A method, system and device for extracting knowledge attributes of mathematical resources

Abstract

The present invention provides a method, system and device for extracting knowledge attributes of mathematical resources, which relates to the fields of natural language processing and online learning. The method includes: extracting knowledge attributes from mathematical expressions in mathematical learning resources; sorting the knowledge attributes according to their importance; determining whether the number of knowledge attributes is greater than a given threshold; when it is greater than the threshold, fusing the cyclic knowledge attributes in the knowledge attributes, and then determining whether the number of knowledge attributes is greater than the threshold again; when it is greater than the threshold, eliminating the redundant knowledge attributes with low importance; when it is less than the threshold, returning to the upper-level judgment until the number of knowledge attributes is less than or equal to the threshold, and ending the judgment. On the basis of effectively extracting the knowledge attributes of mathematics, this method also sorts and filters the knowledge attributes of specific learning resources according to their importance, which is convenient for the platform to quickly locate the knowledge attributes contained in the mathematical learning resources and for users to quickly search for relevant mathematical learning resources through the knowledge attributes in the huge learning resources.

Description

Technical Field

[0001] The present invention relates to the fields of natural language processing and online learning, and in particular, to a method, system and device for extracting knowledge attributes of mathematical resources. Background Art

[0002] With the rapid development of information technology, online learning, as a new educational scenario, has shown extremely strong vitality. However, with the widespread application of online education, its disadvantages have been exposed to users and have become increasingly serious. In the online learning scenario, the number of learning resources in the system is huge and updated frequently, and the quality of the learning resources themselves is also uneven. It has become a difficult problem to quickly find the learning resources that learners are interested in among the huge learning resources.

[0003] The most core part of the learning resources is the several most important knowledge attributes contained therein, which are also the parts that learners need to master most through the learning resources. The knowledge attributes contained in a mathematics learning resource are different, and the importance of each knowledge attribute to the learning resource is also different. Only by quickly and accurately identifying the important knowledge attributes corresponding to each learning resource can learning resources beneficial to the learning progress of users be provided to users, thereby saving the time for users to select learning resources and improving learning efficiency. Therefore, how to efficiently and accurately extract the first few important knowledge attributes in mathematics learning resources has become an urgent problem to be solved. Summary of the Invention

[0004] In view of the technical problems existing in the prior art, the present invention provides a method, system and device for extracting knowledge attributes of mathematical resources, aiming to sort the importance of the knowledge attributes contained in mathematics learning resources and retain several of the most important knowledge attributes to represent the mathematical resources, so as to make effective use of the mathematical resources.

[0005] According to a first aspect of the present invention, there is provided a method for extracting knowledge attributes of mathematical resources, the method comprising the following steps:

[0006] Extract knowledge attributes from the mathematical expressions in the mathematics learning resources;

[0007] Sort the importance of the knowledge attributes;

[0008] Judge whether the number of knowledge attributes is greater than a given threshold; when it is greater than the threshold, fuse the cyclic knowledge attributes in the knowledge attributes, and then judge whether the number of knowledge attributes is greater than the threshold again; when it is greater than the threshold, eliminate the redundant knowledge attributes with low importance; when it is less than the threshold, return to the upper-level judgment until the number of knowledge attributes is less than or equal to the threshold, and end the judgment.

[0009] On the basis of the above technical solutions, the present invention can also be improved as follows.

[0010] Optionally, extracting knowledge attributes from mathematical expressions in the mathematical learning resources includes:

[0011] Performing mathematical expression extraction to obtain the mathematical expressions in the mathematical resources, segmenting the mathematical expressions, and extracting the knowledge attributes in the mathematical expressions.

[0012] Optionally, ranking the importance levels of the knowledge attributes includes:

[0013] Using the TF-IDF algorithm to test and measure the importance of each knowledge attribute in the set of knowledge attributes of the learning resources.

[0014] Optionally, the tf-idf exponential expression of the knowledge attribute for the learning resource is:

[0015]

[0016] In the formula, represents the importance measure of the knowledge attribute for the current learning resource; L i represents the learning resource; j represents the order of the knowledge attribute in the knowledge attributes of the current learning resource, q j represents the j-th knowledge attribute; represents the set of mathematical resources; tf(j, L i ) is the word frequency of the knowledge attribute q j ; is the inverse document frequency of the knowledge attributes of the learning resource.

[0017] Optionally, the cyclic knowledge attributes in the fused knowledge attributes include: forming a sequence of the knowledge attributes of the obtained mathematical resources in sequence, performing cyclic search on all subsequences of the knowledge attribute sequence, fusing the same knowledge attribute subsequences, first eliminating single-knowledge cycles and then eliminating multi-knowledge cycles, and first eliminating the knowledge attributes with more cycles and then eliminating the knowledge attributes with fewer cycles.

