A teaching intelligent service method, electronic device and storage medium

Abstract

This invention discloses an intelligent teaching service method, electronic device, and storage medium. The method involves acquiring a set of jump vectors for P target teaching objects, each set containing J jump vectors. Nearest-neighbor clustering is performed on the J jump vectors in the set to generate K jump vector clusters. The Gini coefficients of the J jump vectors in the K jump vector clusters are calculated. The Gini coefficients of the K jump vector clusters are compared with a threshold value to perform structural jump filtering on the K jump vector clusters until S structural jump clusters are obtained. Based on the S structural jump clusters, the jump knowledge points are located. This invention, by constructing jump vectors for target teaching objects on adjacent knowledge points and using nearest-neighbor clustering to classify all jump vectors, achieves clear location of jump knowledge points in the teaching path, providing support for content feedback acquisition and micro-reconstruction of knowledge points.

Description

Technical Field

[0001] This invention relates to the field of teaching services, specifically to an intelligent teaching service method. Background Technology

[0002] With the development of intelligent teaching platforms, personalized teaching driven by behavioral data has become an important trend in teaching systems. Current technologies typically analyze students' learning behaviors, answer records, and knowledge mastery to recommend learning paths and adapt content. However, in actual teaching, the cognitive span between different knowledge points is uneven. Some knowledge points may experience abrupt changes in learning behavior due to gaps in cognitive structure or insufficient knowledge transfer, manifesting as a surge in error rates, abnormally long learning times, or increased reliance on prompts, thus affecting the overall teaching pace and learning outcomes.

[0003] Existing solutions primarily focus on content recommendation and learning assessment, lacking a mechanism to identify "skipped knowledge points." This makes it difficult to pinpoint key knowledge point pairs causing drastic changes in students' learning states within the overall learning path, thus impacting the optimization of teaching content structure and the dynamic adjustment of learning paths. Furthermore, existing methods often rely on single-point indicators to assess individual differences, lacking aggregated behavioral modeling and failing to characterize structural learning obstacles representative of a group. Therefore, this invention provides an intelligent teaching service method. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an intelligent teaching service method that solves the technical problems mentioned in the background by accurately identifying skipped knowledge points.

[0005] In a first aspect, the present invention provides an intelligent teaching service method, comprising the following steps:

[0006] S1. Obtain the jump vector set of P target teaching objects; where the jump vector set contains J jump vectors;

[0007] S2. Perform nearest neighbor clustering on the J jump vectors in the jump vector set to generate K jump vector clusters;

[0008] S3. Calculate the Gini coefficients of J jump vectors in a cluster of K jump vectors;

[0009] S4. Compare the Gini coefficient of the K jump vector clusters with the coefficient threshold;

[0010] If the Gini coefficient is less than the coefficient threshold, it is determined that there are no significant structural jumps in the learning path of the target learner; otherwise, structural jump screening is performed on the K jump vector clusters until S structural jump clusters are obtained.

[0011] S5. Based on S structural skip clusters, locate the skip knowledge points.

[0012] In some specific embodiments, obtaining the jump vector set of P target teaching objects includes:

[0013] S1-1: Obtain the learning mastery degree of M original teaching subjects in N knowledge points in the teaching subject;

[0014] S1-2. Select P target teaching objects from M original teaching objects whose learning mastery of N knowledge points all exceed the mastery threshold.

[0015] S1-3. For each target teaching object, obtain the behavioral performance vector sequence of the target teaching object on N knowledge points; wherein, the behavioral performance vector sequence represents the ordered sequence of behavioral performance vectors of the target teaching object on N knowledge points;

[0016] S1-4. Calculate the state distance between adjacent behavior performance vectors in the behavior performance vector sequence;

[0017] S1-5. If the state distance exceeds the distance threshold, the corresponding adjacent behavior vectors are concatenated into a jump vector.

[0018] S1-6. Traverse the P target teaching objects and repeatedly obtain the jump vectors until a jump vector set containing J jump vectors is generated.

