain the number of degree supplies of the school, and the number of degree supplies can be used as input of a division model of the school.

In some embodiments, the degree prediction model is a Long Short-Term Memory (LSTM) network, and the raw data of each cell includes at least one of the following information at the current time: population flow data, birth rate, population following migration, entrance rate; each dimension of the feature vector of each cell is used to represent an item of information in the raw data of each cell.

The degree prediction model may also be, for example, a Bi-directional long and short Term Memory network (Bi-directional Long Short-Term Memory, bi-LSTM).

It should be noted that Long Short-Term Memory (LSTM) is suitable for modeling time series data. For example, the word representations are combined into a sentence representation, and an addition method may be employed, i.e., a method of adding all word representations, or an averaging method, etc., but these methods do not take into consideration the order of words in the sentence. Such as the sentence "i do not feel good". The "no" word is negative of the following "good", i.e. the emotion polarity of the sentence is devaluation. The LSTM model can better capture the dependency relationship of a longer distance. Because LSTM can learn which information to memorize and which information to forget through the training process. Modeling sentences with LSTM has a problem: the back-to-front information cannot be encoded. In more fine-grained classification, five classification tasks such as recognition for a strong degree, recognition for a weak degree, neutrality, detraction for a weak degree, detraction for a strong degree require attention to interactions between emotion words, degree words, negatives. However, BI-LSTM may well overcome the above-described problems. As an example, "this restaurant is dirty and not good, where" no-go "is a modification to the degree of" dirty ", BI-LSTM networks can better capture BI-directional semantic dependencies.

It should be noted that LSTM networks and BI-LSTM networks are merely examples of the present application and should not be construed as limiting the application. In other words, the model for feature extraction according to the present application is not limited to using LSTM network or BI-LSTM network, and in other alternative embodiments, other depth models may be used as the level prediction model, for example, a convolutional neural network (Convolutional Neural Network) or a gate-controlled loop unit (Gated Recurrent Unit, GRU) may be used as the level prediction model, so as to construct a level prediction model capable of identifying data in feature vectors and obtaining the level demand number.

Illustratively, the degree prediction model includes, but is not limited to: a traditional learning model, an integrated learning model, or a deep learning model. Alternatively, conventional learning models include, but are not limited to: a tree model (regression tree) or a logistic regression (logistic regression, LR) model; for example, the regression tree model may be an autoregressive moving average model (autoregressive moving average model, ARMA); for another example, the logistic regression model may be, for example, a linear model, an exponentiation curve model, and a compound curve model; the ensemble learning model includes, but is not limited to: an improved model of gradient lifting algorithm (XGBoost) or a random forest model; deep learning models include, but are not limited to: long Short-Term Memory (LSTM) or neural networks. Of course, in other embodiments of the present application, other machine learning type models may be used, as the application is not limited in detail.

It should be noted that the conventional learning model may have the following technical problems:

(1) For the regression tree model, when the outside is greatly changed, larger deviation is often brought, the middle-short term effect is good, and the current data cannot be considered only by using the historical data in the degree prediction. (2) For the logistic regression model, the data to be input satisfies the assumption of the logistic growth process, and the results obtained by selecting different curve functions are different, so it is difficult to select an appropriate curve function.

In this embodiment, the prediction of the degree demand number of each cell is completed by introducing a deep learning model, and the raw data for predicting the degree demand number may be panel data, which includes not only historical information (i.e., time series data including but not limited to local entrance policy, population flow data, historical entrance data, etc. data of multiple dimensions) but also other influencing factors of the current time period (such as purchase data of the study rooms of the cells near the school, etc.), and the recurrent neural network and its variant (LSTM) may well process such panel data highly related to the time series and insensitive to the missing values, so that the prediction accuracy can be improved.

Of course, in other alternative embodiments, the degree-prediction model may also be a cyclic neural network (Recurrent Neural Network, RNN) or other depth model.

It should be noted that, although RNNs are also good at processing time-series data and can learn previous information and predict current results, there is a long-term dependence problem that information with too long intervals cannot be remembered, and thus LSTM models, i.e., long-term memory models, are derived, mainly for solving the problems of gradient extinction and gradient explosion in the long-sequence training process. LSTM mainly adds a number of gate structures to erase or add information learning capabilities of the model, such as forget gates to determine which information is lost at this stage, input gates to determine which information is updated, etc.

Fig. 5 is a schematic block diagram of a degree prediction model 400 provided by an embodiment of the present application.

