Infant personalized learning path planning method and system

Abstract

The invention discloses an infant personalized learning path planning method and system, and relates to the technical field of education, the system comprises a fuzzy planning module, an algorithm implementation module and an algorithm optimization module, the fuzzy planning module is used for investigating an infant just entering a school and rapidly forming a personalized fuzzy learning path, and the algorithm optimization module is used for optimizing the learning path. The algorithm implementation module is used for refining a learning path by utilizing a reinforcement learning algorithm and a collaborative filtering algorithm according to the learning process of the child, and switching the two algorithms according to the collected learning data; the algorithm optimization module is used for optimizing the switching opportunity of the two algorithms and the forming mode of the fuzzy learning path according to the actual learning performance of the child. And the fuzzy programming module comprises a learning data storage module, a learning path analysis module, an influence evaluation module, a involved surface calculation module, an inspection point generation module and a fuzzy path generation module. The method has the characteristic of combining the advantages of two algorithms.

Description

Technical Field

[0001] The present invention relates to the field of educational technology, and specifically to a method and system for planning personalized learning paths for young children. Background Technique

[0002] In the field of artificial intelligence-assisted education, collaborative filtering and reinforcement learning are both very effective learning path planning algorithms. Collaborative filtering is suitable for recommending personalized learning resources and methods based on students' historical feedback data to achieve a tailored solution. Reinforcement learning sets reward signals according to students' current learning performance and optimizes learning resources and methods in real time to maximize long-term benefits.

[0003] However, both algorithms face problems when applied to the planning of young children's learning paths. There is little historical feedback data for young children, and it is difficult to capture personal characteristics. Therefore, it is difficult for the collaborative filtering algorithm to form a tailored solution. The learning process of young children is not as regular as that of older students, and it is difficult to maintain attention for a long time, which will affect learning performance. The process of real-time optimization is difficult to remain stable for a long time and is easily deviated from the original goal. Therefore, it is necessary to design a method and system for planning personalized learning paths for young children that combines the advantages of the two algorithms. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for planning personalized learning paths for young children to solve the problems raised in the above background technique.

[0005] To solve the above technical problems, the present invention provides the following technical solutions: A method and system for planning personalized learning paths for young children, including a fuzzy planning module, an algorithm implementation module, and an algorithm optimization module. The fuzzy planning module is used to examine newly enrolled young children and quickly form a personalized fuzzy learning path. The algorithm implementation module is used to refine the learning path using the reinforcement learning algorithm and the collaborative filtering algorithm according to the learning process of young children and switch between the two algorithms based on the collected learning data. The algorithm optimization module is used to optimize the switching timing of the two algorithms and the formation method of the fuzzy learning path according to the actual learning performance of young children.

[0006] According to the above technical solution, the fuzzy planning module includes a learning data storage module, a learning path analysis module, an influence evaluation module, a coverage calculation module, an inspection point generation module, and a fuzzy path generation module. The learning data storage module and the influence evaluation module are both electrically connected to the learning path analysis module. The coverage calculation module and the influence evaluation module are both electrically connected to the inspection point generation module. The inspection point generation module is electrically connected to the fuzzy path generation module. The learning data storage module is used to store and manage historical early childhood learning data and learning path data. The learning path analysis module is used to disassemble and analyze the stored learning paths. The influence evaluation module is used to identify key data that significantly affects the formation of personalized learning paths. The coverage calculation module is used to calculate the coverage of learning data for each learning path. The inspection point generation module is used to generate inspection points for newly enrolled children. The fuzzy path generation module is used to generate fuzzy learning paths based on the inspection results of each student's inspection points; The algorithm implementation module includes a collaborative filtering algorithm module, a reinforcement learning algorithm module, a data accumulation judgment module, and an algorithm switching module. The collaborative filtering algorithm module generates personalized and accurate learning paths for the accumulated data. The collaborative filtering algorithm module and the reinforcement learning algorithm module are electrically connected to the algorithm switching module. The data accumulation judgment module is electrically connected to the algorithm switching module. The reinforcement learning algorithm module continuously adjusts the learning path according to the real-time learning performance of the children. The data accumulation judgment module determines whether the current learning data is sufficient to decide when to switch the algorithm or adjust the learning path. The algorithm switching module realizes the switching between the reinforcement learning algorithm and the collaborative filtering algorithm; The algorithm optimization module includes a learning outcome evaluation module, a switching timing adjustment module, and a refinement degree adjustment module. The learning outcome evaluation module is electrically connected to the switching timing adjustment module and the refinement degree adjustment module. The fuzzy path generation module is electrically connected to the refinement degree adjustment module. The switching timing adjustment module is electrically connected to the algorithm switching module. The learning outcome evaluation module evaluates the learning outcomes of the children and provides data support to optimize the learning path. The switching timing adjustment module is used to dynamically adjust the timing of algorithm switching. The refinement degree adjustment module is used to adjust the refinement degree when generating the fuzzy path.

