Big data-based personalized learning path recommendation method and system

Abstract

The invention relates to the technical field of big data analysis, and discloses a personalized learning path recommendation method and system based on big data, and the system comprises a data collection processing module, a knowledge graph construction module, a path optimization engine and a dynamic adjustment module. The system can dynamically trigger learning path reconstruction, improve the path matching degree in the middle and later periods of the learning process, set cognitive load constraint terms in a path optimization objective function, avoid continuous arrangement of high-complexity knowledge points, ensure that the learning process conforms to the cognitive psychology law, dynamically update knowledge attribute parameters through a matrix transformation mechanism, and improve the learning efficiency. The learning path is kept in an optimal state all the time, learning feedback data is continuously collected, the system can dynamically trigger learning path reconstruction, the path matching degree in the middle and later periods of the learning process is improved, a cognitive load constraint item is set in a path optimization objective function, and it is guaranteed that the learning process conforms to the cognitive psychology law.

Description

Technical Field

[0001] The present invention relates to the technical field of big data analysis, and in particular to a method and system for recommending personalized learning paths based on big data. Background Art

[0002] With the rapid development of internet technology, online education, as an emerging teaching model, has gained widespread application. Online education breaks the constraints of time and space, allowing students to access educational resources anytime and anywhere, greatly improving the flexibility and convenience of learning. Personalized learning path recommendation technology, a core component of intelligent education, has been widely used in online education platforms and smart teaching systems in recent years.

[0003] Although existing technologies have made certain progress, they still have the following key defects: 1. Insufficient dynamic adaptability: The paths generated by existing systems are mostly static sequences, and the path structure cannot be dynamically adjusted based on real-time learning feedback, resulting in a significant decrease in learning efficiency as progress progresses.

[0004] 2. Cognitive load imbalance: Traditional methods ignore the quantitative management of cognitive load, and often result in highly complex knowledge points being arranged continuously.

[0005] 3. Single optimization dimension: Existing optimization algorithms mostly focus on a single goal and have not established a multi-objective collaborative optimization mechanism such as knowledge mastery, cognitive load, and interest preferences.

[0006] 4. Insufficient utilization of knowledge graphs: Most systems only use knowledge graphs to represent basic topological relationships, and fail to deeply integrate attribute parameters such as cognitive complexity and interdisciplinary associations, resulting in a lack of scientificity in path generation.

[0007] Therefore, the present invention provides a personalized learning path recommendation method and system based on big data, which can integrate the dynamic knowledge graph of cognitive load quantification and multi-dimensional ability assessment, design a multi-objective collaborative optimization mechanism, and realize dynamic path reconstruction based on real-time feedback. Summary of the Invention

[0008] (1) Technical problems solved In response to the shortcomings of the existing technology, the present invention provides a method and system for recommending personalized learning paths based on big data, which solves the problems raised in the above background technology.

[0009] (2) Technical solution To achieve the above objectives, the present invention provides the following technical solutions: a method and system for recommending personalized learning paths based on big data, the method comprising the following steps: S1. Collect learners' multi-dimensional learning characteristic data, including learning behavior data, ability level data, and interest preference data; S2. Preprocessing and feature extraction are performed on the multi-dimensional learning feature data to generate a learner feature vector; S3. Construct a subject knowledge graph, which includes knowledge point nodes, inter-node relationships, and knowledge attribute parameters; S4. Performing a learning needs analysis based on the learner's feature vector and knowledge graph to generate a target knowledge point set and a weak knowledge point set; S5. Generate an initial learning path based on the target knowledge point set and the weak knowledge point set in combination with preset learning rules; S6. Dynamically optimize the initial learning path using an intelligent optimization algorithm to generate a personalized learning path; S7. Adjusting the personalized learning path in real time based on the learning feedback data, and outputting an optimized learning path; Wherein, the S6 includes: S61, establishing a path optimization objective function, wherein the function includes parameters of knowledge mastery, learning time, and cognitive load weight; S62. Searching for the optimal solution of the objective function in the path solution space by improving the Osprey optimization algorithm; S63. Map the optimal solution into a knowledge point sequence to generate a personalized learning path.

[0010] Preferably, the S1 specifically includes: S11. Collect learning behavior data through learning terminals, including learning time, answer accuracy, and resource access frequency; S12. Collect ability level data through the ability assessment module, including knowledge mastery and cognitive ability scores; S13. Collect interest preference data through the preference analysis engine, including subject orientation and resource type preference; S14. Integrate the learning behavior data, ability level data, and interest preference data into a learner feature matrix.