As shown in FIG. 5, the degree prediction model 400 may include a data acquisition and cleansing module 410, a training module 420, and a testing module 430. The data acquisition and cleansing module 410 may include a data acquisition module 411, a data screening module 412, a data cleansing module 413, and a data preprocessing module 414.

The data acquisition and cleansing module 410 may be configured to perform feature construction on the raw data of each of the m cells to obtain a feature vector of each cell. For example, the data acquisition and cleansing module 410 may be configured to perform feature construction on the raw data of each of the m cells at the historical time, so as to obtain a feature vector of each cell, and further obtain a training data set and a test data set of the LSTM. For another example, the data acquisition and cleansing module 410 may be configured to perform feature construction on the original data of each of the m cells at the current time, so as to obtain a feature vector of each cell, and further be configured to predict a feature vector of the number of level requirements of each cell. Illustratively, the raw data for each cell includes at least one of the following information: population flow data, birth rate, population following migration, entrance rate; the degree prediction model can be used for carrying out feature construction on the original data of each cell in m cells to obtain a feature vector of each cell; wherein each dimension of the feature vector of each cell may be used to represent an item of information in the raw data of each cell.

The data acquisition module 411 mainly acquires raw data of each cell. For example, the raw data for each cell includes at least one of the following information: population flow data, birth rate, population following migration, entrance rate, etc. For another example, the raw data of each cell is panel data containing time series information, which may include not only historical information, i.e. time series data, but also other influencing factors of the current time phase, such as purchase data of a study room near each school, etc.

The data filtering module 412 may select complete and valid data and features from the data collected by the data collection module 411 according to the missing rate and the history experience, and then output the selected data or features to the data cleaning module 413.

The data cleaning module 413 may be configured to perform operations such as repeated data removal, outlier processing, and missing value processing on the received data or feature; for example, the data cleansing module 413 may discard features with too large feature values according to the feature distribution, for example discard the outlier of the previous 1/m, m may be set to 10000, specifically according to the application scenario, and for missing features, for example, may fill continuous features with a mean value, and fill discrete features with constants as data of separate categories. Of course, the data cleaning module 413 may also learn to derive derived features, e.g., feature combinations or derivations may be made by feature transformation, feature squaring, feature addition and subtraction. The features in the original data may also be discretized or encoded, for example, continuous features may be box-wise discretized, and discrete features may be one-hot encoded. Of course, in other alternative embodiments, the data cleansing module 413 may discard data with too many missing features. As an example, the data cleansing module 413 may discard features whose number of feature deletions is greater than or equal to a deletion threshold, e.g., the deletion threshold may be equal to the product of the sample data amount and n, which may be a value greater than 0 and less than 1, e.g., n may be set to 0.4. In a specific implementation, the value of n is set according to the application scenario, and if the number of missing features of a certain data exceeds the missing threshold, the data cleaning module 413 may filter the data, i.e. delete or discard the features in the data.

The data preprocessing module 414 may be configured to perform missing value interpolation and data normalization on the cleaned feature to obtain a feature vector of each cell. Further, when the feature obtained by the data preprocessing module 414 is data of each cell at a historical time, the data preprocessing module 414 may further divide the obtained feature vector of each cell into a training data set and a test data set. For example, the data preprocessing module 414 may divide the resulting feature vectors for each cell into a training data set and a test data set at an 8:2 or other ratio.

Training module 420 may be used to train an LSTM network. For example, the training module 420 may train the LSTM network based on the training data set output by the data preprocessing module 414 in the data acquisition and cleansing module 410. The specific training process can be realized as follows: the feature vectors (e.g., X ₁ Feature vector at time X _t Feature vector at time point), the feature vector (e.g., X) of each cell at the history time point ₁ Feature vector at time X _t Feature vectors at time instants) are input into the LSTM network, which automatically updates the optimized model parameters a continuously in the back propagation, and in continuous training, when the error rate is smaller than a given threshold, the final LSTM network is obtained, i.e. the LSTM network determining the model parameters a (note that the parameters a at each time instant are shared here). Compared with the traditional statistical model, the statistical model requires consistent data distribution and is difficult to process the panel data with both time and section, and the LSTM network can effectively overcome the defects.

