Knowledge point association exercise recommendation method and device, equipment and storage medium

Abstract

The application discloses a knowledge point association exercise recommendation method and device, equipment and a storage medium, comprising: determining a target knowledge point pair and a corresponding confusion knowledge point according to a knowledge graph; determining at least one answer accuracy rate of the target knowledge point pair and at least one answer accuracy rate of the confusion knowledge point according to a question database; wherein one answer accuracy rate of the target knowledge point pair and one answer accuracy rate of the confusion knowledge point constitute a sample; stratifying at least one sample of the target knowledge point pair based on a classification model, and the output of the classification model is a probability value; determining an influence effect value of a preceding knowledge point pair current knowledge point in each layer sample according to the answer accuracy rate of the target knowledge point pair, and the target knowledge point pair comprises the preceding knowledge point and the current knowledge point; and determining whether to recommend the exercise associated with the preceding knowledge point according to the influence effect value and a recommendation threshold.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to the field of data processing, and in particular to a knowledge point associated exercise recommendation method and device, equipment and storage medium. BACKGROUND

[0002] In the online learning process, the student is usually recommended a knowledge point with weak mastery based on a knowledge graph. A common recommendation method is to first determine the weak knowledge point of the student, and then recommend relevant knowledge point exercises to the student according to the weak knowledge point and the prerequisite knowledge point of the weak knowledge point. However, the influence degree of different prerequisite knowledge points on the current knowledge point is not the same, and the existing method is to give the closeness degree between knowledge points according to the structure of the knowledge graph (for example, the distance between nodes, the number of associated knowledge points), so that the closeness degree between knowledge points is determined under the premise of the knowledge graph. In fact, the exercises under the knowledge point are related to the content of the question bank, and the correlation degree between knowledge points is different under different question banks, and the correlation degree between knowledge points will change with the update of the question bank. In addition, for different students, the influence of the same prerequisite knowledge point on the current knowledge point may also be different. Therefore, the existing knowledge point related exercise recommendation method is not conducive to the student's mastery of the current knowledge point. SUMMARY

[0003] The present application provides a knowledge point associated exercise recommendation method, device, equipment and storage medium, which can automatically perform personalized exercise recommendation for different situations of students, and selecting the prerequisite knowledge point related exercises by judging the influence degree is also more conducive to improving the student's mastery of the current knowledge point.

[0004] In a first aspect, the embodiments of the present application provide a knowledge point associated exercise recommendation method, which comprises:

[0005] determining a target knowledge point pair and a confusion knowledge point corresponding to the target knowledge point pair according to a knowledge graph;

[0006] determining at least one answer accuracy rate of the target knowledge point pair and at least one answer accuracy rate of the confusion knowledge point according to a question database;

[0007] wherein one answer accuracy rate of the target knowledge point pair and one answer accuracy rate of the confusion knowledge point constitute a sample;

[0008] layering at least one sample of the target knowledge point pair based on a classification model, and the output of the classification model is a probability value;

[0009] determining an influence effect value of the prerequisite knowledge point on the current knowledge point in each layer sample according to the answer accuracy rate of the target knowledge point pair, and the target knowledge point pair comprises the prerequisite knowledge point and the current knowledge point;

[0010] determine whether to recommend the exercise associated with the pre-knowledge point according to the influence effect value and a recommendation threshold.

[0011] In a second aspect, the embodiments of the present application further provide a knowledge point associated exercise recommendation device, which comprises:

[0012] The determining module is configured to determine a target knowledge point pair and a confusion knowledge point corresponding to the target knowledge point pair according to a knowledge graph.

[0013] The determining module is further configured to determine at least one answer accuracy rate of the target knowledge point pair and at least one answer accuracy rate of the confusion knowledge point according to a question database.

[0014] The at least one answer accuracy rate of the target knowledge point pair and the at least one answer accuracy rate of the confusion knowledge point constitute a sample.

[0015] The hierarchical module is configured to hierarchize the at least one sample of the target knowledge point pair based on a classification model, and an output of the classification model is a probability value.

[0016] The determining module is configured to determine an influence effect value of a pre-knowledge point on a current knowledge point in each hierarchical sample according to an answer accuracy rate of the target knowledge point pair, and the target knowledge point pair comprises the pre-knowledge point and the current knowledge point.

