Neural Network-Based Localized Self-Learning Tutoring Management System and Method

Abstract

The present invention relates to the field of artificial intelligence education technology, and specifically discloses a localization self-study tutoring management system based on a neural network, including a user side and a server side; the user side is used to collect the label information of the user and provide the user with teaching content for learning; the server side is used to push teaching content to the user according to the label information of the user; the label information of the user includes an intelligence dimension label K2, a learning time dimension label K3, and an environment dimension label K4. The present invention can screen out teaching content corresponding to textbooks suitable for the user himself / herself.

Description

Technical Field

[0001] The present invention belongs to the technical field of artificial intelligence education, and particularly relates to a local self-study tutoring management system and method based on a neural network. Background Art

[0002] In traditional teaching, different regions have different textbooks. It is not the case that all local students are suitable for learning the local textbooks. The matching degree between textbooks and students is a very difficult value to calculate.

[0003] With the development of the Internet, online education technology has become increasingly mature, and a large amount of teaching content has been produced. However, the teaching content is mainly created based on textbooks and has the characteristics of various regions.

[0004] For a single student, if learning textbooks from different regions, different learning results are bound to occur.

[0005] Facing users in different regions, there are different versions of textbooks. Different versions of textbooks represent different difficulties and teaching progressions. If the learning content does not match the user's version of textbooks, it will surely affect their learning effect.

[0006] The emergence of teaching content enables students to learn textbooks from different regions. However, it is very difficult to screen out the teaching content corresponding to the textbooks suitable for themselves from among numerous teaching contents. Summary of the Invention

[0007] The technical problem to be solved by the present invention is to provide a local self-study tutoring management system based on a neural network for screening out the teaching content corresponding to the textbooks suitable for users in view of the deficiencies in the above-mentioned prior art.

[0008] The first aspect of the present invention discloses a local self-study tutoring management system based on a neural network, including a user side and a server side;

[0009] The user side is used to collect the label information of the user and provide the user with teaching content for learning;

[0010] The server side is used to push teaching content to the user according to the label information of the user;

[0011] The label information of the user includes an intelligence dimension label K2, a learning time dimension label K3, and an environment dimension label K4;

[0012] When the server side pushes teaching content to the user according to the label information of the user, the following steps are included:

[0013] Step 1, retrieve a textbook version label K1;

[0014] Step 2: Input the textbook version label K1, the intelligence dimension label K2, the learning time dimension label K3, and the environment dimension label K4 into the trained neural network model; output the score S of the user on the corresponding textbook version predicted by the neural network model.

[0015] Step 3: Determine whether there is still the next textbook version label K1. If so, go to Step 4; if not, go to Step 5.

[0016] Step 4: Retrieve the next textbook version label K1, and then go to Step 2.

[0017] Step 5: Screen out the teaching content corresponding to the textbook version label K1 with the highest score S and push it to the user.

[0018] For the above localization self-study tutoring management system based on neural network, the server is also used to generate a neural network model, which specifically includes the following steps:

[0019] Step 1: Collect a number of sample data, where the sample data includes the score S of the sample object in a specific exam, the textbook version label K1 corresponding to the specific exam, the intelligence dimension label K2 of the sample object, the learning time dimension label K3, and the environment dimension label K4.

[0020] Step 2: Randomly divide a number of the sample data into training data and test data according to a certain proportion.

[0021] Step 3: Use the training data to train the neural network model; during training, the score S of the sample object in the specific exam is used as the true value of the output, and the textbook version label K1, the intelligence dimension label K2, the learning time dimension label K3, and the environment dimension label K4 are used as inputs.

[0022] Step 4: Test the trained neural network model. If the test passes, output the neural network model y = f(x), where y is the predicted score; x = [K1, K2, K3, K4]; if the test fails, go to Step 2.

[0023] For the above localization self-study tutoring management system based on neural network, when collecting the textbook version label K1 of the specific exam, collect the standard book numbers related to the specific exam that the sample object has studied, and generate the textbook version label K1 based on the combination of multiple standard book numbers.

