Intelligent question generation method for knowledge point analysis

Abstract

The invention relates to the technical field of question generation, and discloses an intelligent question generation method for knowledge point analysis, and the method comprises the steps: carrying out the knowledge point labeling and entity labeling of mathematical questions, and carrying out the clustering processing of the mathematical questions investigating the same knowledge based on an entity labeling result; performing template extraction on the mathematical question cluster by adopting an association rule template mining mode; and generating variable entity content in the question template and an assignment result of the invariable entity by using a variable entity content generation model, and filling the variable entity content and the assignment result into the question template to form a question. According to the method, mathematical questions which are consistent in investigation question knowledge, similar in related theorem principle and similar in question structure are merged into the same mathematical question cluster by utilizing an entity labeling result, so that entity-based mathematical question clustering is realized; and assigning the immutable entities by using a variable entity content generation model, performing semantic iteration on variable entity contents of the variable entities to obtain the variable entity contents with consistent context semantics, and performing topic generation.

Description

Technical Field

[0001] The present invention relates to the technical field of question generation, and in particular to an intelligent question generation method based on knowledge point analysis. Background Art

[0002] With the rapid development of information technology, the field of education is also undergoing profound changes. Especially in mathematics education, traditional teaching methods are gradually unable to meet the needs of students for personalized learning. In traditional mathematics education, teachers usually rely on textbooks and existing question banks to design questions. This method has many limitations. First, the singleness and fixedness of the questions make it difficult to adapt to the learning progress and abilities of different students; second, teachers often lack in-depth analysis of knowledge points when designing questions, resulting in unsatisfactory learning results. Therefore, there is an urgent need for an intelligent way to generate math questions to make education more personalized and efficient. Summary of the invention

[0003] In view of this, the present invention proposes an intelligent question generation method based on knowledge point analysis, which extracts question templates of different knowledge points by analyzing the frequency of occurrence of principles and theorems in different knowledge points, generates question descriptions in the question templates by natural language processing, and generates mathematics questions that meet the needs of knowledge education through intelligent means to improve learning effects.

[0004] To achieve the above object, the present invention provides a method for generating intelligent questions based on knowledge point analysis, comprising the following steps: S1: Obtain math problems, annotate the math problems with knowledge points and entities, cluster the math problems that test the same knowledge based on the entity annotation results, and obtain multiple math problem clusters that test the same knowledge, wherein the clustering process adopts an entity-based problem structure clustering method.

[0005] S2: Use association rule template mining to extract templates from math question clusters, obtain question templates for math question clusters, label entity variables in question templates, and obtain variable entities and immutable entities in question templates.

[0006] S3: Construct a variable entity content generation model, and use the variable entity content generation model to generate variable entity content in the title template and the assignment results of immutable entities. The variable entity content generation model takes the context immutable entity of the variable entity in the title template as input, and takes the variable entity content and the assignment results of the immutable entity as output, wherein the variable entity content generation model is implemented using a sequence-to-sequence method based on context content.

[0007] S4: Fill the variable entity content and the assignment results of the immutable entity into the question template to form a question, and obtain the question generation results of different knowledge points.

[0008] As a further improvement method of the present invention: Optionally, in step S1, obtaining a math problem and annotating the math problem with knowledge points and entities includes: Get N groups of math problems, and perform knowledge point annotation and entity annotation on each group of math problems. The knowledge point annotation and entity annotation process of the nth group of math problems is as follows: Mark the knowledge tested by the nth group of math questions, where the knowledge tested by the math questions includes geometry, algebra, number theory, mathematical analysis, probability and statistics. .

[0009] Mark the theorems and principles involved in the nth group of math problems, and combine the knowledge marking results, theorem and principle marking results to form the knowledge point marking results of the nth group of math problems : .

[0010] in: Indicates the knowledge point annotation results of the nth group of math problems. represents the knowledge tested by the nth group of math problems, Represents the set of theorems and principles involved in the nth group of mathematical problems.

[0011] Identify the key elements in the nth group of math problems, where the key element types include numbers, variables, geometric shapes, operators, and units, annotate the identified key elements and the types corresponding to the key elements, and obtain the entity annotation results of the nth group of math problems : .

[0012] in: represents the cth key element in the nth set of math problems, Represents key elements type, represents the total number of key elements in the nth group of math problems, , key element types 1-5 are numbers, variables, geometric shapes, operators, and units in order; in the embodiment of the present invention, if , then the key element The position of the key element in the nth group of math problems is higher than that of the key element. Appears in the front position.