[0019] In some specific embodiments, obtaining the learning mastery degree of M original students in a teaching subject on N knowledge points includes:

[0020] S1-1-1 Obtain several behavioral performance parameters of the original teaching object at each knowledge point;

[0021] S1-1-2. Normalize several behavioral performance parameters to obtain several behavioral performance characteristics of the original teaching object at each knowledge point;

[0022] S1-1-3. Weighted summation of several behavioral characteristics to generate the original learning mastery of the students at each knowledge point;

[0023] S1-1-4. Repeatedly obtain the learning mastery of the original teaching subjects on N knowledge points until the learning mastery of M original teaching subjects on N knowledge points is obtained.

[0024] In some specific embodiments, obtaining the behavioral performance vector sequence of the target learning object on N knowledge points includes:

[0025] S1-3-1. Based on the behavioral characteristics of the original teaching object at each knowledge point, determine the behavioral characteristics of the target teaching object at each knowledge point;

[0026] S1-3-2. Based on the behavioral characteristics of the target teaching object at each knowledge point, construct a behavioral performance vector of the target teaching object at each knowledge point;

[0027] S1-3-3. Obtain the behavioral performance vectors of the target teaching object on N knowledge points, and sort the behavioral performance vectors of the N knowledge points according to the teaching order of the N knowledge points to obtain the behavioral performance vector sequence.

[0028] In some specific embodiments, nearest neighbor clustering is performed on the J jump vectors in the jump vector set to generate K jump vector clusters, including:

[0029] S2-1. Randomly select K jump vectors from the jump vector set as initial cluster centers;

[0030] S2-2. For each jump vector, calculate its Euclidean distance to the K initial cluster centers to obtain the K Euclidean distances for each jump vector.

[0031] S2-3. From the K Euclidean distances of each jump vector, select the smallest Euclidean distance;

[0032] S2-4. Distribute the J jump vectors equally to the initial cluster centers with the minimum Euclidean distance to obtain K initial clusters;

[0033] S2-5. Calculate the centroid of the jump vector in each initial cluster and determine K centroids;

[0034] S2-6. Use K centroids as the cluster centers for the next round;

[0035] S2-7. Calculate the change distance between the next cluster center and the initial cluster center;

[0036] S2-8. If the change distance is greater than the threshold, the jump vector set is repeatedly traversed, and the J jump vectors are assigned to the next round cluster center that is the smallest Euclidean distance from them, until the change distance of the cluster center is less than the threshold.

[0037] S2-9. Define the cluster centers whose change distance is less than the threshold and their assigned jump vectors as the jump vector clusters to generate K jump vector clusters.

[0038] In some specific embodiments, calculating the Gini coefficients of J jump vectors in a cluster of K jump vectors includes:

[0039] S3-1. Obtain the number of jump vectors contained in each jump vector cluster, and the total number of jump vectors in the K jump vector clusters;

[0040] S3-2. Calculate the Gini coefficients of the J jump vectors based on the number of jump vectors contained in each jump vector cluster and the total number of jump vectors in the K jump vector clusters.

[0041] The formula for calculating the Gini coefficient is:

[0042] ;

[0043] Where G is the Gini coefficient of the J jump vectors, K represents the number of jump vector clusters, and J represents the total number of jump vectors in the jump vector set. This represents the number of jump vectors contained in the i-th jump vector cluster.

[0044] In some specific embodiments, structural jump screening is performed on the K jump vector clusters until S structural jump clusters are obtained, including:

[0045] S4-1. Obtain the average Euclidean distance from the jump vector to the cluster center in each jump vector cluster.

[0046] S4-2. Calculate the structural jump index of each jump vector cluster based on the Gini coefficients of the J jump vectors, the number of jump vectors contained in each jump vector cluster, and the average Euclidean distance of each jump vector cluster.

[0047] The formula for calculating the structural jump index is:

[0048] ;

[0049] Among them, the The structural jump index represents the clustering of the i-th jump vector. This indicates that a logarithmic transformation of the Gini coefficient is performed to adjust the intensity of the Gini coefficient's influence and avoid excessively high G values ​​that could cause index skewness. This represents the frequency proportion adjustment term for the i-th jump vector cluster. It represents the average Euclidean distance of the i-th jump vector cluster, which is used to measure the stability within the jump vector cluster; To prevent small constants from being divided by zero, This represents the intra-cluster compactness of the i-th jump vector cluster; a larger value indicates higher consistency among the jump vectors. and These represent the weighting factors for the frequency percentage adjustment term and the intra-cluster compactness adjustment term, respectively.