[0024] For the above localization self-study tutoring management system based on neural network, the neural network model is a GRU neural network model; when training the GRU neural network model, let the weight parameter set of the GRU neural network model be w i and the bias parameter set be b i; Set the loss function threshold L'. When the true value of the loss function L < L', output a set of weight parameter sets w i and the bias parameter set is b i ;

[0025] Each set of weight parameter sets w i and the bias parameter set b i corresponds to a GRU neural network model y = f(x);

[0026] When the server pushes teaching content to the user based on the user's label information, based on each GRU neural network model y = f(x), perform steps 1 - 5 once; then reverse - parse the multiple selected textbook version labels K1 respectively, and each textbook version label K1 generates several standard book numbers;

[0027] Eliminate the standard book numbers that the user has already learned from the several standard book numbers corresponding to each textbook version label K1; select the single textbook version label K1 corresponding to the least remaining standard book numbers, and push the textbook teaching content corresponding to the remaining standard book numbers of this textbook version label K1 to the user.

[0028] In the above - mentioned neural - network - based local self - learning tutoring management system, when training the GRU neural network model, the flower pollination algorithm is used to randomly generate several pollen grains Q i , and each pollen grain Q represents a set of weight parameter sets w i and the bias parameter set is b i ; The initial population is H, the maximum number of iterations is T, the cross - pollination probability is P, 0 < P < 1, and the search boundary is not set.

[0029] In the above - mentioned neural - network - based local self - learning tutoring management system, the cross - pollination probability is t is the current number of iterations, N is the number of pollen grains Q that currently satisfy L < L' z obtained, when the number of pollen grains Q z is 0, N = 1.

[0030] In the above - mentioned neural - network - based local self - learning tutoring management system, the user - side collects the learning time - dimension label K3 of the user, which refers to the estimated learning duration for a specific exam.

[0031] The second aspect of the present invention discloses a neural - network - based local self - learning tutoring management method, including the following steps:

[0032] Step 1, collect the label information of the user; the label information of the user includes the intelligence - dimension label K2, the learning time - dimension label K3, and the environment - dimension label K4;

[0033] Step 2: Push teaching content to the user according to the user's tag information, which specifically includes the following steps:

[0034] Step 2-1: Retrieve a textbook version tag K1;

[0035] Step 2-2: Input the textbook version tag K1, the intelligence dimension tag K2, the learning time dimension tag K3, and the environment dimension tag K4 into the trained neural network model; output the score S of the user on the corresponding textbook version predicted by the neural network model;

[0036] Step 2-3: Determine whether there is still a next textbook version tag K1. If so, go to Step 4; if not, go to Step 5;

[0037] Step 2-4: Retrieve the next textbook version tag K1, and then go to Step 2;

[0038] Step 2-5: Screen out the teaching content corresponding to the textbook version tag K1 with the highest score S and push it to the user.

[0039] The third aspect of the present invention discloses an electronic device, including: a memory and a processor, and the processor is connected to the memory;

[0040] The memory is used to store programs;

[0041] The processor calls the program stored in the memory to execute the localization self-study tutoring management method based on the neural network described in the second aspect above.

[0042] The fourth aspect of the present invention discloses a computer-readable storage medium, characterized in that a computer program is stored thereon, and when the computer program is run by the computer, it executes the localization self-study tutoring management method based on the neural network described in the second aspect above.

[0043] The present invention has the following advantages compared with the prior art: By predicting the user's final score from four dimensions of the textbook version tag K1, the intelligence dimension tag K2, the learning time dimension tag K3, and the environment dimension tag K4, the present invention can obtain a relatively accurate final score, and then inversely deduce the most suitable textbook version tag for the user through the final score, and finally obtain the appropriate teaching content to push to the user, with accurate pushing.

[0044] Next, through the drawings and embodiments, the technical solutions of the present invention will be further described in detail. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a flowchart when the server pushes teaching content to the user according to the user's tag information.