[0017] The recommending module is configured to determine whether to recommend the exercise associated with the pre-knowledge point according to the influence effect value and a recommendation threshold.

[0018] In a third aspect, the embodiments of the present application further provide a computer device, which comprises a memory and a processor, and when the memory stores a computer program and the processor executes the computer program, a knowledge point associated exercise recommendation method provided by the embodiments of the present application is implemented.

[0019] In a fourth aspect, the embodiments of the present application further provide a computer readable storage medium, which stores a computer program, and when the computer program is executed by a processor, a knowledge point associated exercise recommendation method provided by the embodiments of the present application is implemented.

[0020] The embodiment of the present application provides a knowledge point associated exercise recommendation method, device and equipment and storage medium, including: determining a target knowledge point pair and a confused knowledge point corresponding to the target knowledge point pair according to a knowledge graph; determining at least one answer accuracy rate of the target knowledge point pair and at least one answer accuracy rate of the confused knowledge point according to a question database; wherein one answer accuracy rate of the target knowledge point pair and one answer accuracy rate of the confused knowledge point constitute a sample; stratifying at least one sample of the target knowledge point pair based on a classification model, and the output of the classification model is a probability value; determining an influence effect value of a preceding knowledge point pair current knowledge point in each layer sample according to the answer accuracy rate of the target knowledge point pair, and the target knowledge point pair includes the preceding knowledge point and the current knowledge point; and determining whether to recommend the exercise associated with the preceding knowledge point according to the influence effect value and a recommendation threshold. Through the above method, personalized exercise recommendation can be automatically performed according to different situations of students, and selecting the exercise related to the preceding knowledge point according to the influence degree is also beneficial to improving the mastery of the current knowledge point by the students. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 It is a knowledge point associated exercise recommendation method flowchart provided by the embodiment of the present application.

[0022] Figure 2 It is a schematic diagram of a knowledge graph provided by the embodiment of the present application.

[0023] Figure 2a It is another schematic diagram of a knowledge graph provided by the embodiment of the present application.

[0024] Figure 3 It is a knowledge point associated exercise recommendation device structure schematic diagram provided by the embodiment of the present application.

[0025] Figure 4 It is a structure schematic diagram of a computer device provided by the embodiment of the present application. DETAILED DESCRIPTION

[0026] The present application will be further described in detail below in combination with the drawings and embodiments. It can be understood that the specific embodiments described herein are only used to explain the present application, but not to limit the present application. In addition, it should be noted that, in order to facilitate the description, only the parts related to the present application are shown in the drawings, but not all the structures.

[0027] In addition, in the embodiment of the present application, the words such as "optionally" or "exemplarily" are used to represent an example, illustration or description. Any embodiment or design scheme described as "optionally" or "exemplarily" in the embodiment of the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. In fact, the words such as "optionally" or "exemplarily" are used to present the relevant concept in a specific way.

[0028] The solution provided in this application can be applied to exercise recommendation scenarios under interrelated knowledge points. The influence relationship between exercises under each knowledge point is as follows:

[0029] Suppose that the knowledge point "Solving Linear Equations in Two Variables" includes Exercise A: "In the linear equation 5x - 3y = 16, if x and y are opposites, find the values ​​of x and y." The knowledge point "Applications of Linear Equations in Two Variables" includes Exercise B: "A rectangle's length is reduced by 3cm and its width is increased by 2cm, becoming a square. The areas of the two shapes are equal. What are the length and width of the original rectangle?" and Exercise C: "According to the rules of basketball: each game has a winner and a loser, with 2 points for a win and 1 point for a loss. In a middle school basketball league, a team played 12 games, winning x games and losing y games, scoring 20 points. Please set up a system of equations." Based on the content of the questions, Exercise B involves setting up and solving a system of equations, while Exercise C only involves setting up a system of equations. Therefore, exercises A and B are highly correlated, while exercises A and C are less correlated. If exercises similar to B account for a higher proportion of the knowledge point "Application of Linear Equations in Two Variables," then the correlation between the knowledge points "Solving Linear Equations in Two Variables" and "Application of Linear Equations in Two Variables" is higher. Practicing exercises under the knowledge point "Solving Linear Equations in Two Variables" will help in learning exercises under the knowledge point "Application of Linear Equations in Two Variables." However, if exercises C account for a higher proportion of the knowledge point "Application of Linear Equations in Two Variables," then the correlation between the knowledge points "Solving Linear Equations in Two Variables" and "Application of Linear Equations in Two Variables" will decrease. Practicing exercises under the knowledge point "Solving Linear Equations in Two Variables" may not quickly improve the mastery of exercises under the knowledge point "Application of Linear Equations in Two Variables."