[0007] A method for planning an early childhood personalized learning path includes the following steps: S0. The system stores and collects historical early childhood learning data of children who have participated in learning education and entered the system, including inspection points, learning data, learning paths, and learning outcome evaluation data; S1. Use the learning path analysis module to disassemble and analyze the stored learning paths, disassemble the learning paths into individual basic path nodes, identify key learning data that has a great impact on the formation of personalized learning paths and covers a wide range, and find the examination points for evaluating children's learning achievements corresponding to the key learning data; S2. When children enter school, use these examination points to evaluate their basic strengths and interests, and form a personalized fuzzy learning path; S3. Accumulate learning data based on children's real-time learning performance, use the reinforcement learning algorithm to continuously improve and adjust the learning path. When enough learning data is accumulated in a certain field, switch to the collaborative filtering algorithm to generate a personalized precise learning path, and continue to use the reinforcement learning algorithm to accumulate data in other fields; S4. Calculate the interest index of the child's current learning path, adjust the refinement degree when generating the fuzzy path according to the average interest index of all children in a certain field, and adjust the timing of algorithm switching according to the child's learning achievements in a certain field.

[0008] According to the above technical solution, in S1, when analyzing the learning path, it specifically includes the following steps: S1-1. The basic path nodes that make up the learning path are composed of various knowledge and skill points connected in series and in parallel. One knowledge and skill point is paired with multiple learning activity forms. The learning path priorities of the knowledge and skill points connected in series are different, and the learning path priorities of the knowledge and skill points connected in parallel are the same; S1-2. Count the knowledge and skill points of all learning paths. The determining factor for the coverage is the frequency of occurrence of the current knowledge and skill point, and the determining factor for the influence is the number of knowledge and skill points that the knowledge and skill point can branch into. Count the frequency of occurrence of each knowledge and skill point to obtain , where is the number of knowledge and skill points. Count the number of knowledge and skill points branched from each knowledge and skill point to obtain . Let the serial number of the current knowledge and skill point be , then its consideration weight in forming the fuzzy learning path , where is the coverage weight, is the influence weight. The higher the influence, the closer it is to simple and initial knowledge and skills. The higher the coverage, the more general this knowledge and skill is.

[0009] According to the above technical solution, in S2, forming a personalized fuzzy learning path specifically includes: S2-1. When evaluating strengths, select the weight High knowledge and skill points, as the basic components of the fuzzy learning path, are selected when generating the assessment method Knowledge and skill points higher than the set value as the assessment points. Randomly select one of all the learning activity forms corresponding to these knowledge and skill points as the design prototype of the assessment point. The learning activity forms include interactive games, physical operations, group discussions, and paper-and-pencil exercises. Based on the design prototype, use algorithms to generate the design assessment method. During the assessment, play the first part of the learning activity form that requires teacher instruction to students, let children complete the second part independently, and the teacher evaluates the completion degree of children for the second part as the evaluation result of the strong points; S2-2. When evaluating the interest points, the teacher asks open-ended questions to children and their parents, combines the degree of interest of children during the evaluation of strong points, and uses the system to generate weights to score the interest scale of high knowledge and skill points. According to the evaluation results of strong points and the scoring results of the interest scale, generate the basic components of the current child's fuzzy learning path, and connect the basic components to each other to form a fuzzy learning path.