The test module 430 may be used to test the LSTM network. For example, the test module 430 may test the LSTM network based on the training module 420 based on the test data set output by the data preprocessing module 414 in the data acquisition and cleansing module 410. The specific test process can be realized as follows: test module 430 based on the LSTM network obtained by training module 420, test module 430 determines a feature vector (e.g., C _t Feature vector at time) is input into the LSTM network to obtain the number of degree demands (e.g. C) at historical time _t The number of degree demands at that moment) is denoted as D. Further, the LSTM network may obtain the number of degree demands at the historical time (e.g. C _t Number of degree demands at a time point) is input to the error calculation module 431, so that the error calculation module 431 calculates the number of degree demands at a history time point (e.g., C _t Number of degree demands at a time point) if the number of degree demands at a history time point (e.g., C _t The number of degree demands at a moment) is too large, the LSTM network may be continuously trained by the training module 420 to obtain an LSTM network capable of accurately predicting the number of degree demands.

After the LSTM network is trained, the number of the degree demands of each cell may be predicted by using the trained LSTM network, specifically, the data acquisition and cleaning module 410 may perform feature construction on the original data of each of the m cells at the current time, so as to obtain a feature vector of each cell, and then, the feature vector of each cell is input to the trained LSTM network, so as to obtain the number of the degree demands of each cell.

It should be understood that fig. 5 is only an example of the present application and should not be construed as limiting the present application. For example, the construction manner of the present application for performing the feature construction on the original data of each cell is not particularly limited. For example, in other alternative embodiments, the original image data may be converted to the image feature by feature engineering, as the application is not limited in detail.

The following is an exemplary description of the relevant terms involved in feature engineering.

The characteristics are as follows: the information extracted from the data and useful for predicting the result may be text or data. Characteristic engineering: a process of using knowledge about the data domain to create features that enable machine learning algorithms to achieve optimal performance.

Feature engineering (Feature Engineering) is the process of converting raw data into features that better express the nature of the problem, so that the application of these features to a predictive model can improve model prediction accuracy for invisible data. In short, the features that have a significant impact on the dependent variable y are called independent variables x, which are the features that the purpose of feature engineering is to discover. Because the good characteristics have stronger flexibility, the training can be performed by using a simple model, and excellent results can be obtained. "engineering should be done to improve his/her mind, and" engineering of a feature "can be understood as the process of" benefiting his/her mind ". The purpose of feature engineering is to screen out better features and obtain better training data. Feature engineering is a topic that does not seem worth discussing. But plays a critical role in the success or failure of machine learning. Machine learning algorithms have been successful in many cases by creating engineered features that a learner can understand. In one implementation, feature engineering may refer to feature extraction and analysis of raw data through complex feature engineering, and then training the extracted and analyzed features with a conventional autoregressive integral moving average model (Autoregressive Integrated Moving Average Model, ARIMA) or an isolated forest algorithm, where the training-based model is used to predict LTV.

The feature engineering may include feature extraction, feature construction, feature selection, etc. Feature extraction: the raw data is converted into a set of features with obvious physical or statistical significance or kernel. And (3) feature construction: is a new feature constructed manually in the original data. Feature selection: and selecting a group of feature subsets with the most statistical significance from the feature set, thereby achieving the effect of dimension reduction.

Of course, the embodiment of the present application is not limited to the specific form of the original data. As an example, the raw data may be constructed as tabular data to obtain a feature matrix, based on which feature extraction may be performed using principal component analysis (Principal Component Analysis, PCA) to create new features for characterizing the LTV.

For feature extraction, the object of feature extraction is raw data (raw data), i.e., raw features, which are used to construct new features, i.e., to convert raw data into a set of features that have a distinct physical or statistical meaning. The original data may include static data and dynamic data, wherein the static data may be data related to attribute information of a cell, such as an area of the cell or a number of floors and a layer height of the cell, and the dynamic characteristics may be dynamic data of an entrance policy, birth rate, and the like.

For feature construction, feature construction refers to the artificial construction of new features from raw data. In particular, the potential forms and data structures of the new features can be determined from the real data samples in terms of the predicted number of academic demands to enable better application in the predictive model. The feature construction requires strong insight and analysis capability, requiring that we be able to find some features of physical significance from the raw data. Assuming that the original data is tabular data, the new features may be created using either mixed or combined attributes, or the original features may be decomposed or sliced to create new features.