[0030] In the aforementioned scenario, the impact of the same prior knowledge point on subsequent knowledge points may differ for students with different characteristics. For example, if a student's spatial reasoning ability is average, practicing the knowledge point of "point coordinates" will help them understand the "properties of rotation." However, if a student has strong spatial reasoning ability, practicing the knowledge point of "point coordinates" is unnecessary. Based on this, this application provides a method for recommending knowledge-point related exercises. This method can recommend various types of exercises to students in a targeted manner based on their different characteristics. Figure 1 As shown, the method may include, but is not limited to, the following steps:

[0031] S101. Determine the target knowledge point pair and the corresponding confused knowledge points based on the knowledge graph.

[0032] The knowledge graph in this embodiment is a directed graph. Based on the pointing relationships between knowledge points in the knowledge graph, the target knowledge point pair can be determined. The target knowledge point pair includes a preceding knowledge point and a current knowledge point; that is, a knowledge point pair can be understood as two adjacent knowledge points with a preceding-following pointing relationship. For example... Figure 2 As shown in the figure, based on the direction of the arrows, it can be determined that "like radicals" is a prerequisite knowledge point for "mixed operations of radicals". "Properties and simplifications of radicals" is a prerequisite knowledge point for both "like radicals" and "mixed operations of radicals". In other words, "like radicals" and "mixed operations of radicals" constitute a knowledge point pair.

[0033] It is understandable that since a current knowledge point may have multiple prerequisite knowledge points, for example, "like radicals" and "properties and simplifications of radicals" are both prerequisite knowledge points for "mixed operations of radicals", when selecting "mixed operations of radicals" as the current knowledge point, multiple target knowledge point pairs can be determined based on this knowledge point.

[0034] In this application's embodiments, the obfuscated knowledge points can be understood as knowledge points that affect the influence between the target knowledge point and its preceding and current knowledge points. For example... Figure 2 As shown, "Properties and Simplifications of Radicals" is a prerequisite for both "Like Radicals" and "Mixed Operations of Radicals". If students who have mastered "Like Radicals" have also mastered "Properties and Simplifications of Radicals", while those who have not mastered "Like Radicals" have not mastered "Properties and Simplifications of Radicals", then when choosing "Like Radicals" and "Mixed Operations of Radicals" as the target knowledge pair, it is necessary to consider not only "Like Radicals" but also the impact of "Properties and Simplifications of Radicals" as a confusing knowledge point on "Like Radicals" and "Mixed Operations of Radicals".

[0035] Similarly, in a vast knowledge graph, a target knowledge point pair may have multiple confused knowledge points. This application embodiment uses a target knowledge point pair and a confused knowledge point as the object to illustrate the method of recommending knowledge point association exercises.

[0036] S102. Based on the question database, determine at least one answer accuracy rate for the target knowledge point and at least one answer accuracy rate for the confused knowledge point.

[0037] The aforementioned question database stores exercises associated with each knowledge point in the knowledge graph. Students can use this database to complete exercises related to each knowledge point. Similarly, the database can also store student identifiers (IDs), with each student's ID corresponding to their answer accuracy rate after completing each question. Since a knowledge point may be associated with multiple exercises, students may generate multiple answer accuracy rates when completing exercises related to that knowledge point. Therefore, the average of the answer accuracy rates for the multiple exercises associated with that knowledge point can be used as the overall answer accuracy rate for that knowledge point.

[0038] Furthermore, multiple students may practice exercises related to the same knowledge point. In this case, a knowledge point (e.g., a confused knowledge point, a current knowledge point, or a prerequisite knowledge point) will have multiple answer accuracy rates. Each answer accuracy rate corresponds to a student ID. Therefore, based on the question database, at least one answer accuracy rate for the target knowledge point pair and one answer accuracy rate for the confused knowledge point can be obtained.

[0039] In this embodiment, the answer accuracy rate of a target knowledge point pair and the answer accuracy rate of a confused knowledge point constitute a sample. The answer accuracy rate of the target knowledge point pair can be understood as the answer accuracy rate of the current knowledge point and the answer accuracy rate of the preceding knowledge point.