[0018] Optionally, eliminating redundant knowledge attributes with low importance levels includes:

[0019] If the number of knowledge attributes corresponding to the learning resource is less than or equal to the threshold, the extraction process of the knowledge attributes of the learning resource ends; if the number of knowledge attributes corresponding to the learning resource exceeds the given threshold, the knowledge attributes with low importance levels are sequentially deleted to make the number of knowledge attributes corresponding to the learning resource less than or equal to the threshold.

[0020] Optionally, eliminating redundant knowledge attributes with low importance levels further includes:

[0021] For an ordered sequence with the number of knowledge attributes in the learning resources greater than the threshold, re - sort it in descending order according to the importance test metric values, and sequentially eliminate the knowledge attributes at the end of the sorting and update the value of the number of knowledge attributes until the number of knowledge attributes is equal to the threshold.

[0022] According to the second aspect of the present invention, there is provided a system for extracting knowledge attributes of mathematical resources, including a data processing module and a logical judgment module, wherein,

[0023] The data processing module is used to extract knowledge attributes from mathematical expressions in mathematical learning resources and sort the knowledge attributes according to their importance levels.

[0024] The logical judgment module is used to judge whether the number of knowledge attributes is greater than a given threshold; when it is greater than the threshold, fuse the cyclic knowledge attributes in the knowledge attributes, and then judge whether the number of knowledge attributes is greater than the threshold again; when it is greater than the threshold, eliminate the redundant knowledge attributes with low importance levels; when it is less than the threshold, return to the upper - level judgment until the number of knowledge attributes is less than or equal to the threshold, and end the judgment.

[0025] According to the third aspect of the present invention, there is provided an electronic device, including a memory and a processor, and the processor is used to implement the steps of a method for extracting knowledge attributes of mathematical resources when executing a computer program stored in the memory.

[0026] According to the fourth aspect of the present invention, there is provided a computer - readable storage medium, on which a computer program is stored, and the computer management - type program is used to implement the steps of a method for extracting knowledge attributes of mathematical resources when executed by a processor.

[0027] The technical effects and advantages of the present invention:

[0028] The present invention proposes a method, a system and a device for extracting knowledge attributes of mathematical resources. On the basis of effectively extracting mathematical knowledge attributes, this method also sorts and filters the knowledge attributes of specific learning resources according to their importance levels, which is convenient for the platform to quickly locate the knowledge attributes contained in the mathematical learning resources and for users to quickly search for relevant mathematical learning resources through the knowledge attributes in the vast learning resources. This can improve the accuracy of knowledge extraction and enhance the efficiency of finding the mathematical resources corresponding to the knowledge attributes.

[0029] Other features and advantages of the present invention will be described in the following specification, and, in part, will become apparent from the specification or will be understood by implementing the present invention. The objectives and other advantages of the present invention can be achieved and obtained through the structures pointed out in the specification, the claims and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1It is the overall flowchart of a method for extracting knowledge attributes of mathematical resources according to an embodiment of the present invention;

[0031] Figure 2 It is a schematic diagram for extracting expressions according to an embodiment of the present invention;

[0032] Figure 3 It is a schematic diagram for extracting knowledge attributes according to an embodiment of the present invention;

[0033] Figure 4 It is the flowchart of cyclic fusion of knowledge attributes according to an embodiment of the present invention;

[0034] Figure 5 It is the flowchart for judging the existence of a cycle in the judgment sequence according to an embodiment of the present invention;

[0035] Figure 6 It is the flowchart for eliminating knowledge attributes according to an embodiment of the present invention. Specific embodiments

[0036] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0037] To solve the deficiencies of the prior art, the technical solution of the present invention discloses a method for extracting knowledge attributes of mathematical resources, and the method includes the following steps:

[0038] Extract knowledge attributes from mathematical expressions in mathematical learning resources;

[0039] Sort the knowledge attributes according to the degree of importance;

[0040] Judge whether the number of knowledge attributes exceeds the threshold; when it is greater than the threshold, fuse the cyclic knowledge attributes in the knowledge attributes, and then judge again whether the number of knowledge attributes is greater than the threshold; when it is greater than the threshold, eliminate the redundant knowledge attributes with low importance; when it is less than the threshold, return to the upper-level judgment until the number of knowledge attributes is less than or equal to the threshold, and end the judgment.

[0041] It should be noted that in this embodiment, mathematical resources refer to natural language texts containing mathematical expressions, such as axiom texts, theorem texts, and exercise texts containing mathematical expressions. Knowledge attributes refer to the mathematical knowledge contained in mathematical resources, such as reduction of fractions, trigonometric functions, and Taylor's theorem. For example, if a math exercise resource contains knowledge such as trigonometric functions, Green's formula, and indefinite integrals, then this math exercise can be called a mathematical resource, and trigonometric functions, Green's formula, indefinite integrals, etc. can be called knowledge attributes. In this embodiment, the knowledge attribute set is represented by It is indicated that the knowledge attribute of learning resource L i exists and is represented by vector Q i It is indicated that the knowledge attribute vector is represented as where N Q represents the total number of knowledge attributes of the mathematical resource set; the value of q ij is the number of times the knowledge attribute q j appears in L i , and q ij is used to measure the knowledge attribute q i of learning resource L j .