[0046] Figure 2 Flowchart for generating a neural network model by the server Detailed implementation mode

[0047] Example 1

[0048] A localization self-study tutoring management system based on a neural network, including a user side and a server side;

[0049] The user side is used to collect the label information of the user and provide the user with learning teaching content; the label information of the user includes an intelligence dimension label K2, a learning time dimension label K3, and an environment dimension label K4;

[0050] It should be noted that the user side collects the learning time dimension label K3 of the user, which refers to the estimated learning duration for a specific exam. Specifically, the learning duration is the teaching content duration of all associated teaching materials corresponding to a specific exam. For example, if the specific exam is the judicial exam, the teaching content duration is the total duration of the teaching content learning durations corresponding to each law.

[0051] The environment dimension label K4 refers to the environment label when learning teaching content, such as classroom, bedroom, subway, library, study room, etc.;

[0052] The intelligence dimension label K2 is the intelligence evaluation score of the user. Intelligence evaluation is an existing technology and will not be elaborated here.

[0053] The server side is used to push teaching content to the user according to the label information of the user;

[0054] As Figure 1 shown, when the server side pushes teaching content to the user according to the label information of the user, it includes the following steps:

[0055] Step 1, retrieve a teaching material version label K1;

[0056] Step 2, input the teaching material version label K1, the intelligence dimension label K2, the learning time dimension label K3, and the environment dimension label K4 into the trained neural network model; output the score S of the user on the corresponding teaching material version predicted by the neural network model;

[0057] Step 3, determine whether there is still a next teaching material version label K1. If so, enter Step 4. If not, enter Step 5;

[0058] Step 4, retrieve the next teaching material version label K1, and then enter Step 2;

[0059] Step 5, screen out the teaching content corresponding to the teaching material version label K1 with the highest score S and push it to the user.

[0060] It should be noted that taking high school mathematics as an example, there are several versions such as People's Education Edition A, People's Education Edition B, Jiangsu Education Edition, and Beijing Normal University Edition. The corresponding label K1 of the textbooks of each version is input into the neural network model for score prediction. If People's Education Edition A has the highest score, it can be considered that People's Education Edition A is the most suitable for users to learn. The teaching content corresponding to People's Education Edition A is pushed to the users.

[0061] As Figure 2 shown, in another embodiment of the present invention, the server is further configured to generate a neural network model, specifically including the following steps:

[0062] Step1: Collect a number of sample data, where the sample data includes the score S of the sample object in a specific exam, the textbook version label K1 corresponding to the specific exam, the intelligence dimension label K2 of the sample object, the learning time dimension label K3, and the environment dimension label K4;

[0063] Step2: Randomly divide the number of sample data into training data and test data according to a certain ratio;

[0064] Step3: Use the training data to train the neural network model; during training, the score S of the sample object in the specific exam is used as the true value of the output, and the textbook version label K1, the intelligence dimension label K2, the learning time dimension label K3, and the environment dimension label K4 are used as inputs;

[0065] Step4: Test the trained neural network model. If the test passes, the neural network model y = f(x) is output, where y is the predicted score; x = [K1, K2, K3, K4]; if the test fails, go to Step2.

[0066] It should be noted that taking the neural network model for predicting the scores of college entrance examination physics as an example, when using sample data, the scores of each candidate from all over the country in the national college entrance examination papers over the years and the corresponding textbook version label K1, intelligence dimension label K2, learning time dimension label K3, and environment dimension label K4 are extracted. Here, the intelligence dimension label K2 takes the average score of the sample object in junior high school physics exams as the IQ score, the learning time dimension label K3 takes the teaching time as the label K3, and the environment dimension label K4 is the classroom.

[0067] In this embodiment, when collecting the textbook version label K1 of the specific exam, the standard book numbers related to the specific exam learned by the sample object are collected, and the textbook version label K1 is generated based on the combination of multiple standard book numbers. For example, if the textbook version label K1 represents the label of the high school physics textbook in area A, the standard book numbers of each teaching book corresponding to the three years of high school physics are combined to generate the textbook version label K1.