[0013] All the math problems with the same knowledge under investigation constitute the math problem set of that knowledge, where the math problem set of the i-th knowledge is : ; .

[0014] in: represents the sth group of math problems that examine the ith type of knowledge, represents the total number of math questions that test the i-th type of knowledge; express The set of theorems and principles involved; express Entity annotation results of express The cth key element of Represents key elements The key element type, express The number of key elements.

[0015] Optionally, the mathematical problems that test the same knowledge are clustered based on the entity labeling results to obtain multiple mathematical problem clusters that test the same knowledge, including: The clustering process of the math problems testing the i-th type of knowledge is as follows: S11: Calculate the set of math questions The distance between any two sets of math problems in and The distance between them is: .

[0016] in: Representing math problems and The distance between ; Representing math problems and The distance between key elements in the Representing theorems and principle sets and The intersection between Representing theorems and principle sets and The union between express The set of theorems and principles involved; Indicates the number of theorems and principles in the set; express Entity annotation results of Represents an exponential function with a natural constant as its base.

[0017] S12: Based on the distance between math problems, the cluster center information weight of each group of math problems is calculated. The cluster center information weight is : ; ; .

[0018] in: For math problems density; Represents a set of math problems Chinese and Mathematics Topics The number of math problems whose distance is less than a preset distance threshold; Represents a set of math problems Chinese and Mathematics Topics A set of math problems whose distances are less than a preset distance threshold. , Represents a set of math problems Any math problem in Representing math problems The distance between Representing math problems The clustering center coefficient of Represents a set of math problems Medium density greater than Mathematical problems; Indicates the preset maximum clustering center coefficient.

[0019] S13: Selecting math problems with cluster center information weights higher than a preset threshold as math problem cluster centers, merging and clustering math problems that are not math problem cluster centers and the nearest math problem cluster centers to form math problem clusters, where the set of math problem clusters that examine the i-th knowledge is : .

[0020] in: represents the kth cluster of mathematical questions that examine the i-th knowledge, Represents the total number of math question clusters that test the i-th type of knowledge.

[0021] Optionally, in step S2, the template extraction of the math question cluster is performed by using an association rule template mining method, including: The templates of mathematics question clusters are extracted by association rule template mining, and multiple question templates of mathematics question clusters are obtained. The template extraction process is as follows: S21: Clusters of math questions The math problems in the sentence are segmented and stop words are removed to obtain the phrase sequence of the math problems, where the phrase sequence contains phrases marked as key elements and phrases that are not key elements. The math problem cluster The phrase sequence of the mth group of math problems is : .

[0022] in: Represents a sequence of phrases The hth phrase in , Represents a sequence of phrases The total number of phrases in ; , Represents a cluster of math problems Number of math problems in like , then the phrase As the key element, if , then the phrase Not a key element.

[0023] S22: replacing the phrases that are key elements in the phrase sequence with the key element type to obtain a phrase sequence after the key elements are replaced.

[0024] S23: Multiple phrases in the phrase sequence are combined into a phrase set, and the support of each phrase set is calculated, where the number of key element types in each phrase set is at least , the upper limit of the number of phrases in the phrase set is , the lower limit is , is the preset phrase set threshold, where the support of phrase set Q is : ; .

[0025] in: represents the U phrases in the phrase set Q, , U represents the total number of phrases in the phrase set Q; Represents a cluster of math problems Contains the phrase , and the number of math problems in which the order of phrase appearance is consistent with the phrase set Q.

[0026] S24: Take the phrase set with support higher than the preset support threshold as the math problem cluster The title template is used, and the phrases in the title template are used as entity variables.

[0027] Optionally, the step of labeling entity variables in the topic template to obtain variable entities and immutable entities in the topic template includes: Mark the entity variables of the key element type in the question template as immutable entities; Mark entity variables that are not key element types in the question template as variable entities.

[0028] Optionally, constructing a variable entity content generation model in step S3 includes: A variable entity content generation model is constructed, wherein the variable entity content generation model takes a title template as input and takes variable entity content and assignment results of immutable entities as output, wherein the variable entity content generation model is implemented using a sequence-to-sequence method based on contextual content, and the variable entity content generation model includes an input layer, a context extraction layer, an iterative generation layer and an output layer, wherein the input layer is used to receive the title template and entity variable annotation results, the context extraction layer is used to extract contextual immutable entities of variable entities in the title template and assign values ​​to the contextual immutable entities, the iterative generation layer iteratively generates variable entity content based on the contextual immutable entities to obtain variable entity content of variable entities in the title template, and the output layer outputs the assignment results of the immutable entities and the variable entity content.