[0050] S4-3. Obtain the structural jump index of the K jump vector clusters;

[0051] S4-4. Sort the structural jump indices of the K jump vector clusters to generate a structural jump sequence;

[0052] S4-5. Select the top S structural jump clusters from the structural jump sequences.

[0053] In some specific embodiments, skip knowledge points are located based on S structural skip clusters, including:

[0054] S5-1. Extract structural jump vectors from S structural jump clusters;

[0055] S5-2, Obtaining adjacent knowledge points of structural jump vectors;

[0056] S5-3. Define the next knowledge point after the pairing of adjacent knowledge points as the jumping knowledge point.

[0057] S5-4. Obtain feedback from the target learners regarding the skipped knowledge points, and reconstruct the knowledge points based on this feedback.

[0058] Compared with existing technologies, the present invention provides an intelligent teaching service method. This method constructs jump vectors for target learners on adjacent knowledge points and uses nearest-neighbor clustering to classify all jump vectors. This solves the problem of difficulty in uniformly modeling jump performance due to large individual behavioral differences, allowing jump vectors with similar behavioral characteristics to cluster together, providing a structural basis for subsequent identification. Based on the clustering results, a Gini coefficient is introduced to measure the distribution concentration of jump vectors in each cluster. A structural jump index is constructed by combining multiple dimensions such as jump frequency ratio and average distance within the cluster, achieving quantitative screening of jump clusters and effectively improving the accuracy of identifying jump behaviors with group commonalities. By extracting the knowledge point pairs corresponding to jump vectors in the structural jump clusters and defining the latter knowledge point as the jump knowledge point, the method achieves clear positioning of jump knowledge points in the teaching path, providing support for content feedback acquisition and knowledge point micro-reconstruction.

[0059] In a second aspect, the present invention provides an electronic device comprising a memory and a processor, the memory storing at least one computer-executable instruction, the processor being configured to execute the computer-executable instruction, the computer-executable instruction being executed by the processor to implement the teaching intelligent service method described in the first aspect.

[0060] Thirdly, the present invention provides a computer-readable storage medium on which a computer program is stored, wherein the computer program is executed by a processor to implement the teaching intelligent service method described in the first aspect.

[0061] Compared with the prior art, the beneficial effects of the electronic device and its storage medium of the present invention are the same as those of the above-described intelligent teaching service method, so they will not be repeated here. Attached Figure Description

[0062] Figure 1 This is a flowchart illustrating an intelligent teaching service method according to the present invention;

[0063] Figure 2 This is a schematic diagram illustrating the process of generating jump vector clusters according to the present invention;

[0064] Figure 3 This is a schematic diagram of the selection process for the structural skip clusters described in this invention. Detailed Implementation

[0065] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0066] Please see Figures 1 to 3 This invention provides a method for intelligent teaching services, comprising the following steps:

[0067] S1. Obtain the jump vector set of P target teaching objects; where the jump vector set contains J jump vectors;

[0068] It should be noted that the jump vector represents the joint behavior vector constructed when the target learner's behavior differs significantly in the learning tasks of two adjacent knowledge points; specifically, the jump vector is formed by splicing the learner's behavior vectors on the previous and subsequent knowledge points, and is used to characterize the transitional behavior of changes in learning style.

[0069] S2. Perform nearest neighbor clustering on the J jump vectors in the jump vector set to generate K jump vector clusters;

[0070] S3. Calculate the Gini coefficients of J jump vectors in a cluster of K jump vectors;

[0071] S4. Compare the Gini coefficient of the K jump vector clusters with the coefficient threshold;

[0072] If the Gini coefficient is less than the coefficient threshold, it is determined that there are no significant structural jumps in the learning path of the target learner; otherwise, structural jump screening is performed on the K jump vector clusters until S structural jump clusters are obtained.

[0073] S5. Based on S structural skip clusters, locate the skip knowledge points.