[0068] In this embodiment, the neural network model is a GRU neural network model; when training the GRU neural network model, the weight parameter set of the GRU neural network model is set as w i and the bias parameter set is b i ; a loss function threshold L' is set. When the true value L of the loss function is less than L', a set of weight parameter sets w i and the bias parameter set is b i are output;

[0069] Each set of weight parameter sets w i and the bias parameter set b i corresponds to a GRU neural network model y = f(x);

[0070] When the server pushes teaching content to the user based on the user's label information, based on each GRU neural network model y = f(x), steps 1 - 5 are executed once; then, multiple selected textbook version labels K1 are respectively reverse - parsed, and each textbook version label K1 generates a number of standard book numbers;

[0071] The standard book numbers that the user has already studied among the number of standard book numbers corresponding to each textbook version label K1 are excluded; the single textbook version label K1 corresponding to the least remaining standard book numbers is selected, and the teaching content of the textbooks corresponding to the remaining standard book numbers of this textbook version label K1 is pushed to the user.

[0072] It should be noted that different from the traditional neural network model here, it is not necessary to find the global optimal solution, that is, a unique set of weight parameter sets w i and the bias parameter set b i is not required, because in many actual cases, the user may have already studied the teaching content corresponding to some textbooks, and the user does not necessarily want to get the highest score. Therefore, multiple GRU neural network models y = f(x) that meet the requirements are trained, and finally multiple textbook version labels K1 can be obtained. Finally, the single textbook version label K1 corresponding to the least remaining standard book numbers is selected, which can avoid the user from repeating learning.

[0073] The standard book numbers that the user has already studied can be collected through the user terminal. For example, when the user selects the target as the national college entrance examination physics paper through the user terminal, the server sends textbooks such as the People's Education Edition and the Jiangsu Education Edition corresponding to the national college entrance examination physics paper to the user terminal for the user to select the textbooks that have been studied, so that the standard book numbers that the user has already studied can be obtained.

[0074] In this embodiment, when training the GRU neural network model, the flower pollination algorithm is adopted, and a number of pollen grains Q i are randomly generated, and each pollen grain Q represents a set of weight parameter sets w i and the bias parameter set is b i; The initial population is H, the maximum number of iterations is T, the cross-pollination probability is P, 0 < P < 1, and no search boundary is set.

[0075] It should be noted that since a general GRU neural network model is not required, no search boundary is set, so that multiple local optimal solutions can be obtained, that is, multiple GRU neural network models can be obtained.

[0076] In this embodiment, the cross-pollination probability is t is the current iteration number, and N is the number of pollen grains Q that currently satisfy L < L'. z When the number of pollen grains Q z is 0, N = 1.

[0077] It should be noted that the traditional cross-pollination probability P is a fixed value, which can evolve the population relatively stably. However, the invention does not require stable evolution because the ultimate goal is not to obtain a unique solution. By setting the cross-pollination probability to as the iteration number increases, the cross-pollination probability P can be reduced, avoiding too many pollen grains Q that satisfy L < L'. z .

[0078] Embodiment 2

[0079] A localization self-study tutoring management method based on a neural network includes the following steps:

[0080] Step 1, collect the label information of the user; the label information of the user includes the intelligence dimension label K2, the learning time dimension label K3, and the environment dimension label K4;

[0081] Step 2, push teaching content to the user according to the label information of the user, specifically including the following steps:

[0082] Step 2-1, retrieve a textbook version label K1;

[0083] Step 2-2, input the textbook version label K1, the intelligence dimension label K2, the learning time dimension label K3, and the environment dimension label K4 into the trained neural network model; output the score S of the user on the corresponding textbook version predicted by the neural network model;

[0084] Step 2-3, determine whether there is still a next textbook version label K1. If so, enter Step 4. If not, enter Step 5;

[0085] Step 2-4, retrieve the next textbook version label K1, and then enter Step 2;

[0086] Step 2-5, screen out the teaching content corresponding to the textbook version label K1 with the highest score S and push it to the user.

[0087] It should be noted that taking high school mathematics as an example, there are several versions such as People's Education Edition A, People's Education Edition B, Jiangsu Education Edition, and Beijing Normal University Edition. The corresponding label K1 of the textbooks of each version is input into the neural network model for score prediction. If People's Education Edition A has the highest score, it can be considered that People's Education Edition A is the most suitable for users to learn. The teaching content corresponding to People's Education Edition A is pushed to the users.