[0040] S103. Based on the classification model, stratify at least one sample of the target knowledge point pair.

[0041] The classification model in this embodiment can be a classification model with output probabilities, that is, the output of the model is a probability value. For example, the classification model can be a logistic regression model, a random forest, etc. The accuracy rate of answering the target knowledge point pair and the accuracy rate of answering the confused knowledge point contained in each sample are input into the classification model to obtain the probability value corresponding to the sample.

[0042] Based on the obtained probability values, at least one sample is divided into M layers, with each layer spaced 1 / M apart. That is, [0, 1 / M) is one layer, [1 / M, 2 / M) is another layer, ..., until [(M-1) / M, 1] is the last layer. Furthermore, when stratifying the samples, the sample size within each layer needs to be considered. For example, the number of samples in both the experimental and control groups in each layer is greater than K, where M and K are empirical values, taken as integers greater than 0.

[0043] The experimental group sample can be understood as all samples where the accuracy rate of answering questions on prerequisite knowledge points is greater than or equal to a first threshold. This first threshold is used to set the students' mastery level of the knowledge point; for example, it can be set to 100%, meaning the number of accurate answers to the current knowledge point when the students have fully mastered the prerequisite knowledge point. The control group sample can be understood as all samples where the accuracy rate of answering questions on prerequisite knowledge points is less than or equal to a second threshold. This second threshold is used to set the degree to which students have not mastered the prerequisite knowledge point; for example, if the accuracy rate of answering questions on prerequisite knowledge points is less than 15%, the students are considered not to have mastered the prerequisite knowledge point. Under this premise, the corresponding accuracy rates of all students practicing the current knowledge point and the corresponding accuracy rates of answering questions on confused knowledge points constitute all samples under this condition.

[0044] After stratifying the target knowledge points into corresponding samples, it can be assumed that the confusing knowledge points in each layer are similar, that is, the confusing variables are controlled to the same amount. Therefore, the experimental group and the control group samples in each layer can be compared.

[0045] S104. Determine the impact value of prior knowledge points on current knowledge points in each layer of samples based on the accuracy rate of answering the target knowledge points.

[0046] For example, this step can be implemented by determining the proportion p1 of the number of samples in each layer whose current knowledge point answer accuracy is greater than or equal to the third threshold in the experimental sample and the proportion p2 in the control sample, and determining the effect value of the prior knowledge points on the current knowledge point in each layer of samples based on the difference between p1 and p2.

[0047] For example, assuming the prerequisite knowledge point is labeled X and the current knowledge point is labeled Y, then p(Y=1|X=1) can be understood as the proportion of students in the experimental group whose accuracy rate for answering questions on the current knowledge point is greater than or equal to the third threshold when they have mastered the prerequisite knowledge point, i.e., p1 above. p(Y=1|X=0) can be understood as the proportion of students in the control group whose accuracy rate for answering questions on the current knowledge point is greater than or equal to the third threshold when they have not mastered the prerequisite knowledge point, i.e., p2 above. Therefore, the difference between p1 and p2 is p(Y=1|X=1) - p(Y=1|X=0), which is the effect value of the prerequisite knowledge point on the current knowledge point in each layer of the sample.

[0048] S105. Determine whether to recommend exercises related to prerequisite knowledge points based on the impact effect value and recommendation threshold.

[0049] After obtaining the influence effect value of the preceding knowledge points on the current knowledge point in each layer of samples based on the above step S104, the influence effect values ​​of each layer can be accumulated to obtain the average influence effect value of the preceding knowledge points on the current knowledge point in the target knowledge point pair. That is, let the average influence effect value be 'effect'.

[0050]

[0051] Where, m i Let N represent the number of samples in the i-th layer, N represent the total number of samples, and M represent the total number of layers.

[0052] The larger the value of the calculated effect, the greater the influence of the prior knowledge point X on the current knowledge point Y.

[0053] If the average impact value is greater than or equal to the recommended threshold, it indicates that the prerequisite knowledge points have a significant impact on the current knowledge points. Practicing the prerequisite knowledge points will help students master the current knowledge points, and in this case, exercises related to the prerequisite knowledge points can be recommended to students. Conversely, if the average impact value is less than the recommended threshold, it indicates that the prerequisite knowledge points have a relatively small impact on the current knowledge points. Practicing the prerequisite knowledge points will not be of much help in learning and mastering the current knowledge points, and in this case, exercises related to the prerequisite knowledge points should not be recommended to students.