[0029] Optionally, the generating the variable entity content in the title template and the assignment result of the immutable entity by using the variable entity content generating model includes: S31: The input layer receives the question template B and the entity variable annotation results : ; .

[0030] in: represents the vth entity variable in the question template B, , Indicates the total number of entity variables in question template B; Entity variable The annotation results of Represents entity variables is a variable entity, Represents entity variables An immutable entity.

[0031] S32: The context extraction layer extracts the context immutable entities of the variable entities in the title template B. The context immutable entities of the variable entities include the preceding immutable entity and the following immutable entity. The preceding immutable entity and the following immutable entity are extracted in the following manner: obtain the preceding nearest immutable entity that is in front of the variable entity and closest to the variable entity in the title template, and obtain the following nearest immutable entity that is in the back of the variable entity and closest to the variable entity in the title template, and use the entity sequence between the variable entity and the preceding nearest immutable entity as the preceding immutable entity of the variable entity, and use the entity sequence between the variable entity and the following nearest immutable entity as the following immutable entity of the variable entity.

[0032] For mutable entities ,but The above immutable entity is , the following immutable entity is : ; .

[0033] in: express The nearest immutable entity above, express The nearest immutable entity below.

[0034] S33: Assign the context immutable entity, where the above immutable entity And the following immutable entities The assignment method is: Get The key element type is used to generate phrases that match the key element type. Assign a value to obtain the above immutable entity after the assignment ; Get The key element type is used to generate phrases that match the key element type. Assign a value to obtain the following immutable entity after the assignment ; In the embodiment of the present invention, any context-immutable entity or context-immutable entity will only be assigned once, and will not be reassigned until a group of questions are generated.

[0035] S34: The iterative generation layer iteratively generates variable entity content based on the context immutable entity to obtain variable entity content of the variable entity in the topic template, where the variable entity The iterative generation process of the variable entity content is: S341: Yes as well as Perform word vector representation to obtain word vector sequence ; In an embodiment of the present invention, the word vector representation method is the word2vec model.

[0036] S342: Set the current iteration number of the word vector sequence corresponding to the variable entity content to t, the initial value of t is 0, and the maximum value is Max, then the variable entity The t-th iteration result of the word vector sequence corresponding to the variable entity content is , , , They represent the previous word vector sequence and the following word vector sequence obtained in the t-th iteration respectively.

[0037] S343: Perform semantic iteration processing by adding random perturbations to the word vector sequence corresponding to the variable entity content, wherein the semantic iteration processing formula is: .

[0038] in: T stands for transpose; represents the semantic convolution matrix; represents the convolution operator; ReLU activation function. represents a random noise sequence; Represents the identity matrix.

[0039] According to the order of the variable entity contents in the title template, semantic iteration processing is performed on the word vector sequence corresponding to the variable entity contents in sequence, and the word vector that has completed the semantic iteration processing overwrites the current word vector that has not completed the semantic iteration processing.

[0040] S344: Let t = t + 1, return to step S343, until the maximum number of iterations is reached, and the word vector sequence obtained by the semantic iteration processing at this time Convert to phrase sequence form, and convert the above word vector sequence The last phrase of the corresponding phrase sequence is used as a variable entity The variable entity content.

[0041] S35: Extract the assignment results of all immutable entities in step S33 and the variable entity contents of all variable entities in step S34. The output layer outputs the assignment results of the immutable entities and the variable entity contents as a set of title template contents to be filled, and returns to step S33 to re-assign the immutable entities and generate the title template contents to be filled.

[0042] Optionally, in step S4, filling the variable entity content and the assignment result of the immutable entity into the question template to form a question includes: The variable entity content generated by the variable entity content generation model and the assignment results of the immutable entity are filled into the corresponding question template to form a group of questions, and the operation is repeated to generate question generation results under different question templates that test different knowledge.

[0043] Optionally, The calculation process is: initialization ,in , , Representing math problems The number of key elements; right Perform recursion, where Representing math problems Before Key elements and math problems Before The distance between key elements is calculated by the recursive formula: .

[0044] in: Representing math problems No. The key element type of each key element; Representing math problems No. The key element type of each key element; Indicates selection The maximum value in .

[0045] Based on the distance obtained by recursion, the math problem is calculated and The distance between key elements in: .

[0046] in: Representing math problems and The distance between key elements in .