[0088] It should be noted that the neural network model is generated through the following steps:

[0089] Step1. Collect a number of sample data, where the sample data includes the score S of the sample object in a specific exam, the textbook version label K1 corresponding to the specific exam, the intelligence dimension label K2 of the sample object, the learning time dimension label K3, and the environment dimension label K4;

[0090] Step2. Randomly divide a number of the sample data into training data and test data according to a certain proportion;

[0091] Step3. Use the training data to train the neural network model; during training, the score S of the sample object in the specific exam is used as the true value of the output, and the textbook version label K1, the intelligence dimension label K2, the learning time dimension label K3, and the environment dimension label K4 are used as inputs;

[0092] Step4. Test the trained neural network model. If the test passes, output the neural network model y = f(x), where y is the predicted score; x = [K1, K2, K3, K4]; if the test fails, go to Step2.

[0093] It should be noted that taking the neural network model for predicting the scores of college entrance examination physics as an example, when using sample data, extract the scores of each candidate from all over the country in the national college entrance examination papers over the years and the corresponding textbook version label K1, intelligence dimension label K2, learning time dimension label K3, and environment dimension label K4. Here, the intelligence dimension label K2 takes the average score of the sample object in junior high school physics exams as the IQ score, the learning time dimension label K3 takes the teaching time as the label K3, and the environment dimension label K4 is the classroom.

[0094] In this embodiment, when collecting the textbook version label K1 of the specific exam, collect the standard book numbers related to the specific exam that the sample object has studied, and generate the textbook version label K1 based on the combination of multiple standard book numbers. For example, if the textbook version label K1 represents the label of the high school physics textbook in Area A, then combine the standard book numbers of each teaching book corresponding to the three years of high school physics to generate the textbook version label K1.

[0095] In this embodiment, the neural network model is a GRU neural network model; when training the GRU neural network model, the weight parameter set of the GRU neural network model is set as w i and the bias parameter set is b i ; a loss function threshold L' is set. When the true value L of the loss function is less than L', a set of weight parameter sets w i and the bias parameter set is b i are output;

[0096] Each set of weight parameter sets w i and the bias parameter set b i corresponds to a GRU neural network model y = f(x);

[0097] When pushing teaching content to the user based on the user's label information, based on each GRU neural network model y = f(x), steps 1 and 2 are executed once; then the multiple selected textbook version labels K1 are respectively reverse-parsed, and each textbook version label K1 generates a number of standard book numbers;

[0098] The standard book numbers that the user has already learned among the number of standard book numbers corresponding to each textbook version label K1 are excluded; the single textbook version label K1 corresponding to the least remaining standard book numbers is selected, and the teaching content of the textbooks corresponding to the remaining standard book numbers of this textbook version label K1 is pushed to the user.

[0099] It should be noted that different from the traditional neural network model here, it is not necessary to find the global optimal solution, that is, a unique set of weight parameter sets w i and the bias parameter set b i are not required, because in many actual cases, the user may have already learned the teaching content corresponding to some textbooks, and the user does not necessarily want to get the highest score. Therefore, multiple GRU neural network models y = f(x) that meet the requirements are trained. Finally, multiple textbook version labels K1 can be obtained, and finally the single textbook version label K1 corresponding to the least remaining standard book numbers is selected, which can avoid the user from repeating learning.

[0100] The standard book numbers that the user has already learned can be collected through the user terminal. For example, when the user selects the target as the national college entrance examination physics paper through the user terminal, the server sends textbooks such as the People's Education Edition and the Jiangsu Education Edition corresponding to the national college entrance examination physics paper to the user terminal for the user to select the textbooks that have been learned, so that the standard book numbers that the user has already learned can be obtained.

[0101] In this embodiment, when training the GRU neural network model, the flower pollination algorithm is adopted, and a number of pollen grains Q are randomly generated i , and each pollen grain Q represents a set of weight parameter sets w i and the bias parameter set is b i; The initial population is H, the maximum number of iterations is T, the cross-pollination probability is P, where 0 < P < 1, and no search boundary is set.

[0102] It should be noted that since a general GRU neural network model is not required, no search boundary is set, so that multiple local optimal solutions can be obtained, that is, multiple GRU neural network models can be obtained.