[0010] According to the above technical solution, in S3, the specific method for improving and adjusting the learning path is: S3-1. After the teacher completes the basic teaching work of a certain knowledge and skill point for all children, classify the children according to the types of nodes on the learning path, so that children with the same nodes receive advanced teaching for this knowledge and skill point. During the teaching process, the teacher enters the positive and negative performances of each child in the teaching process into the system. Positive performances include answering questions correctly, actively participating in learning activity forms, and having a test score higher than the average. Negative performances include lack of attention, answering questions incorrectly, and having a test score lower than the average. Enter the degree of enthusiasm of all children for this knowledge and skill point , where is the number of children. If the current child number is , the more positive performances, the bigger, and ; S3-2. According to the reinforcement learning algorithm, match more learning resources for children with a high degree of enthusiasm. Connect multiple extended knowledge and skill points based on this as nodes at the current knowledge and skill point. Do not extend the knowledge and skill points for children with a low degree of enthusiasm at this knowledge and skill point. Continue to carry out basic teaching for children with the same nodes according to the curriculum arrangement; S3-3. Select Knowledge and skill points higher than the set value as a field. As the teaching process continues, when the number of nodes connected by a child in a certain field exceeds the switching threshold When, in the current field, a collaborative filtering algorithm is used to generate a refined learning path based on the positive and negative performances of young children during the teaching process, and then the learning path of this young child in the current field follows the refined learning path for teaching. After that, there is no need to optimize the learning path in real time, and it can maintain long-term stability. Since it is obtained separately for different fields, and a young child can match multiple fields, the generation strategy of the learning path is more flexible and more in line with the characteristics of the current young child itself.

[0011] According to the above technical solution, in S4, the adjustment of the refinement degree during fuzzy path generation specifically includes: S4-1. By collecting the positivity degrees of all knowledge and skill points collected from all young children in the teaching process in S3-2 , according to the change trend of the positivity degree to adjust the size of the set value , and optimize the set value at fixed intervals. Suppose it is adjusted times, and the optimized set value after the -th adjustment; in the formula is the set value at the -th adjustment, is the positivity degree collected at the -th adjustment, is the weight coefficient of the size of the positivity degree itself in the adjustment, is the weight coefficient of the change trend of the positivity degree in the adjustment. When is higher, it means that the student's positivity degree is higher. At this time, the current learning path needs less change and can maintain the status quo. When the student's positivity degree is lower, it means that the learning path needs more change and it is necessary to reduce . At this time, when selecting the fuzzy learning path, the more knowledge and skill points above , the more inspection points need to be considered at the beginning. More data needs to be obtained to inspect the student at the initial stage to improve the applicability of the learning path planning, reduce the space for the reinforcement learning algorithm to play, and obtain a more stable solution. On the contrary, more space is given to the reinforcement learning algorithm, with more uncertainties but higher development potential.

[0012] According to the above technical solution, in S4, the adjustment of the algorithm switching timing specifically includes: S4-2. When adjusting the algorithm switching timing, calculate the sum of the positivity degrees of all knowledge and skill points related to a certain field of the current student and then take the average to obtain , where is the number of knowledge and skill points related to this field. When When it is larger, the switching threshold The higher it is, the later the algorithm is switched, giving more room for the reinforcement learning algorithm to play, and fully giving the child real-time optimization space before the child's interest fades.