[0074] In this embodiment, by constructing jump vectors and using nearest-neighbor clustering and the Gini coefficient, structural jump locations in the learning path of the target learner are identified, improving the ability to model behavioral style mutations during the teaching process. Jump vectors, by concatenating behavioral performance vectors from two adjacent knowledge points, accurately capture behavioral differences in continuous learning tasks. Clustering all jump vectors and introducing the Gini coefficient to measure their distribution concentration facilitates the distinction between sporadic jumps and structural jumps with common characteristics, avoiding the limitations of individual fluctuations interfering with identification results in traditional behavioral analysis. Finally, by comparing the Gini coefficient with a threshold and selecting representative jump clusters, the location of the jump knowledge points is achieved.

[0075] For example, in this embodiment, step S1 further includes:

[0076] S1-1: Obtain the learning mastery degree of M original teaching subjects in N knowledge points in the teaching subject;

[0077] S1-2. Select P target teaching objects from M original teaching objects whose learning mastery of N knowledge points all exceed the mastery threshold.

[0078] Specifically, in this embodiment, the "original teaching object" refers to an individual unit participating in the teaching task, which can be a student, trainee, or other form of learning participant. The "target teaching object" refers to a subset of the original teaching objects that meet certain learning performance conditions (such as the mastery of each knowledge point exceeding a preset threshold).

[0079] S1-3. For each target teaching object, obtain the behavioral performance vector sequence of the target teaching object on N knowledge points; wherein, the behavioral performance vector sequence represents the ordered sequence of behavioral performance vectors of the target teaching object on N knowledge points;

[0080] S1-4. Calculate the state distance between adjacent behavior performance vectors in the behavior performance vector sequence;

[0081] For example, in this embodiment, the adjacent behavior performance vector can be a test behavior feature vector of the target learner on two consecutive knowledge points, including a combination of features such as test scores, completion time, number of error corrections, and number of times prompts are used. Therefore, the state distance can preferably be a weighted Euclidean distance; the weighted Euclidean distance can set corresponding weights according to the degree of influence of different behavioral features on teaching jumps, and has higher interpretability when identifying changes in learning styles.

[0082] S1-5. If the state distance exceeds the distance threshold, the corresponding adjacent behavior vectors are concatenated into a jump vector.

[0083] It should be noted that the jump vector represents the characteristic that the original teaching object has a significant difference in its learning performance in the learning tasks of two adjacent knowledge points, and is used to characterize the switching of learning styles or changes in learning strategies that occur in the process of sequential teaching.

[0084] S1-6. Traverse the P target teaching objects and repeatedly obtain the jump vectors until a jump vector set containing J jump vectors is generated.

[0085] It should be noted that the number of jump vectors for each target learning object may be different, specifically determined by the number of jump vectors in its behavioral performance vector sequence that exceed the distance threshold. Therefore, the resulting J jump vectors are a set of non-equilibrium distributions.

[0086] In this embodiment, high-quality extraction of jumping behavior is achieved by screening teaching subjects with excellent learning mastery and constructing a jump vector set based on their behavioral performance vector sequences. Specifically, a mastery threshold is used to effectively exclude interference from low-quality learning trajectories in jump behavior recognition, and state distance is used to select jump vectors that significantly exceed the threshold, ultimately generating a jump vector set containing J jump vectors.

[0087] For example, step S1-1 further includes:

[0088] S1-1-1 Obtain several behavioral performance parameters of the original teaching object at each knowledge point;

[0089] Among these, several performance parameters include, but are not limited to, test score, test completion time, number of error corrections, and number of prompts used.

[0090] S1-1-2. Normalize several behavioral performance parameters to obtain several behavioral performance characteristics of the original teaching object at each knowledge point;

[0091] S1-1-3. Weighted summation of several behavioral characteristics to generate the original learning mastery of the students at each knowledge point;

[0092] S1-1-4. Repeatedly obtain the learning mastery of the original teaching subjects on N knowledge points until the learning mastery of M original teaching subjects on N knowledge points is obtained.

[0093] In this embodiment, by acquiring multiple behavioral performance parameters and performing normalization and weighting, a learning mastery score is constructed to assess the learning level of the learners. This embodiment can convert raw behavioral data such as test scores, completion time, and number of corrections into a comprehensive mastery score under a unified measurement system, ultimately obtaining the learning mastery score of N knowledge points.