[0103] In this embodiment, the cross-pollination probability is t is the current iteration number, and N is the number of pollen grains Q that currently satisfy L < L'. z When the number of pollen grains Q z is 0, N = 1.

[0104] It should be noted that the traditional cross-pollination probability P is a fixed value, which can stably evolve the population. However, the invention does not require stable evolution because the ultimate goal is not to obtain a unique solution. By setting the cross-pollination probability to as the iteration number increases, the cross-pollination probability P can be reduced, avoiding an excessive number of pollen grains Q that satisfy L < L'. z .

[0105] Embodiment 3

[0106] A computer-readable storage medium, on which a computer program is stored. When the computer program is run by a computer, it executes the neural network-based localized self-study tutoring management method described in Embodiment 2 above.

[0107] Embodiment 4

[0108] An electronic device, comprising: a memory and a processor, the processor is connected to the memory;

[0109] The memory is used to store programs;

[0110] The processor calls the program stored in the memory to execute the neural network-based localized self-study tutoring management method described in Embodiment 2.

[0111] It should be noted that the electronic device can be, but is not limited to, a personal computer (PC), a tablet computer, a mobile internet device (MID), etc.

[0112] It should be noted that the processor, the memory, and other components that may appear in the electronic device are electrically connected to each other directly or indirectly to achieve data transmission or interaction. For example, the processor, the memory, and other possible components can be electrically connected to each other through one or more communication buses or signal lines.

[0113] It should be noted that each embodiment in this specification is described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. For the same or similar parts among the embodiments, reference can be made to each other.

[0114] In several embodiments provided in this application, it should be understood that the disclosed systems and methods can also be implemented in other ways. The system embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions, and operations of the systems, methods, and computer program products according to multiple embodiments of this application. In this regard, each block in the flowchart or block diagram can represent a module, a program segment, or a part of code, and the module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order from that marked in the accompanying drawings. For example, two consecutive blocks can actually be executed substantially in parallel, and they can sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, as well as the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0115] In addition, in each embodiment of this application, the functional modules can be integrated together to form an independent part, or each module can exist alone, or two or more modules can be integrated to form an independent part.

[0116] When the above-mentioned functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a laptop, a server, a mobile phone, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs.

[0117] The above are only the preferred embodiments of the present invention and do not impose any limitations on the present invention. Any simple modifications, changes, and equivalent structural changes made to the above embodiments according to the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. A neural network-based localized self-study tutoring management system, characterized in that, Including the user side and the server side; The user terminal is used to collect user tag information and provide users with teaching content learning; The server is used to push teaching content to users according to their tag information; The user's label information includes an intelligence dimension label K2, a learning time dimension label K3, and an environment dimension label K4; The server side pushes teaching content to the user according to the user's tag information, including the following steps: Step 1, retrieve a textbook version label K1; Step 2: Input the textbook version label K1, intelligence dimension label K2, learning time dimension label K3 and environment dimension label K4 into the trained neural network model; output the score S of the user on the corresponding textbook version predicted by the neural network model; Collect the textbook version label K1, collect the standard book numbers related to the specific exam that the sample subject has studied, and generate the textbook version label K1 based on a combination of multiple standard book numbers; The neural network model is a GRU neural network model; when training the GRU neural network model, the weight parameter set of the GRU neural network model is set to wi and the bias parameter set is set to bi; the loss function threshold L' is set, and when the true value of the loss function L<L', a set of weight parameter set wi and bias parameter set bi is output; Each set of weight parameter set wi and bias parameter set bi corresponds to a GRU neural network model y=f(x); When the server pushes teaching content to the user according to the user's tag information, based on each GRU neural network model y=f(x), steps 1 to 5 are executed once; then the multiple textbook version tags K1 screened out are reversely parsed respectively, and each textbook version tag K1 generates several standard book numbers; Eliminate the standard book numbers that the user has learned from the several standard book numbers corresponding to each textbook version label K1; select the single textbook version label K1 that has the least corresponding remaining standard book number, and push the textbook teaching content corresponding to the remaining standard book numbers of the textbook version label K1 to the user; Step 3: determine whether there is a next textbook version label K1, if yes, go to step 4, if no, go to step 5; Step 4, retrieve the next textbook version label K1, and then go to step 2; Step 5: Filter out the teaching content corresponding to the textbook version label K1 with the highest score S and push it to the user.