[0047] In order to solve the above problem, the present invention provides an electronic device, the electronic device comprising: A memory storing at least one instruction; Communication interface, enabling electronic equipment to communicate; and The processor executes the instructions stored in the memory to implement the intelligent question generation method for knowledge point analysis described above.

[0048] In order to solve the above problems, the present invention also provides a computer-readable storage medium, in which at least one instruction is stored. The at least one instruction is executed by a processor in an electronic device to implement the intelligent question generation method for knowledge point analysis described above.

[0049] Compared with the prior art, the present invention proposes an intelligent question generation method based on knowledge point analysis, which has the following advantages: First, this scheme proposes a method for extracting question templates, which obtains a large number of math questions, and performs knowledge point annotation and entity annotation respectively. Phrases involving numbers, variables, geometric shapes, operators and units in math questions are marked as key elements as entity annotation results. The entity annotation results are used to cluster the math questions based on entity structure, and math questions that test consistent question knowledge, involve similar theorem principles and have similar question structures are merged into the same math question cluster to achieve entity-based math question clustering. The association rule template mining method is used to extract templates from the math question cluster, and a set of math question phrases with support higher than a threshold is extracted as a question template to achieve question template extraction that tests different knowledge.

[0050] At the same time, this solution eliminates a question generation method, which regards the question template phrases that are key elements as immutable entities, and other phrases as variable entities. The immutable entities are assigned values ​​using the variable entity content generation model combined with the key element type of the immutable entity, and the variable entity content of the variable entity is semantically iterated to obtain variable entity content that is consistent with the context semantics. The variable entity content and the assignment results of the immutable entity are filled into the question template to form a question, thereby realizing the automatic generation of questions for different knowledge points. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 A flowchart of an intelligent question generation method for knowledge point analysis provided in one embodiment of the present invention.

[0052] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0053] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0054] The embodiment of the present application provides an intelligent question generation method for knowledge point analysis. The execution subject of the intelligent question generation method for knowledge point analysis includes but is not limited to at least one of the electronic devices such as a server, a terminal, etc. that can be configured to execute the method provided by the embodiment of the present application. In other words, the intelligent question generation method for knowledge point analysis can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc.

[0055] Embodiment 1: A method for generating intelligent questions based on knowledge point analysis includes the following steps: S1: Obtain math problems, annotate the math problems with knowledge points and entities, and cluster the math problems that test the same knowledge based on the entity annotation results to obtain multiple math problem clusters that test the same knowledge.

[0056] In the step S1, a math problem is obtained, and knowledge point and entity annotations are performed on the math problem, including: Get N groups of math problems, and perform knowledge point annotation and entity annotation on each group of math problems. The knowledge point annotation and entity annotation process of the nth group of math problems is as follows: Mark the knowledge tested by the nth group of math questions, where the knowledge tested by the math questions includes geometry, algebra, number theory, mathematical analysis, probability and statistics. .

[0057] Mark the theorems and principles involved in the nth group of math problems, and combine the knowledge marking results, theorem and principle marking results to form the knowledge point marking results of the nth group of math problems : .

[0058] in: Indicates the knowledge point annotation results of the nth group of math problems. represents the knowledge tested by the nth group of math problems, Represents the set of theorems and principles involved in the nth group of mathematical problems.

[0059] Identify the key elements in the nth group of math problems, where the key element types include numbers, variables, geometric shapes, operators, and units, annotate the identified key elements and the types corresponding to the key elements, and obtain the entity annotation results of the nth group of math problems : .

[0060] in: represents the cth key element in the nth set of math problems, Represents key elements type, represents the total number of key elements in the nth group of math problems, , key element types 1-5 are numbers, variables, geometric shapes, operators, and units in order; in the embodiment of the present invention, if , then the key element The position of the key element in the nth group of math problems is higher than that of the key element. Appears in the front position.

[0061] All the math problems with the same knowledge under investigation constitute the math problem set of that knowledge, where the math problem set of the i-th knowledge is : ; .

[0062] in: represents the sth group of math problems that examine the ith type of knowledge, represents the total number of math questions that test the i-th type of knowledge; express The set of theorems and principles involved; express Entity annotation results of express The cth key element of Represents key elements The key element type, express The number of key elements.

[0063] The mathematical questions that examine the same knowledge are clustered based on the entity labeling results to obtain multiple mathematical question clusters that examine the same knowledge, including: The clustering process of the math problems testing the i-th type of knowledge is as follows: S11: Calculate the set of math questions The distance between any two sets of math problems in and The distance between them is: .