For feature selection, the ranking may be based on feature importance, and then feature selection may be based on feature ranking results. Assuming that there is a standard table data, each row (i.e., category item) of the table data represents an observation sample data, each column (i.e., feature item) in the table data is a feature, among the features, some features carry abundant information and some features carry little information, and the features carrying little information belong to irrelevant data (irrelvant data), the feature importance can be measured by the correlation (feature importance) between the feature item and the category item, and then feature selection is performed based on the feature ranking result.

It should be noted that, the embodiment of the present application does not limit the specific flow related to the feature engineering. By way of example, the flows involved in feature engineering include, but are not limited to: binning, one-Hot Encoding, feature Hashing, nesting, taking logarithms (Log Transformation), feature Scaling, normalization, or feature interaction (Feature Interaction). Of course, other procedures may be included, and embodiments of the present application are not limited thereto.

It should be further noted that, the time series data related to the present application is intended to be used to characterize data on a time axis, where the data on the time axis may include data at a plurality of historical moments, and a type of the data at each historical moment may be the same as or different from a type of the data at the current moment. For example, the method provided by the application can be applied to various school district division scenes, such as school district division scenes of various grades.

The preferred embodiments of the present application have been described in detail above with reference to the accompanying drawings, but the present application is not limited to the specific details of the above embodiments, and various simple modifications can be made to the technical solution of the present application within the scope of the technical concept of the present application, and all the simple modifications belong to the protection scope of the present application. For example, the specific features described in the above embodiments may be combined in any suitable manner, and in order to avoid unnecessary repetition, various possible combinations are not described further. As another example, any combination of the various embodiments of the present application may be made without departing from the spirit of the present application, which should also be regarded as the disclosure of the present application.

It should be further understood that, in the various method embodiments of the present application, the sequence numbers of the foregoing processes do not mean the order of execution, and the order of execution of the processes should be determined by the functions and internal logic of the processes, and should not constitute any limitation on the implementation process of the embodiments of the present application.

The method provided by the embodiment of the application is described above, and the device provided by the embodiment of the application is described below.

Fig. 6 is a schematic block diagram of a device 500 for dividing a school district according to an embodiment of the present application.

As shown in fig. 6, the school zone partition means 500 may include:

a first obtaining unit 510 for obtaining a degree prediction model obtained by learning;

a construction unit 520, configured to obtain a feature vector of each of the m cells by performing feature construction on the original data of the each cell; m is an integer greater than 0;

a prediction unit 530, configured to input the feature vector of each cell to the level prediction model, to obtain a level requirement number of each cell;

a second acquisition unit 540 for acquiring the number of degree supplies of each of the n schools; n is an integer greater than 0;

a dividing unit 550 for inputting the number of degree demands of each cell and the number of degree supplies of each school into a division model, and obtaining division results for representing division of the m cells into the n schools.

In some embodiments, the dividing unit 550 is specifically configured to:

determining whether a j-th school of the n schools is an alternative school of an i-th cell of the m cells;

determining at least one school zone division mode of the ith cell based on whether the jth school is an alternative school to the ith cell;

calculating at least one loss value of the ith cell in the at least one learning division mode by using an objective function in the learning prediction model; the at least one loss value is used for evaluating the loss of the cells with divided ranks in the m cells when the ith cell is divided according to the at least one division mode;

and determining an alternative school used by the learning division mode with the minimum loss value in the at least one loss value as the school divided for the ith cell.

In some embodiments, the dividing unit 550 is specifically configured to:

determining parameter x _ij The method comprises the steps of carrying out a first treatment on the surface of the If the jth school is the candidate school of the ith cell, the x _ij The value of (2) is a first numerical value; if the jth school is not the candidate school for the ith cell, the x _ij The value of (2) is a second value; the first value is greater than the second value;

Based on the x _ij And the distance d to school of the ith cell relative to the jth school _ij A loss value of a current division pattern of the at least one degree division pattern is determined.

In some embodiments, the dividing unit 550 is specifically configured to:

based on the x _ij D is at least one of _ij And determining a loss value for the current learn zone division mode from:

the distance to school of the first i-1 cells in the n cells, the average distance to school of the first i-1 cells, the relaxation factor of the school in which the first i-1 cells are located, and the relaxation factor of the j-th school; wherein the relaxation factor is used to represent the number of students that the school needs to accommodate elastically.