[0013] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: The present invention first collects all the learning data of children who have participated in learning and education in history and enters them into the system, finds out the key learning data that has a great impact on the personalized customization plan and covers a wide range, and matches the corresponding inspection methods, inspects the children who have just entered school and quickly forms a personalized fuzzy learning path; According to the child's subsequent learning performance, the reinforcement learning algorithm is used to gradually improve the learning path until enough learning data is accumulated in a certain field, and then switched to the collaborative filtering algorithm to generate a personalized accurate learning path in this field, which helps to stabilize and refine the learning path. The reinforcement learning algorithm is continued to accumulate data in other fields, which helps to optimize the plan in real time according to the positive feedback mechanism. By combining the two algorithms at different times, the advantages of the two algorithms are fully utilized, and it is more suitable for the learning path planning of children. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification, and are used to explain the present invention together with the embodiments of the present invention, and do not constitute a limitation to the present invention. In the drawings: Figure 1 is a schematic diagram of the overall module structure of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0015] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0016] Please refer to Figure 1 , the present invention provides a technical solution: a method and system for planning a personalized learning path for children, including a fuzzy planning module, an algorithm implementation module, and an algorithm optimization module. The fuzzy planning module is used to inspect children who have just entered school and quickly form a personalized fuzzy learning path. The algorithm implementation module is used to refine the learning path according to the learning process of children using the reinforcement learning algorithm and the collaborative filtering algorithm, and switch the two algorithms according to the collected learning data. The algorithm optimization module is used to optimize the switching timing of the two algorithms and the formation method of the fuzzy learning path according to the actual learning performance of children; The fuzzy planning module includes a learning data storage module, a learning path analysis module, an influence evaluation module, a coverage calculation module, an inspection point generation module, and a fuzzy path generation module. The learning data storage module and the influence evaluation module are both electrically connected to the learning path analysis module. The coverage calculation module and the influence evaluation module are both electrically connected to the inspection point generation module. The inspection point generation module is electrically connected to the fuzzy path generation module. The learning data storage module is used to store and manage historical early childhood learning data and learning path data. The learning path analysis module is used to disassemble and analyze the stored learning paths. The influence evaluation module is used to identify key data that significantly affects the formation of personalized learning paths. The coverage calculation module is used to calculate the coverage of learning data for each learning path. The inspection point generation module is used to generate inspection points for newly enrolled children. The fuzzy path generation module is used to generate fuzzy learning paths based on the inspection results of each student's inspection points; The algorithm implementation module includes a collaborative filtering algorithm module, a reinforcement learning algorithm module, a data accumulation judgment module, and an algorithm switching module. The collaborative filtering algorithm module generates personalized and precise learning paths for the accumulated data. The collaborative filtering algorithm module and the reinforcement learning algorithm module are electrically connected to the algorithm switching module. The data accumulation judgment module is electrically connected to the algorithm switching module. The reinforcement learning algorithm module continuously adjusts the learning path according to the real-time learning performance of children. The data accumulation judgment module determines whether the current learning data is sufficient to decide when to switch the algorithm or adjust the learning path. The algorithm switching module realizes the switching between the reinforcement learning algorithm and the collaborative filtering algorithm; The algorithm optimization module includes a learning outcome evaluation module, a switching timing adjustment module, and a refinement degree adjustment module. The learning outcome evaluation module is electrically connected to the switching timing adjustment module and the refinement degree adjustment module. The fuzzy path generation module is electrically connected to the refinement degree adjustment module. The switching timing adjustment module is electrically connected to the algorithm switching module. The learning outcome evaluation module evaluates the learning outcomes of children and provides data support to optimize the learning path. The switching timing adjustment module is used to dynamically adjust the timing of algorithm switching. The refinement degree adjustment module is used to adjust the refinement degree during the generation of fuzzy paths; The method for planning personalized learning paths for children includes the following steps: S0. The system stores and collects historical early childhood learning data of children who have participated in learning and education and are entered into the system, including inspection points, learning data, learning paths, and learning outcome evaluation data; S1. Use the learning path analysis module to disassemble and analyze the stored learning paths, disassemble the learning paths into each basic path node, identify key learning data that has a large impact and wide coverage on the formation of personalized learning paths, and find the inspection points for evaluating children's learning outcomes corresponding to the key learning data; S2. When children start school, use these assessment points to evaluate their basic strengths and interests, and form a personalized fuzzy learning path. S3. Accumulate learning data based on children's real-time learning performance, and continuously improve and adjust the learning path using the reinforcement learning algorithm. When enough learning data has been accumulated in a certain field, switch to the collaborative filtering algorithm to generate a personalized precise learning path, and continue to use the reinforcement learning algorithm to accumulate data in other fields. S4. Calculate the interest index of the child's current learning path, adjust the refinement degree when generating the fuzzy path according to the average interest index of all children in a certain field, and adjust the timing of algorithm switching according to the child's learning achievements in a