[0064] in: Representing math problems and The distance between ; Representing math problems and The distance between key elements in the Representing theorems and principle sets and The intersection between Representing theorems and principle sets and The union between express The set of theorems and principles involved; Indicates the number of theorems and principles in the set; express Entity annotation results of Represents an exponential function with a natural constant as its base.

[0065] S12: Based on the distance between math problems, the cluster center information weight of each group of math problems is calculated. The cluster center information weight is : ; ; .

[0066] in: For math problems density; Represents a set of math problems Chinese and Mathematics Topics The number of math problems whose distance is less than a preset distance threshold; Represents a set of math problems Chinese and Mathematics Topics A set of math problems whose distances are less than a preset distance threshold. , Represents a set of math problems Any math problem in Representing math problems The distance between Representing math problems The clustering center coefficient of Represents a set of math problems Medium density greater than Mathematical problems; Indicates the preset maximum clustering center coefficient.

[0067] S13: Selecting math problems with cluster center information weights higher than a preset threshold as math problem cluster centers, merging and clustering math problems that are not math problem cluster centers and the nearest math problem cluster centers to form math problem clusters, where the set of math problem clusters that examine the i-th knowledge is : .

[0068] in: represents the kth cluster of mathematical questions that examine the i-th knowledge, Represents the total number of math question clusters that test the i-th type of knowledge.

[0069] S2: Use association rule template mining to extract templates from math question clusters, obtain question templates for math question clusters, label entity variables in question templates, and obtain variable entities and immutable entities in question templates.

[0070] In the step S2, the template extraction of the math question cluster is performed by using the association rule template mining method, including: The templates of mathematics question clusters are extracted by association rule template mining, and multiple question templates of mathematics question clusters are obtained. The template extraction process is as follows: S21: Clusters of math questions The math problems in the sentence are segmented and stop words are removed to obtain the phrase sequence of the math problems, where the phrase sequence contains phrases marked as key elements and phrases that are not key elements. The math problem cluster The phrase sequence of the mth group of math problems is : .

[0071] in: Represents a sequence of phrases The hth phrase in , Represents a sequence of phrases The total number of phrases in ; , Represents a cluster of math problems Number of math problems in like , then the phrase As the key element, if , then the phrase Not a key element.

[0072] S22: replacing the phrases that are key elements in the phrase sequence with the key element type to obtain a phrase sequence after the key elements are replaced.

[0073] S23: Multiple phrases in the phrase sequence are combined into a phrase set, and the support of each phrase set is calculated, where the number of key element types in each phrase set is at least , the upper limit of the number of phrases in the phrase set is , the lower limit is , is the preset phrase set threshold, where the support of phrase set Q is : ; .

[0074] in: represents the U phrases in the phrase set Q, , U represents the total number of phrases in the phrase set Q; Represents a cluster of math problems Contains the phrase , and the number of math problems where the order of phrase appearance is consistent with the phrase set Q; Mark the entity variables of the key element type in the question template as immutable entities; Mark entity variables that are not key element types in the question template as variable entities.

[0075] S3: Construct a variable entity content generation model, and use the variable entity content generation model to generate variable entity content in the question template and the assignment results of immutable entities.

[0076] The variable entity content generation model is constructed in the S3 step, including: A variable entity content generation model is constructed, wherein the variable entity content generation model takes a title template as input and takes variable entity content and assignment results of immutable entities as output, wherein the variable entity content generation model is implemented using a sequence-to-sequence method based on contextual content, and the variable entity content generation model includes an input layer, a context extraction layer, an iterative generation layer and an output layer, wherein the input layer is used to receive the title template and entity variable annotation results, the context extraction layer is used to extract contextual immutable entities of variable entities in the title template and assign values ​​to the contextual immutable entities, the iterative generation layer iteratively generates variable entity content based on the contextual immutable entities to obtain variable entity content of variable entities in the title template, and the output layer outputs the assignment results of the immutable entities and the variable entity content.

[0077] The method of using the variable entity content generation model to generate variable entity content in the topic template and the assignment result of the immutable entity includes: S31: The input layer receives the question template B and the entity variable annotation results : ; .

[0078] in: represents the vth entity variable in the question template B, , Indicates the total number of entity variables in question template B; Entity variable The annotation results of Represents entity variables is a variable entity, Represents entity variables An immutable entity.