[0094] For example, steps S1-3 further include:

[0095] S1-3-1. Based on the behavioral characteristics of the original teaching object at each knowledge point, determine the behavioral characteristics of the target teaching object at each knowledge point;

[0096] S1-3-2. Based on the behavioral characteristics of the target teaching object at each knowledge point, construct a behavioral performance vector of the target teaching object at each knowledge point;

[0097] S1-3-3. Obtain the behavioral performance vectors of the target teaching object on N knowledge points, and sort the behavioral performance vectors of the N knowledge points according to the teaching order of the N knowledge points to obtain the behavioral performance vector sequence.

[0098] In this embodiment, a structured sequence of behavioral performance vectors is formed by sorting the behavioral performance vectors of the target learner at each knowledge point based on the teaching sequence. This embodiment can preserve the sequential logic of teaching tasks while ensuring the complete expression of features, providing ordered input for distance calculation of skipped behaviors and identification of sequence mutations.

[0099] For example, in this embodiment, step S2 further includes:

[0100] S2-1. Randomly select K jump vectors from the jump vector set as initial cluster centers;

[0101] S2-2. For each jump vector, calculate its Euclidean distance to the K initial cluster centers to obtain the K Euclidean distances for each jump vector.

[0102] S2-3. From the K Euclidean distances of each jump vector, select the smallest Euclidean distance;

[0103] S2-4. Distribute the J jump vectors equally to the initial cluster centers with the minimum Euclidean distance to obtain K initial clusters;

[0104] S2-5. Calculate the centroid of the jump vector in each initial cluster and determine K centroids;

[0105] S2-6. Use K centroids as the cluster centers for the next round;

[0106] S2-7. Calculate the change distance between the next cluster center and the initial cluster center;

[0107] S2-8. If the change distance is greater than the threshold, the jump vector set is repeatedly traversed, and the J jump vectors are assigned to the next round cluster center that is the smallest Euclidean distance from them, until the change distance of the cluster center is less than the threshold.

[0108] S2-9. Define the cluster centers whose change distance is less than the threshold and their assigned jump vectors as the jump vector clusters to generate K jump vector clusters.

[0109] In this embodiment, by performing iterative clustering based on minimum Euclidean distance on the jump vector set, a complete clustering process is completed, from initial cluster center selection, vector assignment, centroid update to convergence determination, ultimately generating K jump vector clusters. This embodiment, while ensuring computational controllability, utilizes multiple rounds of centroid updates to improve the similarity of jump vectors within the same cluster and ensures the distinguishability of the clustering results for different jump behaviors, providing a clear-boundary cluster partitioning basis for structural jump selection.

[0110] For example, in this embodiment, step S3 further includes:

[0111] S3-1. Obtain the number of jump vectors contained in each jump vector cluster, and the total number of jump vectors in the K jump vector clusters;

[0112] S3-2. Calculate the Gini coefficients of the J jump vectors based on the number of jump vectors contained in each jump vector cluster and the total number of jump vectors in the K jump vector clusters.

[0113] The formula for calculating the Gini coefficient is:

[0114] ;

[0115] Where G is the Gini coefficient of the J jump vectors, K represents the number of jump vector clusters, and J represents the total number of jump vectors in the jump vector set. This represents the number of jump vectors contained in the i-th jump vector cluster.

[0116] Specifically, the Gini coefficient, calculated by normalizing and accumulating the pairwise differences in the number of jump vectors within each jump vector cluster, can be used to measure the concentration of jump vector distribution within each cluster. A Gini coefficient close to 0 indicates a relatively uniform distribution of jump vectors across clusters, suggesting that jumping behavior is dispersed and lacks structural characteristics. A Gini coefficient close to 1 indicates that jump vectors are concentrated in a few clusters, suggesting the existence of high-frequency, typical jumping patterns, which can be used to identify structural jump clusters. In other words, the Gini coefficient is used to determine whether jumping behavior is concentrated within clusters. If certain jumping behaviors are repeatedly identified and clustered in a few jumping patterns within the target group, it indicates the presence of common structural jumps in the learning path; conversely, jumping behavior may stem more from individual differences or incidental factors, lacking universal value for optimizing teaching structure.