[0011] Preferably, the S2 includes: S21, performing missing value filling and outlier removal processing on the learner feature matrix; S22, using principal component analysis to reduce the dimension of the processed feature matrix; S23, normalize the reduced-dimensional data into learner feature vectors through feature scaling; .

[0012] Preferably, the S3 includes: S31. Extract knowledge point metadata from the teaching resource library and construct a knowledge point node set: Knowledge point node collection; No. Knowledge point nodes; Total number of knowledge points; S32. Define node association relationships, including prerequisite relationships and difficulty progression

[0013] S33. Configure knowledge attribute parameters for each knowledge point node, including cognitive complexity and learning weights ; S34. Construct the node set, association relationship and attribute parameters into a knowledge graph: Prerequisite relation set; A set of difficulty progression relationships.

[0014] Preferably, the S4 includes: S41: Input the learner feature vector V into the demand prediction model and output the confidence level of the target knowledge point

[0015] S42. Generate a set of weak knowledge points based on the comparison of ability level data and knowledge point mastery threshold ; S43, according to Greater than 0.7 to filter the target knowledge point set .

[0016] S44. Establish learning needs analysis data: Analyze data for learning needs, is the target knowledge point set, A collection of weak knowledge points.

[0017] Preferably, the S5 includes: S51. Define the learning path generation rules: Rule 1: Weak knowledge points take precedence over target knowledge points; Rule 2: Satisfaction The knowledge points of the relationship are arranged in topological order; S52, based on And the generation rules build the initial path: S53, for Configuring learning resource mapping relationships .

[0018] Preferably, the S62 includes: S62.1. Initialize the Osprey population location: in represents the location of the osprey population, and are the lower and upper bounds of the position, is a random number in the interval [0,1]; S62.2. Update location during the exploration phase: μ is the current optimal solution, is the attack constant; S62.3. Update location during development phase: t is the number of iterations; S62.4. Through the fitness function: For knowledge mastery, To save study time, is cognitive load, is the weight coefficient; S62 5. After iteration until convergence, output the optimal path code.

[0019] Preferably, the S7 includes: S71. Real-time collection of learning feedback data, including knowledge point relearning rate and progress deviation value; S72, when the progress deviation value δ is greater than 15%, triggering path adjustment; S73. Update the learner feature vector based on feedback data ; S74, according to Re-execute S4-S6 to generate the optimized path .

[0020] Preferably, the S74 includes: S74 1. Construct feedback adjustment matrix: is the progress deviation value, is the knowledge point relearning rate; S74 2. Update knowledge attribute parameters through matrix transformation: is the original cognitive complexity vector.

[0021] Preferably, the data acquisition and processing module includes: Behavior collection unit, which collects learning behavior data through terminal devices; Ability assessment unit, generating a capability level assessment matrix; Preference analysis unit, extracting interest preference features; Knowledge graph building blocks include: Node generation unit, creating a knowledge point topology network; Relationship configuration unit, defining prerequisite relationships and difficulty relationships; Path optimization engine, including: Rule base, which stores path generation constraints; Intelligent optimization unit, which executes the Osprey optimization algorithm; Dynamic adjustment module, including: Feedback analysis unit, processing learning progress deviation data; Path reconstruction unit, updates the learning path sequence in real time.

[0022] (3) Beneficial effects Compared with the existing technology, the present invention provides a method and system for recommending personalized learning paths based on big data, which has the following beneficial effects: 1. Dynamic adaptability and cognitive load optimization: By continuously collecting learning feedback data, the system can dynamically trigger the reconstruction of the learning path, improve the path matching degree in the later stages of the learning process, set cognitive load constraints in the path optimization objective function, avoid the continuous arrangement of highly complex knowledge points, ensure that the learning process conforms to the laws of cognitive psychology, and dynamically update knowledge attribute parameters through the matrix transformation mechanism to keep the learning path in the optimal state at all times.

[0023] 2. Multi-objective collaborative optimization: By innovatively integrating the three core goals of knowledge mastery, learning time, and cognitive load, we break through the limitations of traditional single-dimensional optimization, improve the bionic optimization algorithm, conduct efficient global search in complex knowledge point networks, improve the quality of path generation, and support flexible adjustment of target weights according to different learning scenarios to meet differentiated learning needs.

[0024] 3. Deep application of knowledge graph: By constructing a knowledge graph containing multiple teaching relationships and scientific attribute parameters, and by deeply mining the implicit connections between knowledge points, it supports the construction of interdisciplinary learning paths, and the sorting algorithm based on graph characteristics greatly improves the efficiency of path planning.