[0079] S32: The context extraction layer extracts the context immutable entities of the variable entities in the title template B. The context immutable entities of the variable entities include the preceding immutable entity and the following immutable entity. The preceding immutable entity and the following immutable entity are extracted in the following manner: obtain the preceding nearest immutable entity that is in front of the variable entity and closest to the variable entity in the title template, and obtain the following nearest immutable entity that is in the back of the variable entity and closest to the variable entity in the title template, and use the entity sequence between the variable entity and the preceding nearest immutable entity as the preceding immutable entity of the variable entity, and use the entity sequence between the variable entity and the following nearest immutable entity as the following immutable entity of the variable entity.

[0080] For mutable entities ,but The above immutable entity is , the following immutable entity is : ; .

[0081] in: express The nearest immutable entity above, express The nearest immutable entity below.

[0082] S33: Assign the context immutable entity, where the above immutable entity And the following immutable entities The assignment method is: Get The key element type is used to generate phrases that match the key element type. Assign a value to obtain the above immutable entity after the assignment ; Get The key element type is used to generate phrases that match the key element type. Assign a value to obtain the following immutable entity after the assignment ; In the embodiment of the present invention, any context-immutable entity or context-immutable entity will only be assigned once, and will not be reassigned until a group of questions are generated.

[0083] S34: The iterative generation layer iteratively generates variable entity content based on the context immutable entity to obtain variable entity content of the variable entity in the topic template, where the variable entity The iterative generation process of the variable entity content is: S341: Yes as well as Perform word vector representation to obtain word vector sequence ; In an embodiment of the present invention, the word vector representation method is the word2vec model.

[0084] S342: Set the current iteration number of the word vector sequence corresponding to the variable entity content to t, the initial value of t is 0, and the maximum value is Max, then the variable entity The t-th iteration result of the word vector sequence corresponding to the variable entity content is , , , They represent the previous word vector sequence and the following word vector sequence obtained in the t-th iteration respectively.

[0085] S343: Perform semantic iteration processing by adding random perturbations to the word vector sequence corresponding to the variable entity content, wherein the semantic iteration processing formula is: .

[0086] in: T stands for transpose; represents the semantic convolution matrix; represents the convolution operator; ReLU activation function. represents a random noise sequence; Represents the identity matrix.

[0087] According to the order of the variable entity contents in the title template, semantic iteration processing is performed on the word vector sequence corresponding to the variable entity contents in sequence, and the word vector that has completed the semantic iteration processing overwrites the current word vector that has not completed the semantic iteration processing.

[0088] S344: Let t = t + 1, return to step S343, until the maximum number of iterations is reached, and the word vector sequence obtained by the semantic iteration processing at this time Convert to phrase sequence form, and convert the above word vector sequence The last phrase of the corresponding phrase sequence is used as a variable entity The variable entity content.

[0089] S35: Extract the assignment results of all immutable entities in step S33 and the variable entity contents of all variable entities in step S34. The output layer outputs the assignment results of the immutable entities and the variable entity contents as a set of title template contents to be filled, and returns to step S33 to re-assign the immutable entities and generate the title template contents to be filled.

[0090] S4: Fill the variable entity content and the assignment results of the immutable entity into the question template to form a question, and obtain the question generation results of different knowledge points.

[0091] In the step S4, the variable entity content and the assignment result of the immutable entity are filled into the topic template to form a topic, including: The variable entity content generated by the variable entity content generation model and the assignment results of the immutable entity are filled into the corresponding question template to form a group of questions, and the operation is repeated to generate question generation results under different question templates that test different knowledge.

[0092] It should be understood that the embodiment is for illustration only and the scope of the patent application is not limited to this structure.

[0093] It should be noted that the serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments. And the terms "including", "comprising" or any other variants thereof in this article are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "including a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.

[0094] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0095] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for generating intelligent questions based on knowledge point analysis, characterized in that: The method comprises: S1: Obtain math problems, annotate the math problems with knowledge points and entities, and cluster the math problems that test the same knowledge based on the entity annotation results to obtain multiple math problem clusters that test the same knowledge; S2: Use association rule template mining to extract templates from math question clusters, obtain question templates of math question clusters, annotate entity variables in question templates, and obtain variable entities and immutable entities in question templates; S3: construct a variable entity content generation model, and use the variable entity content generation model to generate variable entity content in the question template and the assignment result of the immutable entity; S4: Fill the variable entity content and the assignment results of the immutable entity into the question template to form a question, and obtain the question generation results of different knowledge points.