[0054] It should be noted that the above method describes a single target knowledge point pair. In the case of multiple target knowledge point pairs where a current knowledge point has multiple prerequisite knowledge points, if the average impact effect value of multiple prerequisite knowledge points is greater than the recommendation threshold, the multiple prerequisite knowledge points can be sorted in order of magnitude based on the average impact effect value. Exercises associated with each prerequisite knowledge point can then be recommended in order of the magnitude of the influence of each prerequisite knowledge point on the current knowledge point.

[0055] For example, if the current knowledge point is a, and its prerequisite knowledge points are b, c, and d, if knowledge points b and c are not mastered and the influence of knowledge points b and c on knowledge point a is higher than the set recommendation threshold, then the recommended order of related exercises for these two knowledge points is determined according to the degree of influence of knowledge points b and c on knowledge point a.

[0056] This application provides a method for recommending knowledge point-related exercises, including: determining target knowledge point pairs and corresponding confused knowledge points based on a knowledge graph; determining at least one answer accuracy rate for the target knowledge point pair and at least one answer accuracy rate for the confused knowledge point based on a question database; wherein, one answer accuracy rate for the target knowledge point pair and one answer accuracy rate for the confused knowledge point constitute a sample; stratifying at least one sample of the target knowledge point pair based on a classification model, with the output of the classification model being a probability value; determining the influence effect value of prior knowledge points on the current knowledge point in each layer of samples based on the answer accuracy rate of the target knowledge point pair, wherein the target knowledge point pair includes prior knowledge points and the current knowledge point; and determining whether to recommend exercises associated with prior knowledge points based on the influence effect value and a recommendation threshold. This method can automatically provide personalized exercise recommendations based on different student situations, and selecting exercises related to prior knowledge points by judging the degree of influence is more conducive to improving students' mastery of the current knowledge point.

[0057] In one example, the implementation of determining the obfuscated knowledge point corresponding to the target knowledge point in step S101 above may include: if the set of nodes in the knowledge graph satisfies the first condition and the second condition, the node elements contained in the smallest set are determined as the obfuscated knowledge points corresponding to the target knowledge point pair.

[0058] The first condition mentioned above includes that the node elements in the minimum set are not descendant nodes of the preceding knowledge point. The second condition includes that the node elements in the minimum set block all backdoor paths between the preceding knowledge point and the current knowledge point. A backdoor path is a path that connects to the preceding knowledge point and points to the preceding knowledge point among all the connection paths between the preceding knowledge point and the current knowledge point.

[0059] like Figure 2As shown, let the prerequisite knowledge point be "like radicals". The knowledge point "Properties and simplifications of radicals" is not a descendant node of the prerequisite knowledge point "like radicals". Here, a descendant node can be understood as a child node, which satisfies the first condition mentioned above. There are two connection paths between the prerequisite knowledge point "like radicals" and the current knowledge point "mixed operations of radicals": one is "like radicals" — "mixed operations of radicals", and the other is "like radicals" — "Properties and simplifications of radicals" — "mixed operations of radicals". It should be noted that the connection path here only indicates that there is a path connection relationship between the two knowledge points, and does not consider the direction of the connection. Since the path in the first path connected to the prerequisite knowledge point "like radicals" does not point to the prerequisite knowledge point "like radicals", this path is not determined to be a backdoor path. In the second path, the path connected to the prerequisite knowledge point "like radicals", namely the path "like radicals" - "properties and simplifications of radicals", points to the prerequisite knowledge point "like radicals". Therefore, the path "like radicals" - "properties and simplifications of radicals" - "mixed operations of radicals" is identified as the backdoor path.

[0060] Furthermore, the element "Properties and Simplifications of Quadratic Radicals" in the set {"Properties and Simplifications of Quadratic Radicals"} blocks the aforementioned backdoor path between the prerequisite knowledge point "Similar Quadratic Radicals" and the current knowledge point "Mixed Operations of Quadratic Radicals", thus satisfying the second condition. Therefore, the knowledge point "Properties and Simplifications of Quadratic Radicals" that simultaneously satisfies the first and second conditions is identified as a confusing knowledge point.