[0117] In this embodiment, the structural quantification of jumping behavior is achieved by calculating the distribution of jump vectors in each cluster and introducing the Gini coefficient for concentration analysis. The Gini coefficient, as an indicator of the degree of uneven distribution of jump vectors, can identify whether jumping behavior is concentrated in a few clusters, thereby determining whether they share common characteristics within the group.

[0118] For example, in this embodiment, step S4 further includes:

[0119] S4-1. Obtain the average Euclidean distance from the jump vector to the cluster center in each jump vector cluster.

[0120] S4-2. Calculate the structural jump index of each jump vector cluster based on the Gini coefficients of the J jump vectors, the number of jump vectors contained in each jump vector cluster, and the average Euclidean distance of each jump vector cluster.

[0121] The formula for calculating the structural jump index is:

[0122] ;

[0123] Among them, the The structural jump index represents the clustering of the i-th jump vector. This indicates that a logarithmic transformation of the Gini coefficient is performed to adjust the intensity of the Gini coefficient's influence and avoid excessively high G values ​​that could cause index skewness. This represents the frequency proportion adjustment term of the i-th jump vector cluster, reflecting the probability that the jump vector cluster is a "major contributor"; It represents the average Euclidean distance of the i-th jump vector cluster, which is used to measure the stability within the jump vector cluster; To prevent small constants from being divided by zero, This represents the intra-cluster compactness of the i-th jump vector cluster; a larger value indicates higher consistency among the jump vectors. and These represent the weighting factors for the frequency percentage adjustment term and the intra-cluster compactness adjustment term, respectively.

[0124] S4-3. Obtain the structural jump index of the K jump vector clusters;

[0125] S4-4. Sort the structural jump indices of the K jump vector clusters to generate a structural jump sequence;

[0126] S4-5. Select the top S structural jump clusters from the structural jump sequences.

[0127] In this embodiment, a structural jump index is constructed that integrates the Gini coefficient, the mean Euclidean distance within the cluster, and the number of jump vectors to quantify and rank K jump vector clusters. This index logically combines three key factors: the Gini coefficient measures the concentration of jump behavior, reflecting whether the cluster represents a common jumping pattern in the group; the mean Euclidean distance measures the consistency of jump vectors within the cluster, ensuring that the selected clusters have internally stable behavioral characteristics; and the proportion of jump vectors serves as a dominant reference, reflecting the relative influence of the cluster on the overall jump behavior. By integrating these factors, the structural strength of clusters can be determined on a multi-dimensional scale, thereby selecting the top S structural jump clusters and providing a reliable basis for locating jump knowledge points. This method avoids the bias caused by single-factor selection and achieves multi-dimensional discrimination for structural jump identification.

[0128] For example, in this embodiment, step S5 further includes:

[0129] S5-1. Extract structural jump vectors from S structural jump clusters;

[0130] S5-2, Obtaining adjacent knowledge points of structural jump vectors;

[0131] S5-3. Define the next knowledge point after the pairing of adjacent knowledge points as the jumping knowledge point.

[0132] S5-4. Obtain feedback from the target learners regarding the skipped knowledge points, and reconstruct the knowledge points based on this feedback.

[0133] Furthermore, micro-restructuring of knowledge points based on feedback refers to adjusting the teaching content of the skipped knowledge point at the structural level based on the learning behavior characteristics of the target learners and in combination with feedback parameters such as error rate, cue dependence, and correction path. Specifically, this includes:

[0134] The structure or difficulty order of the question options can be rearranged, and transitional exercises or video materials related to the knowledge points before the jump can be inserted. A tiered prompting mechanism or answer guidance logic can be added, or a personalized explanation module can be recommended based on the error cause clustering results. This makes the teaching design of the knowledge points to be jumped more in line with learning behavior, thereby improving its coherence in the teaching sequence.