[0025] 4.Technology integration innovation: By establishing a complete closed loop of "collection-analysis-optimization-feedback", the continuous evolution of learning paths is achieved, and the educational theoretical model is deeply combined with the intelligent optimization algorithm to form a solution with educational significance. The optimized architecture supports the rapid path generation and adjustment of large-scale knowledge point networks. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a schematic diagram of the overall system architecture of the present invention.

[0027] Figure 2 Flow chart of the method of the present invention. DETAILED DESCRIPTION

[0028] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0029] See also Figure 1 A method and system for recommending personalized learning paths based on big data, the method comprising the following steps: S1. Collect learners' multi-dimensional learning characteristic data, including learning behavior data, ability level data, and interest preference data; S2. Preprocessing and feature extraction are performed on the multi-dimensional learning feature data to generate a learner feature vector; S3. Construct a subject knowledge graph, which includes knowledge point nodes, inter-node relationships, and knowledge attribute parameters; S4. Performing a learning needs analysis based on the learner's feature vector and knowledge graph to generate a target knowledge point set and a weak knowledge point set; S5. Generate an initial learning path based on the target knowledge point set and the weak knowledge point set in combination with preset learning rules; S6. Dynamically optimize the initial learning path using an intelligent optimization algorithm to generate a personalized learning path; S7. Adjusting the personalized learning path in real time based on the learning feedback data, and outputting an optimized learning path; Wherein, the S6 includes: S61, establishing a path optimization objective function, wherein the function includes parameters of knowledge mastery, learning time, and cognitive load weight; S62. Searching for the optimal solution of the objective function in the path solution space by improving the Osprey optimization algorithm; S63. Map the optimal solution into a knowledge point sequence to generate a personalized learning path.

[0030] S1 specifically includes: S11. Collect learning behavior data through learning terminals, including learning time, answer accuracy, and resource access frequency; S12. Collect ability level data through the ability assessment module, including knowledge mastery and cognitive ability scores; S13. Collect interest preference data through the preference analysis engine, including subject orientation and resource type preference; S14. Integrate the learning behavior data, ability level data, and interest preference data into a learner feature matrix.

[0031] S2 includes: S21, performing missing value filling and outlier removal processing on the learner feature matrix; S22, using principal component analysis to reduce the dimension of the processed feature matrix; S23. Standardize the reduced-dimensional data into learner feature vectors through feature scaling: S3 includes: S31. Extract knowledge point metadata from the teaching resource library and construct a knowledge point node set: Knowledge point node collection; No. Knowledge point nodes; Total number of knowledge points; S32. Define node association relationships, including prerequisite relationships and difficulty progression

[0032] S33. Configure knowledge attribute parameters for each knowledge point node, including cognitive complexity and learning weights ; S34. Construct the node set, association relationship and attribute parameters into a knowledge graph: Prerequisite relation set; A set of difficulty progression relationships.

[0033] S4 includes: S41: Input the learner feature vector V into the demand prediction model and output the confidence level of the target knowledge point

[0034] S42. Generate a set of weak knowledge points based on the comparison of ability level data and knowledge point mastery threshold ; S43, according to Greater than 0.7 to filter the target knowledge point set .

[0035] S44. Establish learning needs analysis data: Analyze data for learning needs, is the target knowledge point set, A collection of weak knowledge points.

[0036] S5 includes: S51. Define the learning path generation rules: Rule 1: Weak knowledge points take precedence over target knowledge points; Rule 2: Satisfaction The knowledge points of the relationship are arranged in topological order; S52, based on And the generation rules build the initial path: S53, for Configuring learning resource mapping relationships .

[0037] S62 includes: S62.1. Initialize the Osprey population location: in represents the location of the osprey population, and are the lower and upper bounds of the position, is a random number in the interval [0,1]; S62.2. Update location during the exploration phase: μ is the current optimal solution, is the attack constant; S62.3. Update location during development phase: t is the number of iterations; S62.4. Through the fitness function: For knowledge mastery, To save study time, is cognitive load, is the weight coefficient; S62 5. After iteration until convergence, output the optimal path code.

[0038] S7 includes: S71. Real-time collection of learning feedback data, including knowledge point relearning rate and progress deviation value; S72, when the progress deviation value δ is greater than 15%, triggering path adjustment; S73. Update the learner feature vector based on feedback data ; S74, according to Re-execute S4-S6 to generate the optimized path .

[0039] S74 includes: S74 1. Construct feedback adjustment matrix: is the progress deviation value, is the knowledge point relearning rate; S74 2. Update knowledge attribute parameters through matrix transformation: is the original cognitive complexity vector.