2. The method for generating intelligent questions based on knowledge point analysis as claimed in claim 1, characterized in that: The S1 step includes: Get N groups of math problems, and perform knowledge point annotation and entity annotation on each group of math problems. The knowledge point annotation and entity annotation process of the nth group of math problems is as follows: Mark the knowledge tested by the nth group of math questions, where the knowledge tested by the math questions includes geometry, algebra, number theory, mathematical analysis, probability and statistics. ; Mark the theorems and principles involved in the nth group of math problems, and combine the knowledge marking results, theorem and principle marking results to form the knowledge point marking results of the nth group of math problems : ； in: Indicates the knowledge point annotation results of the nth group of math problems. represents the knowledge tested by the nth group of math problems, Represents the set of theorems and principles involved in the nth set of mathematical problems; Identify the key elements in the nth group of math problems, where the key element types include numbers, variables, geometric shapes, operators, and units, annotate the identified key elements and the types corresponding to the key elements, and obtain the entity annotation results of the nth group of math problems : ； in: represents the cth key element in the nth set of math problems, Represents key elements type, represents the total number of key elements in the nth group of math problems, ,Key element types 1-5 are numbers, variables, geometric shapes, operators, and units; All the math problems with the same knowledge under investigation constitute the math problem set of that knowledge, where the math problem set of the i-th knowledge is : ； ； in: represents the sth group of math problems that examine the ith type of knowledge, represents the total number of math questions that test the i-th type of knowledge; express The set of theorems and principles involved; express Entity annotation results of express The cth key element of Represents key elements The key element type, express The number of key elements.

3. The intelligent question generation method for knowledge point analysis according to claim 2, characterized in that: The mathematical questions that examine the same knowledge are clustered based on the entity labeling results to obtain multiple mathematical question clusters that examine the same knowledge, including: The clustering process of the math problems testing the i-th type of knowledge is as follows: S11: Calculate the set of math questions The distance between any two sets of math problems in and The distance between them is: ； in: Representing math problems and The distance between ; Representing math problems and The distance between key elements in the Representing theorems and principle sets and The intersection between Representing theorems and principle sets and The union between express The set of theorems and principles involved; Indicates the number of theorems and principles in the set; express Entity annotation results of represents an exponential function with a natural constant as base; S12: Based on the distance between math problems, the cluster center information weight of each group of math problems is calculated. The cluster center information weight is : ； ； ； in: For math problems density; Represents a set of math problems Chinese and Mathematics Topics The number of math problems whose distance is less than a preset distance threshold; Represents a set of math problems Chinese and Mathematics Topics A set of math problems whose distances are less than a preset distance threshold. , Represents a set of math problems Any math problem in Representing math problems The distance between Representing math problems The clustering center coefficient of Represents a set of math problems Medium density greater than Mathematical problems; Indicates the preset maximum cluster center coefficient; S13: Selecting math problems whose cluster center information weight is higher than a preset threshold as math problem cluster centers, merging and clustering math problems that are not math problem cluster centers and the nearest math problem cluster centers to form math problem clusters, where the set of math problem clusters that examine the i-th knowledge is : ； in: represents the kth cluster of mathematical questions that examine the i-th knowledge, Represents the total number of math question clusters that test the i-th type of knowledge.

4. The method for generating intelligent questions based on knowledge point analysis as claimed in claim 3, characterized in that: The S2 step includes: The templates of mathematics question clusters are extracted by association rule template mining, and multiple question templates of mathematics question clusters are obtained. The template extraction process is: S21: Clusters of math questions The math problems in the sentence are segmented and stop words are removed to obtain the phrase sequence of the math problems, where the phrase sequence contains phrases marked as key elements and phrases that are not key elements. The math problem cluster The phrase sequence of the mth group of math problems is : ； in: Represents a sequence of phrases The hth phrase in , Represents a sequence of phrases The total number of phrases in ; , Represents a cluster of math problems Number of math problems in like , then the phrase As the key element, if , then the phrase Not a key element; S22: replacing the phrases that are key elements in the phrase sequence with the key element type to obtain a phrase sequence after the key elements are replaced; S23: Multiple phrases in the phrase sequence are combined into a phrase set, and the support of each phrase set is calculated, where the number of key element types in each phrase set is at least , the upper limit of the number of phrases in the phrase set is , the lower limit is , is the preset phrase set threshold, where the support of phrase set Q is : ； ； in: represents the U phrases in the phrase set Q, , U represents the total number of phrases in the phrase set Q; Represents a cluster of math problems Contains the phrase , and the number of math problems where the order of phrase appearance is consistent with the phrase set Q; S24: Take the phrase set with support higher than the preset support threshold as the math problem cluster The title template is used, and the phrases in the title template are used as entity variables.