In some embodiments, the dividing unit 550 is specifically configured to:

based on the followingx _ij D is at least one of _ij And the distance to the first i-1 cells, determining a first parameter; the first parameter is used for evaluating the ith cell to perform according to the current learning zone division mode the total distance of the cells with divided disciplines in the m cells is counted when dividing the cells;

based on the x _ij D is at least one of _ij Determining a second parameter from the calibrated distances of the first i-1 cells and the average distance; the second parameter is used for evaluating the difference between the distance to the school of the cells with divided degree in the m cells when the ith cell performs the division of the division according to the current division mode of the division;

Determining a third parameter based on the relaxation factor of the school in which the first i-1 cell is located and the relaxation factor of the j-th school; the third parameter is used for evaluating the ith cell when the ith cell performs the learning partition according to the current learning partition mode the total number of students to be elastically accommodated by the school to which the cell of which the degree is divided among the m cells;

and determining a loss value of the current division mode based on the first parameter, the second parameter and the third parameter.

In some embodiments, the dividing unit 550 is specifically configured to:

based on the constraint of the objective function, it is determined whether the jth school is an alternative school to the ith cell.

In some embodiments, the second obtaining unit 540 is specifically configured to:

determining the number of class supplies for each school based on:

the six-grade existing class number of each school, the difference between the class number set when each school is started and the existing class number of each school, the number of remaining classrooms of each school, and the number of functional classrooms convertible into class classrooms;

the number of degree supplies for each school is determined based on the number of class supplies for each school.

In some embodiments, the degree prediction model is a long-short-term memory LSTM network, and the raw data of each cell includes at least one of the following information at the current time: population flow data, birth rate, population following migration, entrance rate; each dimension of the feature vector of each cell is used to represent an item of information in the raw data of each cell.

It should be understood that apparatus embodiments and method embodiments may correspond with each other and that similar descriptions may refer to the method embodiments. To avoid repetition, no further description is provided here. Specifically, the learning region dividing apparatus 500 may correspond to a corresponding main body in the method 200 for executing the embodiment of the present application, and each unit in the learning region dividing apparatus 500 is for implementing a corresponding flow in the method 200, and for brevity, will not be described herein.

It should also be understood that each unit in the apparatus 500 according to the embodiments of the present application may be separately or all combined into one or several additional units, or some unit(s) thereof may be further split into a plurality of units with smaller functions, which may achieve the same operation without affecting the implementation of the technical effects of the embodiments of the present application. The above units are divided based on logic functions, and in practical applications, the functions of one unit may be implemented by a plurality of units, or the functions of a plurality of units may be implemented by one unit. In other embodiments of the present application, the apparatus 500 may also include other units, and in practical applications, these functions may also be implemented with assistance from other units, and may be implemented by cooperation of a plurality of units. According to another embodiment of the present application, the chemical division 400 according to the embodiment of the present application may be constructed by running a computer program (including a program code) capable of executing the steps involved in the respective methods on a general-purpose computing device of a general-purpose computer including a processing element such as a Central Processing Unit (CPU), a random access storage medium (RAM), a read only storage medium (ROM), and the like, and the chemical division method of the embodiment of the present application is implemented. The computer program may be recorded on a computer readable storage medium, and loaded into an electronic device and executed therein to implement a corresponding method of an embodiment of the present application.

In other words, the units referred to above may be implemented in hardware, or may be implemented by instructions in software, or may be implemented in a combination of hardware and software. Specifically, each step of the method embodiment in the embodiment of the present application may be implemented by an integrated logic circuit of hardware in a processor and/or an instruction in software form, and the steps of the method disclosed in connection with the embodiment of the present application may be directly implemented as a hardware decoding processor or implemented by a combination of hardware and software in the decoding processor. Alternatively, the software may reside in a well-established storage medium in the art such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, and the like. The storage medium is located in a memory, and the processor reads information in the memory, and in combination with hardware, performs the steps in the above method embodiments.

Fig. 7 is a schematic structural diagram of an electronic device 600 provided in an embodiment of the present application.

As shown in fig. 7, the electronic device 600 includes at least a processor 610 and a computer-readable storage medium 620. Wherein the processor 610 and the computer-readable storage medium 620 may be connected by a bus or other means. The computer readable storage medium 620 is used to store a computer program 621, the computer program 621 including computer instructions, and the processor 610 is used to execute the computer instructions stored by the computer readable storage medium 620. Processor 610 is a computing core and a control core of electronic device 600 that are adapted to implement one or more computer instructions, in particular to load and execute one or more computer instructions to implement a corresponding method flow or a corresponding function.