certain field. In S1, when analyzing the learning path, it specifically includes the following steps: S1-1. The basic path nodes that make up the learning path are composed of various knowledge and skill points connected in series and in parallel. One knowledge and skill point is paired with multiple learning activity forms. The knowledge and skill points connected in series have different learning priorities, and the knowledge and skill points connected in parallel have the same learning priorities. S1-2. Count the knowledge and skill points of all learning paths. The determining factor for the scope involved is the frequency of occurrence of the current knowledge and skill point, and the determining factor for the influence is the number of knowledge and skill points that the knowledge and skill point can branch into. Count the frequency of occurrence of each knowledge and skill point to obtain , where is the number of knowledge and skill points. Count the number of knowledge and skill points branched from each knowledge and skill point to obtain . Let the current knowledge and skill point number be , then its consideration weight in forming the fuzzy learning path , where is the weight for the scope involved, is the weight for the influence. The higher the influence, the closer it is to simple and initial knowledge and skills. The higher the scope involved, the more general this knowledge and skill is. In S2, forming a personalized fuzzy learning path specifically includes: S2-1. When evaluating strengths, select knowledge and skill points with a high weight as the basic components of the fuzzy learning path. When generating the assessment method, select higher than the set value Select a knowledge and skill point as the assessment point, and randomly select one of all the learning activity forms corresponding to this knowledge and skill point as the design prototype of the assessment point. The learning activity forms include interactive games, physical operations, group discussions, and paper-and-pencil exercises. Based on the design prototype, use algorithms to generate a design assessment method. During the assessment, play the first part of the learning activity form that requires teacher instruction to the students, let the children complete the second part independently, and the teacher evaluates the completion degree of the children for the second part as the evaluation result of the strong points; S2-2. When evaluating the points of interest, the teacher asks open-ended questions to the children and their parents, and combines the degree of interest of the children during the evaluation of their strong points to generate weights using the system Score using the interest scale of the knowledge and skill points with high scores. According to the evaluation results of the strong points and the score results of the interest scale, generate the basic components of the current child's fuzzy learning path, and connect the basic components to each other to form a fuzzy learning path; In S3, the specific methods for improving and adjusting the learning path are as follows: S3-1. After the teacher completes the basic teaching work on a certain knowledge and skill point for all children, classify the children according to the types of nodes on the learning path, so that children with the same nodes receive advanced teaching for this knowledge and skill point. During the teaching process, the teacher enters the positive and negative performances of each child into the system. Positive performances include answering questions correctly, actively participating in learning activity forms, and having a test score higher than the average. Negative performances include lack of attention, answering questions incorrectly, and having a test score lower than the average. Enter the degree of positivity of all children for this knowledge and skill point , where is the number of children. If the current child number is , the more positive performances, the bigger, and ; S3-2. According to the reinforcement learning algorithm, match more learning resources to children with a high degree of positivity. Connect multiple extended knowledge and skill points based on this as nodes for the current knowledge and skill point, and do not extend the knowledge and skill points for children with a low degree of positivity. Continue to carry out basic teaching for children with the same nodes according to the curriculum arrangement; S3-3. Select higher than the set value of the knowledge and skill points as a field. As the teaching process continues, when the number of nodes connected by a child in a certain field exceeds the switching threshold When learning, the collaborative filtering algorithm is used in the current field to generate a refined learning path based on the positive and negative performance of the child in the teaching process. After that, the learning path of this child in the current field follows the refined learning path for teaching. There is no need to optimize the learning path in real time, and it can maintain long-term stability. Because it is derived from different fields, and a child can match multiple fields, the generation strategy of the learning path is more flexible and more in line with the characteristics of the current child. In S4, the refinement level of fuzzy path generation is adjusted, including: S4-1, by collecting the enthusiasm of all the knowledge and skills points collected by all children in S3-2 during the teaching process , according to the degree of positivity The trend of change to the set value The size of the To optimize, adjust sequence The optimized setting value after adjustment ; In the formula For the The setting value for the first adjustment is For the The positive level collected during the adjustment, is the weight coefficient of the positive degree itself during adjustment, is the weight coefficient of the positive degree change trend in the adjustment. When the level is higher, it means that the students are more motivated. At this time, the current learning path does not need to be changed and the status quo can be maintained. When the level of students is lower, it means that the learning path needs to be changed more and the learning path needs to be lowered. , when the fuzzy learning path is selected, it is higher than The more knowledge and skills a student has, the more points need to be examined at the beginning. In the early stage, more data is needed to examine students in order to improve the applicability of learning path planning. This reduces the room for reinforcement learning algorithms to play in exchange for a more stable solution. On the contrary, it gives reinforcement learning algorithms more room to play, which has more uncertainty but higher development potential. In S4, the timing of algorithm switching is adjusted specifically including: S4-2. When adjusting the timing of algorithm switching, the current student's enthusiasm for all knowledge and skill points involved in a certain field is summed up and averaged to obtain ,in The number of knowledge and skill points involved in this field is The larger the switching threshold The higher it is, the later the algorithm switch is made, giving more room for the reinforcement learning algorithm to play, and fully giving the real-time optimization space to the child before their interest fades.

