Big data-based customized long-term education planning system
Through big data platform and intelligent automation technology, combined with kmeans clustering and LSTM neural network, education planning is dynamically updated, and the experience dependence and data single problem of traditional education planning systems is solved, and the adaptation rate of planning solutions is improved.
Patent Information
- Application Number
- CN202510454479.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-11
- Publication Date
- 2025-07-25
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing educational planning system relies on manual experience, the planning tools are static and the data dimensions are single, resulting in less than 60% of the planning scheme adaptation rate.
The big data platform is used to integrate cross-regional education databases, combine kmeans clustering algorithm and LSTM neural network, dynamically update learning plans, integrate academic data, psychological characteristics and family resources, and optimize decision-making models using reinforcement learning algorithms.
The matching degree between planning schemes and user needs is improved, dynamic response to policy changes and real-time updates are achieved, and the adaptation rate of planning schemes is improved.
Smart Images
Figure CN120374318A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of big data technology, and particularly relates to a customized long-term school entrance planning system based on big data. Background Art
[0002] The following problems generally exist in the current education planning field:
[0003] (1) Experience-dependent decision-making: Traditional consulting services rely on manual experience, and there are problems such as uneven levels of consultants and limited coverage of cases.
[0004] (2) Static planning model: Most mainstream planning tools adopt fixed templates and cannot dynamically respond to policy changes.
[0005] (3) Single data dimension: Existing systems are mostly limited to academic performance analysis, ignoring multiple factors such as interest characteristics (Holland vocational assessment) and family resources (economic support index), resulting in a matching rate of less than 60% for the planning scheme. Summary of the Invention
[0006] In order to at least solve or partially solve the above problems, a customized long-term school entrance planning system based on big data is provided to improve the matching degree between the planning scheme and the real needs of users; intelligent automation is used to analyze and plan data.
[0007] In order to achieve the above purpose, the present invention provides the following technical solutions:
[0008] A customized long-term school entrance planning system based on big data of the present invention includes a database, a planning module, and a learning plan module;
[0009] The database aggregates all cross-regional education databases;
[0010] The planning module includes a user module and a clustering calculation module; the user module sends the school entrance data to the clustering calculation module, and the clustering calculation module inputs the school entrance data into the database and classifies it using the kmeans algorithm to classify the target schools and majors for entrance;
[0011] The learning plan module includes a user academic information module and a data analysis module. The user academic information module sends the current academic data of the user to the data analysis module, and the data analysis module plans the study using the reinforcement learning algorithm and the lstm neural network model, and outputs the daily study plan.
[0012] As a preferred technical solution of the present invention, the school entrance data includes the user's interests, hobbies, family economic situation, and favorite subject types, etc.
[0013] As a preferred technical solution of the present invention, the data analysis module takes academic data as state s, learning plan as action a, and the difference between the academic data of the next stage and that of the previous stage as q value; and constructs an LSTM neural network to obtain the learning plan.
[0014] As a preferred technical solution of the present invention, the learning plan includes the time for learning, learning progress, rest time and entertainment time, and the academic data includes academic performance, learning state and learning mood.
[0015] As a preferred technical solution of the present invention, it further includes an update and push module; the update and push module obtains the latest information on colleges, majors and occupations from the official education website through big data, and automatically updates the database irregularly.
[0016] Compared with the prior art, the beneficial effects of the present invention are as follows:
[0017] The present invention (1) constructs a global knowledge graph
[0018] By establishing a cross-regional education database (covering 31 provinces, municipalities and autonomous regions across the country, with a professional library of more than 2,000 colleges and universities), the system automatically integrates and trains the data.
[0019] (2) Utilize the reinforcement learning algorithm and neural network to construct a dynamic feedback closed loop, and the system automatically optimizes the decision-making model every 14 days, enabling it to be updated in real time according to the current academic status.
[0020] (3) Multi-modal data fusion and clustering architecture
[0021] Integrate multiple aspects such as academic data, psychological characteristics, and family resources (economic support index) for scheme adaptation, extract from big data and classify through clustering algorithms to find suitable target schools. Description of the Drawings
[0022] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention, and do not constitute a limitation to the present invention. In the drawings:
[0023] Figure 1 is a schematic diagram of the overall system module of the present invention;
[0024] In the figure: 1, database; 2, planning module; 3, learning plan module; 4, user module; 5, clustering calculation module; 6, user academic information module; 7, data analysis module; 8, update and push module. Detailed Embodiments
[0025] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.
[0026] Furthermore, if a detailed description of known techniques is not necessary for illustrating the characteristics of the present invention, it will be omitted.
[0027] Example
[0028] like Figure 1 As shown, the present invention provides a customized long-term study planning system based on big data, including a database 1, a planning module 2, and a study plan module 3;
[0029] Database 1 aggregates all cross-regional education databases;
[0030] The planning module 2 includes a user module 4 and a clustering calculation module 5; the user module 4 sends the enrollment data to the clustering calculation module 5, and the clustering calculation module 5 inputs the enrollment data into the database 1 and classifies them using the kmeans algorithm to classify the target enrollment schools and majors;
[0031] The study plan module 3 includes a user academic information module 6 and a data analysis module 7. The user academic information module 6 sends the user's current academic data to the data analysis module 7. The data analysis module 7 uses a reinforcement learning algorithm and an LSTM neural network model to plan the academics and output a daily study plan.
[0032] The education data includes the user's interests, hobbies, family financial situation, and favorite subject types.