[0040] Data acquisition and processing module, including: Behavior collection unit, which collects learning behavior data through terminal devices; Ability assessment unit, generating a capability level assessment matrix; Preference analysis unit, extracting interest preference features; Knowledge graph building blocks include: Node generation unit, creating a knowledge point topology network; Relationship configuration unit, defining prerequisite relationships and difficulty relationships; Path optimization engine, including: Rule base, which stores path generation constraints; Intelligent optimization unit, which executes the Osprey optimization algorithm; Dynamic adjustment module, including: Feedback analysis unit, processing learning progress deviation data; Path reconstruction unit, updates the learning path sequence in real time.

[0041] Example 1: Implementation of a personalized learning path recommendation method Step 1: Multi-dimensional data collection and feature processing 1. Collect three types of core data through the intelligent learning terminal: Learning behavior data: records user interaction behavior characteristics; Ability level data: obtained through diagnostic assessment; Interest preference data: Analyze resource ratings and reviews; 2. Data preprocessing process: Perform integrity check on original data; Apply interpolation methods to fill missing values; Perform normalization transformation to generate feature vectors; Step 2: Knowledge graph construction and relationship definition 1. Create a knowledge point network using mathematics as an example: Define basic knowledge point nodes: algebraic expressions; Establish advanced knowledge nodes: functions and equations; Set up high-level knowledge point nodes: Basics of Calculus; 2. Configure the teaching relationship between nodes: Prerequisite relations: algebraic expressions → functions and equations; Difficulty progression: Functions and equations → Basics of calculus; 3. Set knowledge attribute parameters: Cognitive complexity parameters; Learning weight parameters; Step 3: Learning needs analysis and path generation 1. Target knowledge point identification: Input feature vector to prediction model; Output knowledge point confidence score; Screening high-confidence target knowledge points; 2. Weak point detection: Compare the mastery of each knowledge point with the threshold; Mark the knowledge points that do not meet the requirements; Generate a set of weak knowledge points; 3. Initial path construction: Applying teaching rule engines; Generate base learning sequences; Step 4: Path optimization and dynamic adjustment 1. Intelligent optimization process: Initialize the optimization algorithm population; Perform a two-phase position update; Evaluate paths based on multi-objective functions; 2. Learning process monitoring: Collect learning progress data in real time; Calculate schedule variance indicators; 3. Path reconstruction mechanism: When the deviation exceeds the threshold, the adjustment is initiated; Update the learner feature vector; Regenerate the optimized path; Example 2: System Hardware Implementation 1. Module 1: Data Acquisition and Processing System Hardware configuration plan: Front-end acquisition equipment: equipped with multimodal sensors; Edge computing nodes: deploy lightweight processing units; Cloud storage cluster: distributed data warehouse; 2. Data processing flow: The behavioral acquisition unit captures raw data; The capability assessment unit constructs a multidimensional model; The preference analysis unit generates a feature vector; Module 2: Knowledge Graph Engine 1. System architecture implementation: Storage layer: graph database management node; Computational layer: rule engine processes constraints; Interface layer: provides query services; 2. Typical graph operations: Knowledge point node creation and configuration; Relationship network construction and verification; Discovery of interdisciplinary connection pathways; Module 3: Path Optimization Core 1. Algorithm deployment plan: The rule base integrates expert experience; The optimization unit executes intelligent algorithms; Parallel computing architecture accelerates processing; 2. Dynamically adjust components: Feedback analysis unit monitoring indicators; The path reconstruction unit performs optimization; Module 4: Learning Path Delivery 1. Path presentation format: Visualize learning timeline; Knowledge point dependency diagram; Mobile task reminders; 2. User interaction function: Learning mode switching mechanism; Node order adjustment interface.

[0042] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply the existence of any such actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.

[0043] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A personalized learning path recommendation method based on big data, characterized by: The method comprises the following steps: S1. Collect learners' multi-dimensional learning characteristic data, including learning behavior data, ability level data, and interest preference data; S2. Preprocessing and feature extraction are performed on the multi-dimensional learning feature data to generate a learner feature vector; S3. Construct a subject knowledge graph, which includes knowledge point nodes, inter-node relationships, and knowledge attribute parameters; S4. Performing a learning needs analysis based on the learner's feature vector and knowledge graph to generate a target knowledge point set and a weak knowledge point set; S5. Generate an initial learning path based on the target knowledge point set and the weak knowledge point set in combination with preset learning rules; S6. Dynamically optimize the initial learning path using an intelligent optimization algorithm to generate a personalized learning path; S7. Adjusting the personalized learning path in real time based on the learning feedback data, and outputting an optimized learning path; Wherein, the S6 includes: S61, establishing a path optimization objective function, wherein the function includes parameters of knowledge mastery, learning time, and cognitive load weight; S62. Searching for the optimal solution of the objective function in the path solution space by improving the Osprey optimization algorithm; S63. Map the optimal solution into a knowledge point sequence to generate a personalized learning path.