[0061] like Figure 2a In the knowledge graph shown, let the preceding knowledge point be X and the current knowledge point be Y. Since there are three sets {W}, {Z} and {W, Z} that satisfy the first and second conditions above, we select the smallest set {W} and {Z} from them, and determine the elements W and Z in the smallest set as the two confused knowledge points of the knowledge point pair (X, Y).

[0062] That is, there may be multiple sets of knowledge points in the knowledge graph that satisfy the first and second conditions mentioned above. Therefore, the smallest set is selected as the set of confusing knowledge points, and the elements in this set are used as confusing knowledge points.

[0063] It should be noted that, in the embodiments of this application, those skilled in the art can also add confounding variables according to actual needs. For example, they can simultaneously consider confounding knowledge points and characteristics such as student gender and students' spatial imagination ability. Accordingly, methods such as Double Machine Learning (DML) and causal forest can be used to determine the impact of prior knowledge points on current knowledge points under different confounding variables. The embodiments of this application do not limit this.

[0064] Figure 3 A schematic diagram of a recommended device for knowledge point-related exercises provided in this application embodiment is shown below. Figure 3 As shown, the device includes: a determination module 301, a layering module 302, and a recommendation module 303;

[0065] The determination module is used to identify target knowledge point pairs and corresponding confused knowledge points based on the knowledge graph.

[0066] The determination module is also used to determine, based on the question database, at least one answer accuracy rate for the target knowledge point pair and at least one answer accuracy rate for the confused knowledge point;

[0067] One sample consists of the accuracy rate of answering a question for the target knowledge point and the accuracy rate of answering a question for the confused knowledge point.

[0068] The stratification module is used to stratify at least one sample of the target knowledge point pair based on the classification model, and the output of the classification model is a probability value.

[0069] The determination module is used to determine the impact effect value of the preceding knowledge points on the current knowledge point in each layer of samples based on the answer accuracy of the target knowledge point pair. The target knowledge point pair includes the preceding knowledge point and the current knowledge point.

[0070] The recommendation module is used to determine whether to recommend exercises related to prerequisite knowledge points based on the impact effect value and recommendation threshold.

[0071] In one example, a determination module is used to determine the node elements contained in the minimum set as the confused knowledge points corresponding to the target knowledge point pair, provided that the set of nodes in the knowledge graph satisfies the first and second conditions.

[0072] The knowledge graph is a directed graph. The first condition includes that the node elements in the minimum set are not descendant nodes of the previous knowledge point. The second condition includes that the node elements in the minimum set block all backdoor paths between the previous knowledge point and the current knowledge point. The backdoor path is the path connected to the previous knowledge point that points to the previous knowledge point among all the connection paths between the previous knowledge point and the current knowledge point.

[0073] For example, the stratification module is used to input the answer accuracy of the target knowledge point pair and the answer accuracy of the confused knowledge point contained in each sample into the classification model, and obtain the probability value of the sample based on the classification model; and to divide at least one sample into M layers according to the stratification value M and the number of samples in the layer, where M is an integer greater than 0.

[0074] The above-mentioned conditions for the number of samples within a layer include: the number of experimental samples and the number of control samples in each layer are both greater than K, where K is an integer greater than 0;

[0075] The experimental sample size includes all samples where the accuracy rate of answering questions on prerequisite knowledge points is greater than or equal to the first threshold.

[0076] The control sample size includes all samples where the accuracy rate of answering questions on prerequisite knowledge points is less than or equal to the second threshold.

[0077] In one example, a determination module is used to determine the proportion p1 of the number of samples in each layer where the accuracy rate of answering the current knowledge point is greater than or equal to the third threshold in the experimental sample size and the proportion p2 in the control sample size; and to determine the effect value of the prior knowledge points on the current knowledge point in each layer of samples based on the difference between p1 and p2.

[0078] In one example, the recommendation module is used to accumulate the influence effect values ​​of the preceding knowledge points on the current knowledge point in each layer of samples to obtain the average influence effect value of the preceding knowledge points on the current knowledge point in the target knowledge point pair; if the average influence effect value is greater than or equal to the recommendation threshold, exercises associated with the preceding knowledge points are recommended; or, if the average influence effect value is less than the recommendation threshold, exercises associated with the preceding knowledge points are not recommended.