[0017] First, collect all the historical learning data of children who participated in learning and education and were entered into the system, find out the key learning data that has a great impact on the personalized customization plan and covers a wide range, and match the corresponding inspection methods. Inspect the newly enrolled children and quickly form a personalized fuzzy learning path. According to the children's subsequent learning performance, use the reinforcement learning algorithm to gradually improve the learning path until enough learning data is accumulated in a certain field, and then switch to the collaborative filtering algorithm to generate a personalized accurate learning path in this field, which helps to stabilize and refine the learning path. In other fields, continue to use the reinforcement learning algorithm to accumulate data, which helps to optimize the plan in real time according to the positive feedback mechanism. By combining the two algorithms at different times, the advantages of the two algorithms are fully utilized, which is more suitable for the learning path planning of children.

[0018] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or elements inherent to such process, method, article or device.

[0019] 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 personalized learning path planning system for young children, characterized in that: It includes a fuzzy planning module, an algorithm implementation module, and an algorithm optimization module. The fuzzy planning module is used to examine newly enrolled children and quickly form personalized fuzzy learning paths. The algorithm implementation module is used to refine the learning paths using reinforcement learning algorithms and collaborative filtering algorithms according to the learning process of children, and switch between the two algorithms based on the collected learning data. The algorithm optimization module is used to optimize the switching timing of the two algorithms and the formation method of the fuzzy learning paths according to the actual learning performance of children.