2. The neural network-based localized self-study tutoring management system according to claim 1, wherein The server is also used to generate a neural network model, which specifically includes the following steps: Step 1, collect some sample data, the sample data includes the score S of the sample subject in a specific test, the textbook version label K1 corresponding to the specific test, the intelligence dimension label K2 of the sample subject, the learning time dimension label K3 and the environment dimension label K4; Step 2, randomly dividing the sample data into training data and test data according to a certain ratio; Step 3, use the training data to train the neural network model; during training, the score S of the sample object in a specific test is used as the true value of the output, and the textbook version label K1, intelligence dimension label K2, learning time dimension label K3 and environment dimension label K4 are used as input; Step 4: Test the trained neural network model. If the test passes, output the neural network model y=f(x), where y is the predicted score; x=[K1,K2,K3,K4]; if the test fails, go to Step 2.

3. The localized self-study tutoring management system based on a neural network according to claim 2, characterized in that When the GRU neural network model is trained, a flower pollination algorithm is used to randomly generate a number of pollen grains Qi, each pollen grain Q represents a set of weight parameter sets wi and bias parameter sets bi; the initial population is H, the maximum number of iterations is T, the cross-pollination probability is P, 0<P<1, and no search boundary is set.

4. The neural network-based localized self-study tutoring management system according to claim 3, characterized in that, The cross-pollination probability is , where t is the current iteration number, N is the number of pollen grains Qz that currently satisfy L < L', and when the number of pollen grains Qz is 0, N = 1.

5. The neural network-based localized self-study tutoring management system according to claim 1, characterized in that, The user terminal collects the user's study time dimension label K3, which refers to the estimated study time for a specific exam.

6. A localization self-study tutoring management method based on a neural network, characterized in that, The following steps are involved: Step 1: Collect user tag information; the user tag information includes intelligence dimension tag K2, learning time dimension tag K3 and environment dimension tag K4; Step 2: Pushing teaching content to users based on their tag information, specifically including the following steps: Step 2-1, retrieve a textbook version label K1, collect the standard book numbers related to the specific exam that the sample subject has studied, and generate the textbook version label K1 based on a combination of multiple standard book numbers; Step 2-2, input the textbook version label K1, the intelligence dimension label K2, the learning time dimension label K3 and the environment dimension label K4 into the trained neural network model; Output the score S of the user on the corresponding textbook version predicted by the neural network model, where the neural network model is a GRU neural network model; when training the GRU neural network model, set the weight parameter set of the GRU neural network model to wi and the bias parameter set to bi; set the loss function threshold L', and when the true value of the loss function L<L', output a set of weight parameter set wi and bias parameter set to bi; Each set of weight parameter set wi and bias parameter set bi corresponds to a GRU neural network model y=f(x); Step 2-3, determine whether there is a next textbook version label K1, if yes, go to step 4, if not, go to step 5; Step 2-4, retrieve the next textbook version label K1, and then go to step 2; Step 2-5: Filter out the teaching content corresponding to the textbook version label K1 with the highest score S and push it to the user; When the server pushes teaching content to the user according to the user's tag information, based on each GRU neural network model y=f(x), steps 1 to 5 are executed once; then the multiple textbook version tags K1 screened out are reverse parsed respectively, and each textbook version tag K1 generates several standard book numbers; Eliminate the standard book numbers that the user has learned from the several standard book numbers corresponding to each textbook version label K1; select the single textbook version label K1 that corresponds to the least remaining standard book number, and push the textbook teaching content corresponding to the remaining standard book numbers of the textbook version label K1 to the user.

7. An electronic device, characterized in that, include: A memory and a processor, wherein the processor is connected to the memory; The memory is used to store programs; The processor calls the program stored in the memory to execute the localized self-study tutoring management method based on neural network as claimed in claim 6.

8. A computer-readable storage medium, characterized in that, A computer program is stored thereon, and when the computer program is run by a computer, it executes the neural network-based localized self-study tutoring management method described in claim 6.

Images