5. The method for generating intelligent questions based on knowledge point analysis as claimed in claim 4, characterized in that: The step of labeling entity variables in the topic template to obtain variable entities and immutable entities in the topic template includes: Mark the entity variables of the key element type in the question template as immutable entities; Mark entity variables that are not key element types in the question template as variable entities.

6. The method for generating intelligent questions based on knowledge point analysis according to claim 1, characterized in that: The S3 step includes: A variable entity content generation model is constructed, wherein the variable entity content generation model takes a title template as input and takes variable entity content and assignment results of immutable entities as output, wherein the variable entity content generation model includes an input layer, a context extraction layer, an iterative generation layer and an output layer, wherein the input layer is used to receive the title template and entity variable annotation results, the context extraction layer is used to extract context immutable entities of variable entities in the title template and assign values ​​to the context immutable entities, the iterative generation layer iteratively generates variable entity content based on the context immutable entities to obtain variable entity content of variable entities in the title template, and the output layer outputs the assignment results of the immutable entities and the variable entity content.

7. The method for generating intelligent questions based on knowledge point analysis as claimed in claim 6, characterized in that: The method of using the variable entity content generation model to generate variable entity content in the topic template and the assignment result of the immutable entity includes: S31: The input layer receives the question template B and the entity variable annotation results : ； ； in: represents the vth entity variable in the question template B, , Indicates the total number of entity variables in question template B; Entity variable The annotation results of Represents entity variables is a variable entity, Represents entity variables is an immutable entity; S32: The context extraction layer extracts context immutable entities of the variable entities in the topic template B, where the context immutable entities of the variable entities include the preceding immutable entity and the following immutable entity; For mutable entities ,but The above immutable entity is , the following immutable entity is : ； ； in: express The nearest immutable entity above, express The following nearest immutable entity; S33: Assign the context immutable entity, where the above immutable entity And the following immutable entities The assignment method is: Get The key element type is used to generate phrases that match the key element type. Assign a value to obtain the above immutable entity after the assignment ; Get The key element type is used to generate phrases that match the key element type. Assign a value to obtain the following immutable entity after the assignment ; S34: The iterative generation layer iteratively generates variable entity content based on the context immutable entity to obtain variable entity content of the variable entity in the topic template, where the variable entity The iterative generation process of the variable entity content is: S341: Yes as well as Perform word vector representation to obtain word vector sequence ; S342: Set the current iteration number of the word vector sequence corresponding to the variable entity content to t, the initial value of t is 0, and the maximum value is Max, then the variable entity The t-th iteration result of the word vector sequence corresponding to the variable entity content is , , , Represent the previous word vector sequence and the following word vector sequence obtained in the tth iteration respectively; S343: Perform semantic iteration processing by adding random perturbations to the word vector sequence corresponding to the variable entity content, wherein the semantic iteration processing formula is: ； in: T stands for transpose; represents the semantic convolution matrix; represents the convolution operator; ReLU activation function. represents a random noise sequence; represents the identity matrix; S344: Let t = t + 1, return to step S343, until the maximum number of iterations is reached, and the word vector sequence obtained by the semantic iteration processing at this time Convert to phrase sequence form, and convert the above word vector sequence The last phrase of the corresponding phrase sequence is used as a variable entity The mutable entity content of S35: Extract the assignment results of all immutable entities in step S33 and the variable entity contents of all variable entities in step S34. The output layer outputs the assignment results of the immutable entities and the variable entity contents as a set of title template contents to be filled.

8. The method for generating intelligent questions based on knowledge point analysis as claimed in claim 7, characterized in that: The S4 step includes: The variable entity content generated by the variable entity content generation model and the assignment results of the immutable entity are filled into the corresponding question template to form a group of questions, and the operation is repeated to generate question generation results under different question templates that test different knowledge.

9. The method for generating intelligent questions based on knowledge point analysis as claimed in claim 2, characterized in that: The calculation process is: initialization ,in , , Representing math problems The number of key elements; right Perform recursion, where Representing math problems Before Key elements and math problems Before The distance between key elements is calculated by the recursive formula: ； in: Representing math problems No. The key element type of each key element; Representing math problems No. The key element type of each key element; Indicates selection The maximum value in ; Based on the distance obtained by recursion, the math problem is calculated and The distance between key elements in: ； in: Representing math problems and The distance between key elements in .

Images