2. The personalized learning path planning system for toddlers according to claim 1, characterized in that: The fuzzy planning module includes a learning data storage module, a learning path analysis module, an influence evaluation module, a coverage calculation module, an examination point generation module, and a fuzzy path generation module. The learning data storage module and the influence evaluation module are both electrically connected to the learning path analysis module. The coverage calculation module and the influence evaluation module are both electrically connected to the examination point generation module. The examination point generation module is electrically connected to the fuzzy path generation module. The learning data storage module is used to store and manage historical children's learning data and learning path data. The learning path analysis module is used to disassemble and analyze the stored learning paths. The influence evaluation module is used to identify key data that significantly affects the formation of personalized learning paths. The coverage calculation module is used to calculate the coverage of learning data for each learning path. The examination point generation module is used to generate examination points for newly enrolled children. The fuzzy path generation module is used to generate fuzzy learning paths based on the examination results of each student's examination points. The algorithm implementation module includes a collaborative filtering algorithm module, a reinforcement learning algorithm module, a data accumulation judgment module, and an algorithm switching module. The collaborative filtering algorithm module generates personalized accurate learning paths for the accumulated data. The collaborative filtering algorithm module and the reinforcement learning algorithm module are electrically connected to the algorithm switching module. The data accumulation judgment module is electrically connected to the algorithm switching module. The reinforcement learning algorithm module continuously adjusts the learning path according to the real-time learning performance of children. The data accumulation judgment module determines whether the current learning data is sufficient to decide when to switch algorithms or adjust the learning path. The algorithm switching module realizes the switching between the reinforcement learning algorithm and the collaborative filtering algorithm. The algorithm optimization module includes a learning achievement evaluation module, a switching timing adjustment module, and a refinement degree adjustment module. The learning achievement evaluation module is electrically connected to the switching timing adjustment module and the refinement degree adjustment module. The fuzzy path generation module is electrically connected to the refinement degree adjustment module. The switching timing adjustment module is electrically connected to the algorithm switching module. The learning achievement evaluation module evaluates the learning achievements of children and provides data support to optimize the learning path. The switching timing adjustment module is used to dynamically adjust the switching timing of the algorithm. The refinement degree adjustment module is used to adjust the refinement degree when generating the fuzzy path.

3. Method for planning personalized learning paths for young children, characterized in that: It includes the following steps: S0. The system stores and collects the learning data of children who have historically participated in learning education and been entered into the system, including examination points, learning data, learning paths, and learning achievement evaluation data. S1. Use the learning path analysis module to disassemble and analyze the stored learning paths, disassemble the learning paths into individual basic path nodes, identify key learning data that has a great impact on and a wide coverage of the formation of personalized learning paths, and search for the inspection points when evaluating children's learning achievements corresponding to the key learning data; S2. When children enter school, use these inspection points to evaluate their basic strengths and interests, and form a personalized fuzzy learning path; S3. Accumulate learning data based on children's real-time learning performance, continuously improve and adjust the learning path using the reinforcement learning algorithm. When enough learning data is accumulated in a certain field, switch to the collaborative filtering algorithm to generate a personalized precise learning path, and continue to use the reinforcement learning algorithm to accumulate data in other fields; S4. Calculate the interest index of the child's current learning path, adjust the refinement degree when generating the fuzzy path according to the average interest index of all children in a certain field, and adjust the timing of algorithm switching according to the child's learning achievements in a certain field.

4. The method for planning an individualized learning path for young children according to claim 3, wherein: In S1, when analyzing the learning path, it specifically includes the following steps: S1-1. The basic path nodes that make up the learning path are composed of various knowledge and skill points connected in series and in parallel with each other. One knowledge and skill point is paired with multiple learning activity forms. The learning sequence priorities of the knowledge and skill points connected in series are different, and the learning sequence priorities of the knowledge and skill points connected in parallel are the same; S1-2. Count the knowledge and skill points of all learning paths. The determining factor for the coverage is the frequency of occurrence of the current knowledge and skill point, and the determining factor for the influence is the number of knowledge and skill points that the knowledge and skill point can branch into. Count the frequency of occurrence of each knowledge and skill point to obtain , where is the number of knowledge and skill points. Count the number of knowledge and skill points branched from each knowledge and skill point to obtain . Let the serial number of the current knowledge and skill point be , then its consideration weight in forming the fuzzy learning path , where is the coverage weight,[[]] is the influence weight.