[0135] This embodiment of an intelligent teaching service method constructs jump vectors for target learners on adjacent knowledge points and categorizes all jump vectors using nearest-neighbor clustering. This solves the problem of inconsistent modeling of jump behavior due to large individual behavioral differences, allowing jump vectors with similar behavioral characteristics to cluster together, providing a structural basis for subsequent identification. Based on the clustering results, a Gini coefficient is introduced to measure the distribution concentration of jump vectors in each cluster. A structural jump index is constructed by combining multiple dimensions such as jump frequency ratio and average distance within the cluster, enabling quantitative screening of jump clusters and effectively improving the accuracy of identifying jump behaviors with group commonalities. By extracting the knowledge point pairs corresponding to jump vectors in the structural jump clusters and defining the latter knowledge point as the jump knowledge point, the precise location of jump knowledge points in the teaching path is achieved.

[0136] This invention also provides an electronic device, which includes a memory and a processor. The memory stores at least one computer-executable instruction, and the processor is configured to run the computer-executable instruction. The computer-executable instruction is run by the processor to implement the above-described intelligent teaching service method.

[0137] The electronic device may include a processor, a communications interface, memory, and a communication bus, wherein the processor, communications interface, and memory communicate with each other via the communication bus. The processor can invoke logical instructions stored in the memory to execute a teaching intelligent service method disclosed in this embodiment.

[0138] Furthermore, the logical instructions in the aforementioned memory can be implemented as software functional units and sold or used as independent products, and can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0139] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means.

[0140] In the embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; multiple units or components may be combined or integrated into another system, or some features may be omitted or not performed. Furthermore, the mutual couplings or direct couplings or communication connections shown or discussed may be indirect couplings or communication connections between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.

[0141] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A teaching intelligent service method, characterized in that, include: S1. Obtain the jump vector set of P target teaching objects; where the jump vector set contains J jump vectors; S2. Perform nearest neighbor clustering on the J jump vectors in the jump vector set to generate K jump vector clusters; S3. Calculate the Gini coefficients of J jump vectors in a cluster of K jump vectors; S4. Compare the Gini coefficient of the K jump vector clusters with the coefficient threshold; If the Gini coefficient is less than the coefficient threshold, it is determined that there are no significant structural jumps in the learning path of the target learner; otherwise, structural jump screening is performed on the K jump vector clusters until S structural jump clusters are obtained. S5. Based on S structural skip clusters, locate the skip knowledge points; The process of obtaining the jump vector set for P target teaching objects includes: S1-1: Obtain the learning mastery degree of M original teaching subjects in N knowledge points in the teaching subject; S1-2. Select P target teaching objects from M original teaching objects whose learning mastery of N knowledge points all exceed the mastery threshold. S1-3. For each target teaching object, obtain the behavioral performance vector sequence of the target teaching object on N knowledge points; wherein, the behavioral performance vector sequence represents the ordered sequence of behavioral performance vectors of the target teaching object on N knowledge points; S1-4. Calculate the state distance between adjacent behavior performance vectors in the behavior performance vector sequence; S1-5. If the state distance exceeds the distance threshold, the corresponding adjacent behavior vectors are concatenated into a jump vector. S1-6. Traverse the P target teaching objects and repeatedly obtain the jump vectors until a jump vector set containing J jump vectors is generated. The calculation of the Gini coefficients of J jump vectors in a cluster of K jump vectors includes: S3-1. Obtain the number of jump vectors contained in each jump vector cluster, and the total number of jump vectors in the K jump vector clusters; S3-2. Calculate the Gini coefficients of the J jump vectors based on the number of jump vectors contained in each jump vector cluster and the total number of jump vectors in the K jump vector clusters. The formula for calculating the Gini coefficient is: ; in, Let K be the Gini coefficient of the J jump vectors, K represent the number of jump vector clusters, and J represent the total number of jump vectors in the jump vector set. This represents the number of jump vectors contained in the i-th jump vector cluster; The step of performing structural jump screening on K jump vector clusters until S structural jump clusters are obtained includes: S4-1. Obtain the average Euclidean distance from the jump vector to the cluster center in each jump vector cluster. S4-2. Calculate the structural jump index of each jump vector cluster based on the Gini coefficients of the J jump vectors, the number of jump vectors contained in each jump vector cluster, and the average Euclidean distance of each jump vector cluster. The formula for calculating the structural jump index is: ； in, The structural jump index represents the clustering of the i-th jump vector. This indicates that a logarithmic transformation of the Gini coefficient is performed to adjust the intensity of the Gini coefficient's influence and avoid excessively high G values ​​that could cause index skewness. This represents the frequency proportion adjustment term for the i-th jump vector cluster. It represents the average Euclidean distance of the i-th jump vector cluster, which is used to measure the stability within the jump vector cluster; To prevent small constants from being divided by zero, This represents the intra-cluster compactness of the i-th jump vector cluster; a larger value indicates higher consistency among the jump vectors. and These represent the weighting factors for the frequency percentage adjustment term and the intra-cluster compactness adjustment term, respectively. S4-3. Obtain the structural jump index of the K jump vector clusters; S4-4. Sort the structural jump indices of the K jump vector clusters to generate a structural jump sequence; S4-5. Select the top S structural jump clusters from the structural jump sequences.