[0079] The knowledge point association exercise recommendation device provided in this application embodiment can execute this application. Figure 1 The method for recommending knowledge-point related exercises provided in the embodiments has corresponding functional units for execution and beneficial effects.

[0080] Figure 4 This application provides a schematic diagram of the structure of a computer device, as shown in the embodiment of the present application. Figure 4 As shown, the computer device includes a processor 401, a memory 402, an input device 403, an output device 404, and an interface-based event triggering module 405; the number of processors 401 in the computer device can be one or more. Figure 4 Taking a processor 401 as an example; the processor 401, memory 402, input device 403, and output device 404 in the learning machine can be connected via a bus or other means. Figure 4 Taking the example of a connection between China and Israel via a bus.

[0081] Memory 402, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, as described in the embodiments of this application. Figure 1The program instructions / modules corresponding to the method of recommending knowledge-point-related exercises (e.g., the determination module 301, the hierarchical module 302, and the recommendation module 303 in the knowledge-point-related exercise recommendation device). The processor 401 executes the relevant functions in the computer device by running the software programs, instructions, and modules stored in the memory 402, that is, implementing the above-mentioned method of recommending knowledge-point-related exercises.

[0082] The memory 402 may primarily include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a given function; the data storage area may store data created based on the use of the cloud server, etc. Furthermore, the memory 402 may include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 402 may further include memory remotely located relative to the processor 401, which can be connected to the learning machine / terminal / server via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0083] Input device 403 can be used to receive input digital or character information, and to generate key signal inputs related to user settings and function control of the computer device. Output device 404 may include display devices such as a display screen.

[0084] This application embodiment also provides a storage medium containing computer-executable instructions, which, when executed by a processor of a computer device, are used to perform a method for recommending knowledge-point-associated exercises, the method comprising:

[0085] Identify the target knowledge point pairs and the corresponding confused knowledge points based on the knowledge graph;

[0086] Based on the question database, determine at least one answer accuracy rate for the target knowledge point pair and at least one answer accuracy rate for the confused knowledge point;

[0087] One sample consists of the accuracy rate of answering a question for the target knowledge point and the accuracy rate of answering a question for the confused knowledge point.

[0088] Based on the classification model, at least one sample of the target knowledge point pair is stratified, and the output of the classification model is a probability value;

[0089] The impact of prior knowledge points on current knowledge points in each sample layer is determined based on the accuracy of answering the target knowledge point pairs. The target knowledge point pairs include prior knowledge points and current knowledge points.

[0090] Whether to recommend exercises related to prerequisite knowledge points is determined based on the impact effect value and recommendation threshold.

[0091] Of course, the computer-executable instructions provided in the embodiments of this application are not limited to the methods described above, but can also execute the recommended methods for knowledge point association exercises provided in any embodiment of this application.

[0092] Based on the above description of the implementation methods, those skilled in the art can clearly understand that this application can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0093] It is worth noting that in the embodiments of the above-mentioned knowledge point-related exercise recommendation device, the modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific names of each functional unit are only for easy differentiation and are not used to limit the scope of protection of this application.

[0094] Note that the above are merely preferred embodiments and the technical principles employed in this application. Those skilled in the art will understand that this application is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions can be made without departing from the scope of protection of this application. Therefore, although this application has been described in detail through the above embodiments, this application is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of this application, the scope of which is determined by the scope of the appended claims.

Claims

1. A method for recommending exercises that relate to specific knowledge points, characterized in that, include: Based on the knowledge graph, determine the target knowledge point pairs and the corresponding confused knowledge points of the target knowledge point pairs; Based on the question database, determine at least one answer accuracy rate for the target knowledge point pair and at least one answer accuracy rate for the confused knowledge point; Among them, the accuracy rate of answering a pair of target knowledge points and the accuracy rate of answering a pair of confused knowledge points constitute a sample; At least one sample of the target knowledge point pair is stratified based on a classification model, and the output of the classification model is a probability value. The impact effect value of the preceding knowledge points on the current knowledge point in each layer of samples is determined based on the answer accuracy of the target knowledge point pair. The target knowledge point pair includes the preceding knowledge point and the current knowledge point. Whether to recommend exercises related to the prerequisite knowledge points is determined based on the impact effect value and recommendation threshold. The step of stratifying at least one sample of the target knowledge point pair based on a classification model, wherein the output of the classification model is a probability value, including: Input the answer accuracy of the target knowledge point pair and the answer accuracy of the confused knowledge point contained in each sample into the classification model, and obtain the probability value of the sample based on the classification model; The at least one sample is divided into M layers based on the stratification value M and the number of samples within each layer, with each layer spaced 1 / M apart, where M is an integer greater than 0; the confusing knowledge points in each layer of samples are similar.