5. The method for planning an individualized learning path for young children according to claim 4, wherein: In S2, forming a personalized fuzzy learning path specifically includes: S2-1. When evaluating the strong points, select the weight of high knowledge and skill points as the basic component units of the fuzzy learning path. When generating the assessment methods, select knowledge and skill points higher than the set value as the assessment points. Randomly select one of all the learning activity forms corresponding to these knowledge and skill points as the design prototype of the assessment point. The learning activity forms include interactive games, physical operations, group discussions, and paper-and-pencil exercises. Based on the design prototype, use algorithms to generate the design assessment methods. During the assessment, play the first part of the learning activity form that requires teacher instruction to the students, let the children complete the second part independently, and the teacher evaluates the completion degree of the children for the second part as the assessment result of the strong points; S2-2. When evaluating the points of interest, the teacher asks the children and their parents open-ended questions, combines the degree of interest shown by the children during the evaluation of their strengths, and uses the system to generate weights. Rate them using an interest scale for the knowledge and skill points with high interest. Based on the evaluation results of the strengths and the scores of the interest scale, generate the basic components of the current child's fuzzy learning path, and connect the basic components to form a fuzzy learning path.

6. The method for planning an individualized learning path for young children according to claim 5, wherein: In S3, the specific method for improving and adjusting the learning path is: S3-1. After the teacher completes the basic teaching work on a certain knowledge and skill point for all children, the children are classified according to the types of nodes on the learning path, so that children with the same nodes receive advanced teaching for this knowledge and skill point. During the teaching process, the teacher enters the positive and negative performances of each child in the teaching process into the system. The positive performances include answering questions correctly, actively participating in learning activities, and having a test score higher than the average. The negative performances include lack of attention, answering questions incorrectly, and having a test score lower than the average. The positive degree of all children for this knowledge and skill point is entered , where is the number of children. If the current child number is , the more positive performances, the bigger, and ; S3-2. According to the reinforcement learning algorithm, match more learning resources to children with a high level of enthusiasm, connect multiple extended knowledge and skill points based on this as nodes in parallel at the current knowledge and skill point, do not extend at this knowledge and skill point for children with a low level of enthusiasm, and continue to carry out basic teaching for children with the same nodes according to the curriculum arrangement; S3-3. Selection Knowledge and skill points higher than the set value are regarded as a domain. As the teaching process continues, when the number of nodes connected by a child in a certain domain exceeds the switching threshold , a refined learning path is generated in the current domain according to the positive and negative performances of the child during the teaching process using the collaborative filtering algorithm. After that, the learning path of this child in the current domain follows the refined learning path for teaching.

7. The method for planning an individualized learning path for young children according to claim 6, wherein: In S4, adjusting the refinement degree when generating the fuzzy path specifically includes: S4-1. By collecting the enthusiasm levels of all the knowledge and skill points collected by all the children during the teaching process in S3-2 , according to the changing trend of the enthusiasm level , adjust the magnitude of the set value , and optimize the set value at fixed intervals. Suppose the adjustment has been carried out times, and the optimized set value after the th adjustment is ; where is the set value at the th adjustment, is the enthusiasm level collected at the th adjustment, is the weight coefficient of the magnitude of the enthusiasm level itself during the adjustment, is the weight coefficient of the changing trend of the enthusiasm level during the adjustment. is the weight coefficient of the changing trend of the enthusiasm level during the adjustment.

8. The method for planning a personalized learning path for young children according to claim 7, characterized in that: In S4, adjusting the timing of algorithm switching specifically includes: S4-2. When adjusting the timing of algorithm switching, sum up the enthusiasm of the current student for all knowledge and skill points related to a certain field and then calculate the average value to obtain , where is the number of knowledge and skill points related to this field. When is larger, the switching threshold is increased more, and the algorithm switching is carried out later.

Images