2. The method for recommending personalized learning paths based on big data according to claim 1, characterized in that: Said S1 specifically includes: S11. Collect learning behavior data through learning terminals, including learning time, answer accuracy, and resource access frequency; S12. Collect ability level data through the ability assessment module, including knowledge mastery and cognitive ability scores; S13. Collect interest preference data through the preference analysis engine, including subject orientation and resource type preference; S14. Integrate the learning behavior data, ability level data, and interest preference data into a learner feature matrix.

3. The method and system for recommending personalized learning paths based on big data according to claim 1, characterized in that: The S2 includes: S21, performing missing value filling and outlier removal processing on the learner feature matrix; S22, using principal component analysis to reduce the dimension of the processed feature matrix; S23. Standardize the reduced-dimensional data into learner feature vectors through feature scaling: 。 4. The method for recommending personalized learning paths based on big data according to claim 1, characterized in that: The S3 includes: S31. Extract knowledge point metadata from the teaching resource library and construct a knowledge point node set: Knowledge point node collection; No. Knowledge point nodes; Total number of knowledge points; S32. Define node association relationships, including prerequisite relationships and difficulty progression ; S33. Configure knowledge attribute parameters for each knowledge point node, including cognitive complexity and learning weights ; S34. Construct the node set, association relationship and attribute parameters into a knowledge graph: Prerequisite relation set; A set of difficulty progression relationships.

5. The method for recommending personalized learning paths based on big data according to claim 1, characterized in that: The S4 includes: S41: Input the learner feature vector V into the demand prediction model and output the confidence level of the target knowledge point ; S42. Generate a set of weak knowledge points based on the comparison of ability level data and knowledge point mastery threshold ; S43, according to Greater than 0.7 to filter the target knowledge point set ; S44. Establish learning needs analysis data: Analyze data for learning needs, is the target knowledge point set, A collection of weak knowledge points.

6. The method for recommending personalized learning paths based on big data according to claim 1, characterized in that: The S5 includes: S51. Define the learning path generation rules: Rule 1: Weak knowledge points take precedence over target knowledge points; Rule 2: Satisfaction The knowledge points of the relationship are arranged in topological order; S52, based on And the generation rules build the initial path: S53, for Configuring learning resource mapping relationships .

7. The method for recommending personalized learning paths based on big data according to claim 1, characterized in that: The S62 includes: S62.

1. Initialize the Osprey population location: in represents the location of the osprey population, and are the lower and upper bounds of the position, is a random number in the interval [0,1]; S62.

2. Update location during the exploration phase: μ is the current optimal solution, is the attack constant; S62.

3. Update location during development phase: t is the number of iterations; S62.

4. Through the fitness function: For knowledge mastery, To save study time, is cognitive load, is the weight coefficient; S62 5. After iteration until convergence, output the optimal path code.

8. The method for recommending personalized learning paths based on big data according to claim 1, characterized in that: The S7 includes: S71. Real-time collection of learning feedback data, including knowledge point relearning rate and progress deviation value; S72, when the progress deviation value δ is greater than 15%, triggering path adjustment; S73. Update the learner feature vector based on feedback data ; S74, according to Re-execute S4-S6 to generate the optimized path .

9. The method for recommending personalized learning paths based on big data according to claim 1, characterized in that: The S74 includes: S74 1. Construct feedback adjustment matrix: is the progress deviation value, is the knowledge point relearning rate; S74 2. Update knowledge attribute parameters through matrix transformation: is the original cognitive complexity vector.

10. The big data-based personalized learning path recommendation system according to claim 1, used to implement the methods according to claims 1-9, characterized in that: Data acquisition and processing module, including: Behavior collection unit, which collects learning behavior data through terminal devices; Ability assessment unit, generating a capability level assessment matrix; Preference analysis unit, extracting interest preference features; Knowledge graph building blocks include: Node generation unit, creating a knowledge point topology network; Relationship configuration unit, defining prerequisite relationships and difficulty relationships; Path optimization engine, including: Rule base, which stores path generation constraints; Intelligent optimization unit, which executes the Osprey optimization algorithm; Dynamic adjustment module, including: Feedback analysis unit, processing learning progress deviation data; Path reconstruction unit, updates the learning path sequence in real time.

Images