2. The method according to claim 1, characterized in that, The step of determining the corresponding confused knowledge points based on the knowledge graph includes: If the set of nodes in the knowledge graph satisfies the first and second conditions, the node elements contained in the minimum set are determined as the confused knowledge points corresponding to the target knowledge point pair. Wherein, the knowledge graph is a directed graph, the first condition includes that the node elements in the minimum set are not descendant nodes of the preceding knowledge point, and the second condition includes that the node elements in the minimum set block all backdoor paths between the preceding knowledge point and the current knowledge point, wherein the backdoor path is the path connected to the preceding knowledge point that points to the preceding knowledge point among all connection paths between the preceding knowledge point and the current knowledge point.

3. The method according to claim 1, characterized in that, The conditions for the number of samples within a stratum include: the number of experimental samples and the number of control samples in each stratum are both greater than K, where K is an integer greater than 0; The experimental sample size includes all samples where the accuracy rate of answering questions on prerequisite knowledge points is greater than or equal to the first threshold. The control sample size includes all samples where the accuracy rate of answering questions on prerequisite knowledge points is less than or equal to the second threshold.

4. The method according to claim 3, characterized in that, Based on the accuracy rate of answering the target knowledge point pairs, the influence effect value of the preceding knowledge points on the current knowledge point in each layer of samples is determined, including: Determine the proportion p1 of the number of samples in each layer whose current knowledge point answer accuracy is greater than or equal to the third threshold in the experimental sample size; Determine the proportion (p2) of the number of samples in each layer whose current knowledge point answer accuracy is greater than or equal to the third threshold in the control sample; The effect value of prior knowledge points on current knowledge points in each layer of samples is determined based on the difference between p1 and p2.

5. The method according to any one of claims 1-4, characterized in that, Determining whether to recommend exercises associated with the prerequisite knowledge points based on the impact effect value and recommendation threshold includes: The influence effect values ​​of the preceding knowledge points on the current knowledge point in each layer of samples are summed to obtain the average influence effect value of the preceding knowledge points on the current knowledge point in the target knowledge point pair. If the average impact value is greater than or equal to the recommendation threshold, the exercises associated with the prerequisite knowledge points are recommended. Alternatively, if the average impact value is less than the recommendation threshold, exercises associated with the prerequisite knowledge points are not recommended.

6. A device for recommending exercises that relate knowledge points, characterized in that, include: The determination module is used to determine the target knowledge point pair and the corresponding confused knowledge point based on the knowledge graph. The determining module is further configured to determine, based on the question database, at least one answer accuracy rate for the target knowledge point pair and at least one answer accuracy rate for the confused knowledge point; Among them, the accuracy rate of answering a pair of target knowledge points and the accuracy rate of answering a pair of confused knowledge points constitute a sample; The stratification module is used to stratify at least one sample of the target knowledge point pair based on a classification model, wherein the output of the classification model is a probability value. The determining module is used to determine the influence effect value of the preceding knowledge points on the current knowledge point in each layer of samples based on the answer accuracy of the target knowledge point pair. The target knowledge point pair includes the preceding knowledge point and the current knowledge point. The recommendation module is used to determine whether to recommend exercises related to the prerequisite knowledge points based on the impact effect value and the recommendation threshold. The hierarchical module is specifically used for: The accuracy rates of answering the target knowledge point pair and the confused knowledge point contained in each sample are input into the classification model, and the probability value of the sample is obtained based on the classification model. The at least one sample is divided into M layers according to the stratification value M and the number of samples in each layer, with each layer spaced 1 / M, where M is an integer greater than 0. The confused knowledge points in each layer of samples are similar.

7. A computer device, comprising a memory and a processor, characterized in that, The memory stores a computer program, and when the processor executes the computer program, it implements the method for recommending knowledge point-related exercises as described in any one of claims 1-5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method for recommending knowledge point association exercises as described in any one of claims 1-5.

Images