2. The intelligent teaching service method according to claim 1, characterized in that, To obtain the learning mastery of M original students in a teaching subject across N knowledge points, including: S1-1-1 Obtain several behavioral performance parameters of the original teaching object at each knowledge point; S1-1-2. Normalize several behavioral performance parameters to obtain several behavioral performance characteristics of the original teaching object at each knowledge point; S1-1-3. Weighted summation of several behavioral characteristics to generate the original learning mastery of the students at each knowledge point; S1-1-4. Repeatedly obtain the learning mastery of the original teaching subjects on N knowledge points until the learning mastery of M original teaching subjects on N knowledge points is obtained.

3. The intelligent teaching service method according to claim 2, characterized in that, Obtain the behavioral performance vector sequence of the target learning object on N knowledge points, including: S1-3-1. Based on the behavioral characteristics of the original teaching object at each knowledge point, determine the behavioral characteristics of the target teaching object at each knowledge point; S1-3-2. Based on the behavioral characteristics of the target teaching object at each knowledge point, construct a behavioral performance vector of the target teaching object at each knowledge point; S1-3-3. Obtain the behavioral performance vectors of the target teaching object on N knowledge points, and sort the behavioral performance vectors of the N knowledge points according to the teaching order of the N knowledge points to obtain the behavioral performance vector sequence.

4. The intelligent teaching service method according to claim 3, characterized in that, Perform nearest neighbor clustering on the J jump vectors in the jump vector set to generate K jump vector clusters, including: S2-1. Randomly select K jump vectors from the jump vector set as initial cluster centers; S2-2. For each jump vector, calculate its Euclidean distance to the K initial cluster centers to obtain the K Euclidean distances for each jump vector. S2-3. From the K Euclidean distances of each jump vector, select the smallest Euclidean distance; S2-4. Distribute the J jump vectors equally to the initial cluster centers with the minimum Euclidean distance to obtain K initial clusters; S2-5. Calculate the centroid of the jump vector in each initial cluster and determine K centroids; S2-6. Use K centroids as the cluster centers for the next round; S2-7. Calculate the change distance between the next cluster center and the initial cluster center; S2-8. If the change distance is greater than the threshold, the jump vector set is repeatedly traversed, and the J jump vectors are assigned to the next round cluster center that is the smallest Euclidean distance from them, until the change distance of the cluster center is less than the threshold. S2-9. Define the cluster centers whose change distance is less than the threshold and their assigned jump vectors as the jump vector clusters to generate K jump vector clusters.

5. The intelligent teaching service method according to claim 4, characterized in that, Based on S structural skip clusters, locate the skip knowledge points, including: S5-1. Extract structural jump vectors from S structural jump clusters; S5-2, Obtaining adjacent knowledge points of structural jump vectors; S5-3. Define the next knowledge point after the pairing of adjacent knowledge points as the jumping knowledge point. S5-4. Obtain feedback from the target learners regarding the skipped knowledge points, and reconstruct the knowledge points based on this feedback.

6. An electronic device, characterized in that, The method includes a memory and a processor, wherein the memory stores at least one computer-executable instruction, and the processor is configured to run the computer-executable instruction, wherein the computer-executable instruction, when run by the processor, implements the teaching intelligent service method according to any one of claims 1 to 5.

7. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the intelligent teaching service method according to any one of claims 1 to 5.

Images