By way of example, the processor 610 may also be referred to as a central processor (CentralProcessingUnit, CPU). The processor 610 may include, but is not limited to: a general purpose processor, digital signal processor (Digital Signal Processor, DSP), application specific integrated circuit (Application Specific Integrated Circuit, ASIC), field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware components, or the like.

By way of example, computer readable storage medium 620 may be high speed RAM memory or Non-volatile memory (Non-VolatileMemorye), such as at least one magnetic disk memory; alternatively, it may be at least one computer-readable storage medium located remotely from the aforementioned processor 610. In particular, computer-readable storage media 620 include, but are not limited to: volatile memory and/or nonvolatile memory. The nonvolatile Memory may be a Read-Only Memory (ROM), a Programmable ROM (PROM), an Erasable PROM (EPROM), an Electrically Erasable EPROM (EEPROM), or a flash Memory. The volatile memory may be random access memory (Random Access Memory, RAM) which acts as an external cache. By way of example, and not limitation, many forms of RAM are available, such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (Double Data Rate SDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), and Direct memory bus RAM (DR RAM).

As shown in fig. 6, the electronic device 600 may also include a transceiver 630.

The processor 610 may control the transceiver 630 to communicate with other devices, and in particular, may send information or data to other devices or receive information or data sent by other devices. Transceiver 630 may include a transmitter and a receiver. Transceiver 630 may further include antennas, the number of which may be one or more.

It should be appreciated that the various components in the communication device 600 are connected by a bus system that includes a power bus, a control bus, and a status signal bus in addition to a data bus.

In one implementation, the electronic device 600 may be any electronic device having data processing capabilities; the computer readable storage medium 620 has stored therein first computer instructions; first computer instructions stored in computer-readable storage medium 620 are loaded and executed by processor 610 to implement the corresponding steps in the method embodiment shown in fig. 1; in a specific implementation, the first computer instructions in the computer-readable storage medium 620 are loaded by the processor 610 and perform the corresponding steps, and for avoiding repetition, a detailed description is omitted herein.

According to another aspect of the present application, the embodiment of the present application further provides a computer-readable storage medium (Memory), which is a Memory device in the electronic device 600, for storing programs and data. Such as computer-readable storage medium 620. It is understood that the computer readable storage medium 620 herein may include a built-in storage medium in the electronic device 600, and may include an extended storage medium supported by the electronic device 600. The computer-readable storage medium provides storage space that stores an operating system of the electronic device 600. Also stored in this memory space are one or more computer instructions, which may be one or more computer programs 621 (including program code), adapted to be loaded and executed by the processor 610.

According to another aspect of the application, embodiments of the application also provide a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. Such as computer program 621. At this time, the data processing apparatus 600 may be a computer, and the processor 610 reads the computer instructions from the computer-readable storage medium 620, and the processor 610 executes the computer instructions so that the computer performs the method of dividing a study provided in the above-described various alternatives.

In other words, 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. When loaded and executed on a computer, runs the processes of, or implements the functions of, embodiments of the present application, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a computer network, or other programmable apparatus. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, from one website, computer, server, or data center by wired (e.g., coaxial cable, fiber optic, digital subscriber line (digital subscriber line, DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means.

Those of ordinary skill in the art will appreciate that the elements and process steps of the examples described in connection with the embodiments disclosed herein may be implemented as electronic hardware, or as a combination of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.

Finally, it should be noted that the above is only a specific embodiment of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily think about the changes or substitutions within the technical scope of the present application, and the changes or substitutions are all covered by the scope of the present application. Therefore, the protection scope of the application is subject to the protection scope of the claims.

Claims

1. A method of distinguishing between students, comprising:

acquiring a learning level prediction model obtained through learning;

and inputting the number of the academic demands of each cell and the number of the academic supplies of each school into a division model to obtain division results for dividing the m cells into the n schools.

2. The method of claim 1, wherein the inputting the number of the degree demands of each cell and the number of the degree supplies of each school into the division model for division results representing division of the m cells into the n schools comprises:

determining at least one learning division pattern of the ith cell based on whether the jth school is an alternative school to the ith cell;

calculating at least one loss value of the ith cell in the at least one learning compartment division mode by using an objective function in the degree prediction model; the at least one loss value is used for evaluating the loss of the cells with divided degrees in the m cells when the ith cell is divided according to the at least one division mode;

3. The method of claim 2, wherein calculating at least one loss value for the ith cell in the at least one region-dividing mode using an objective function in the degree prediction model, comprises:

determining parameter x _ij The method comprises the steps of carrying out a first treatment on the surface of the If the jth school is an alternative school of the ith cell, the x _ij The value of (2) is a first numerical value; if the jth school is not the candidate school for the ith cell, the x _ij The value of (2) is a second value; the first value is greater than the second value;

based on the x _ij Distance d to school from the ith cell to the jth school _ij And determining a loss value of a current division mode in the at least one degree division mode.