[0033] Specifically, the clustering calculation module 5 inputs the enrollment data into the database 1 and classifies them using clustering calculation. The classification in the clustering calculation uses the kmeans clustering algorithm to identify the categories. The kmeans clustering algorithm uses the enrollment data as x (m) , put it into the expert database to get the training sample set {x (1) 、x (2) ……x (m)}, and then set k classes and the centroid of each class to μ1, μ2, ..., μ k , repeat the following process until convergence {
[0034] For each sample i, calculate the class it should belong to
[0035]
[0036] For each class j, recalculate the centroid of the class
[0037]
[0038] }
[0039] x (i) belongs to the training sample set {x (1) 、x (2) ……x (m)}, c (i) represents the class closest to the sample i among the k classes, and the value of c (i) is one from 1 to k. The centroid μ j represents our expected value of the center point of the samples belonging to the same class.
[0040] From the above calculations, C (m) can be obtained, that is, the class closest to the sample m among the k classes.
[0041] Thus, the appropriate colleges and majors for the user to be promoted to can be found. After finding the appropriate colleges and majors for promotion, a study plan needs to be formulated.
[0042] When formulating a study plan, since long - term learning will cause poor learning mood and weariness of learning, and the learning efficiency is also low for a long time. Therefore, it is necessary to reasonably plan the learning time, rest time, and entertainment time. Therefore, the study plan includes the time required for learning, learning progress, rest time, and entertainment time, and the academic data includes academic performance, learning status, and learning mood.
[0043] The data analysis module 7 takes the academic data as the state s, the study plan as the action a, and the difference between the academic data in the next stage and the academic data in the previous stage as the q - value; constructs an LSTM neural network to obtain the study plan.
[0044] The specific steps of the reinforcement learning algorithm are as follows;
[0045] Input the current academic data into the LSTM neural network to obtain the Q - values of different actions a in this state, select the action a with the largest Q - value, and the estimated Q - value at this time is Qest(s,a);
[0046] The neural network feeds back a to the academic state to obtain the academic state S′ after passing through a;
[0047] The network state forwards S′ to the knowledge base. The knowledge base obtains the optimal action a′ in the S′ state through a heuristic algorithm, and obtains the Q - value corresponding to this action, that is, the maximum Q - value maxQ(s′), through the neural network. The maximum Q - value and the real - world reward R fed back from the academic state together constitute the target Q - value Qtarg(s,a) = R + γ×maxQ(s′), where γ is the discount factor;
[0048] After obtaining Qtarg(s,a), the latest Q value is obtained through the Q-learning algorithm update formula Q(s,a)new = Qest(s,a) + α × [Qtarg(s,a) - Qest(s,a)], where α is the learning efficiency;
[0049] After obtaining the latest Q value, the mean square error between the new and old Q values is utilized, and the neural network is trained using the backpropagation method;
[0050] After repeating the above steps until the mean square error between the new and old Q values is less than a predetermined threshold, the neural network also converges. At this time, the vector composed of the academic status is used as the input to obtain the Q values corresponding to different actions.
[0051] Then, the action a is selected according to the maximum q value, that is, the learning plan.
[0052] To ensure the timeliness of the data, it also includes an update and push module 8; the update and push module 8 obtains the latest information on colleges, majors, and occupations from the official education website through big data and automatically updates the database irregularly.
[0053] (1) Construction of the global knowledge graph of the present invention
[0054] By establishing a cross-regional education database (covering 31 provinces, municipalities, and autonomous regions across the country and a professional database of more than 2,000 colleges and universities), the system automatically integrates and trains the data.
[0055] (2) Using the reinforcement learning algorithm and neural network to construct a dynamic feedback closed-loop, the system automatically optimizes the decision-making model every 14 days so that it can be updated in real time according to the current academic status.
[0056] (3) Multimodal data fusion and clustering architecture
[0057] Integrate academic data, psychological characteristics, family resources (economic support index), etc. for scheme adaptation, extract from big data and classify through clustering algorithms to find suitable target schools.
[0058] Finally, it should be noted that the above are only the preferred embodiments of the present invention and are not used to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.
Claims
1. A customized long-term education planning system based on big data, characterized in that, It includes a database (1), a planning module (2), and a learning plan module (3); The database (1) aggregates all cross-regional education databases; The planning module (2) includes a user module (4) and a clustering calculation module (5); the user module (4) sends the entrance examination data to the clustering calculation module (5), and the clustering calculation module (5) inputs the entrance examination data into the database (1) and classifies it using the kmeans algorithm to classify the target colleges and majors; The learning plan module (3) includes a user academic information module (6) and a data analysis module (7). The user academic information module (6) sends the user's current academic data to the data analysis module (7), and the data analysis module (7) uses the reinforcement learning algorithm and the lstm neural network model to plan the academic work and outputs the daily learning plan.
2. The customized long-term further education planning system based on big data according to claim 1, characterized in that, The entrance examination data includes the user's interests, hobbies, family economic situation, and favorite subject types, etc.
3. A customized long-term further education planning system based on big data according to claim 1, characterized in that, The data analysis module (7) takes the academic data as the state s, the learning plan as the action a, and the difference between the academic data of the next stage and the academic data of the previous stage as the q value; constructs an LSTM neural network to obtain the learning plan.
4. The customized long-term further education planning system based on big data according to claim 1, characterized in that, The learning plan includes the time required for learning, the learning progress, the rest time, and the entertainment time, and the academic data includes the learning achievements, the learning status, and the learning mood.
5. A customized long-term further education planning system based on big data according to claim 1, characterized in that It further includes an update and push module (8); the update and push module (8) obtains the latest information on colleges, majors, and occupations from the official education website through big data and automatically updates the database irregularly.