4. A method according to claim 3, wherein said x is based on _ij And determining a loss value of a current learning division mode in the at least one learning division mode according to a check-out distance of the ith cell relative to the jth school, wherein the method comprises the following steps:

based on the x _ij D, said d _ij And determining a loss value for the current learn zone division mode from:

the distance to school of the first i-1 cells in the n cells, the average distance to school of the first i-1 cells, the relaxation factor of the school in which the first i-1 cells are located, and the relaxation factor of the jth school; wherein, the relaxation factor is used for representing the number of students that school needs to be accommodated elastically.

5. The method of claim 4, wherein the x is based on _ij And determining a loss value for the current learn zone division mode, comprising:

based on the x _ij D, said d _ij Determining a first parameter according to the distance from the first i-1 cells to the correction; the first parameter is used for evaluating the total distance to school of the cells with divided degrees in the m cells when the ith cell performs the division according to the current division mode;

based on the x _ij D, said d _ij Determining a second parameter from the distance to the correction of the first i-1 cells and the average distance; the second parameter is used for evaluating the difference between the distance to the school of the cells with the divided degree in the m cells when the ith cell performs the division according to the current division mode;

determining a third parameter based on the relaxation factor of the school in which the first i-1 cell is located and the relaxation factor of the j-th school; the third parameter is used for evaluating the total number of students which need to be elastically accommodated by the school to which the cell with the divided degree is divided in the m cells when the ith cell performs the division according to the current division mode;

a loss value for the current learn zone mode is determined based on the first parameter, the second parameter, and the third parameter.

6. The method of claim 2, wherein the determining whether the j-th school of the n schools is an alternative school to the i-th cell of the m cells comprises:

based on the constraint condition of the objective function, determining whether the jth school is an alternative school of the ith cell.

7. The method of claim 6, wherein the first constraint of the objective function is: the number of schools into which the ith cell is divided is less than or equal to a preset threshold.

8. The method of claim 6, wherein the second constraint of the objective function is: the sum of the number of the degree demands of all the cells divided into the jth school and the number of the degree demands of the ith cell is smaller than or equal to the sum of the number of the degree supplies of the jth school and the relaxation factor threshold of the jth school.

9. The method of claim 6, wherein the third constraint of the objective function is: the sum of the number of the degree demands of all the cells divided into the jth school and the number of the degree demands of the ith cell is smaller than or equal to the threshold value of the degree supply number of the jth school and the degree rate of the jth school; wherein the employment rate threshold is determined based on a school equality of the j-th school.

10. The method according to any one of claims 1 to 9, wherein the acquiring the degree supply number of each of the n schools includes:

determining the class offer number of each school based on the following information:

the six-grade existing class number of each school, the difference value between the class number set when each school is started and the existing class number of each school, the residual classroom number of each school and the number of functional classrooms which can be modified into class classrooms;

and determining the grade feed number of each school based on the grade feed number of each school.

11. The method according to any one of claims 1 to 9, wherein the degree prediction model is a long-short-term memory LSTM network, and the raw data of each cell includes at least one of the following information at the current time: population flow data, birth rate, population following migration, entrance rate; each dimension of the feature vector of each cell is used to represent one item of information in the original data of each cell.

12. A learning area dividing apparatus, comprising:

the construction unit is used for carrying out feature construction on the original data of each cell in the m cells to obtain a feature vector of each cell; m is an integer greater than 0;

the prediction unit is used for inputting the characteristic vector of each cell into the academic predictive model to obtain the academic demand number of each cell;

a dividing unit for inputting the number of the degree demands of each cell and the number of the degree supplies of each school to a division model to obtain division results for dividing the m cells into the n schools.

13. An electronic device, comprising:

a processor adapted to execute a computer program;

a computer readable storage medium having stored therein a computer program which, when executed by the processor, implements the method of any of claims 1 to 11.

14. A computer readable storage medium storing a computer program for causing a computer to perform the method of any one of claims 1 to 11.