[0042] Specifically, as shown below, the extraction of each mathematical resource knowledge attribute includes the following steps: Figure 1 Step 1.1: Extract the digital resource expression, including obtaining the mathematical expressions in the mathematical resource, performing segmentation and recombination of the mathematical expressions, and converting the original mathematical expression set

[0043] Step 1.2: Extract the knowledge attribute q in the mathematical expression

[0044] , and the knowledge attribute set is represented as where N j represents the total number of knowledge attributes of the mathematical resource set ; the mathematical resource L Q can be characterized by the knowledge attribute vector Q ; i q i is used to measure the knowledge attribute q of learning resource L ij ; i j ;

[0045] Step 1.3: Perform a multi-knowledge attribute importance measure based on information gain on the knowledge attribute set i in learning resource L; among them, the learning resource is a natural language text, video, audio, etc. containing knowledge that is available for learners to learn. In the present invention, the learning resource specifically refers to a natural language text containing knowledge;

[0046] Step 1.4: Traverse the knowledge attribute order sequence i of learning resource Lto determine whether the number of knowledge attributes N is greater than the given threshold λ; when the number of knowledge attributes N c is greater than the threshold λ, perform knowledge attribute cyclic fusion and knowledge attribute elimination on the knowledge attribute sequence; c

[0047] ​Step 1.5, if after executing Step 1.4, the number of knowledge attributes N i of the learning resource L c does not exceed the given threshold λ, that is, N c ≤λ, the processing operation on the learning resource L i ends; if the number of knowledge attributes of the learning resource L i exceeds the threshold λ, then it is necessary to delete the knowledge attributes with low importance according to the multi-knowledge attribute importance measurement results of Step 1.3.

[0048] In summary, on the basis of effectively extracting the knowledge attributes of mathematics, the method of the present invention also ranks and screens the importance of the knowledge attributes of specific learning resources, facilitating the platform to quickly locate the knowledge attributes contained in the mathematics learning resources and enabling users to quickly search for relevant mathematics learning resources through the knowledge attributes among the huge learning resources, so as to improve the accuracy of knowledge extraction and the efficiency of finding the mathematics resources corresponding to the knowledge attributes.

[0049] Further, extracting the knowledge attributes from the mathematical expressions in the mathematics learning resources specifically includes: performing mathematical expression extraction to obtain the mathematical expressions in the mathematics resources , splitting the mathematical expressions, and extracting the knowledge attributes in the mathematical expressions. Where L i represents the i-th mathematics resource, i = 1, 2,..., m, m is the total number of mathematics resources, 1 ≤ j ≤ n, n is the total number of mathematical expressions of L i , is the j-th mathematical expression of L i ; converting the original set of mathematical expressions into , n' is the total number of mathematical expressions of L i after splitting.

[0050] Further, performing mathematical expression extraction specifically includes: representing the mathematics resources in LaTeX markup language, using regular expression matching rules to screen the content of the mathematics resource L i , and only retaining the set of mathematical expressions , where L i represents the i-th mathematics resource, i = 1, 2,..., m, m is the total number of mathematics resources, n is the total number of mathematical expressions of L i ; splitting the operators inside the mathematical expression according to the equal sign ("=") or inequality signs ("<", ">", "≤", "≥", "≠"), and converting the original set of mathematical expressions i of the mathematics resource L into , n' is the total number of mathematical expressions of L i after splitting and recombination.

[0051] The splitting algorithm for the j-th mathematical expression is as follows: Let the set of splitting operators it contains be OP, denoted as , where α is the number of splitting operators.

[0052] a) When α = 0, it is not split and directly saved as one expression;

[0053] b) When α = 1, it is not split and directly saved as one expression containing a splitting operator;

[0054] c) When α > 1, it is split into a separate set of expressions with the splitting operators as boundaries where β is the total number of expressions split from the j-th original expression of the mathematical resource L i , and the splitting operators are stored in the set of splitting operators OP. At this time, the first and second expressions are connected in order with the splitting operators in OP to form the first expression containing a splitting operator the second and third expressions are connected to form the second expression containing a splitting operator , the third and fourth expressions are connected to form the third expression ……;

[0055] The above processing is performed on all the original mathematical expressions of the mathematical resource L i , and finally a new set of mathematical expressions of the mathematical resource Li is obtained where n′ is the total number of mathematical expressions in this set.

[0056] It should be noted that the extraction of mathematical expressions can use the non-greedy algorithm in the regular expression matching rule to extract the middle characters between pairs of "$" or "$$", and the middle characters between pairs of "\[" and "\]". The specific implementation process is as follows Figure 2 shown. Taking α = 2, β = 3, and j = 1, the mathematical expression of the mathematical resource L i has only , and the set of split mathematical expressions is ; At this time, the first and second expressions are connected in order with the splitting operators and to form the first expression containing a splitting operator , the second and third expressions are connected and to form the second expression , and a new set of mathematical expressions of the mathematical resource L i is obtained

[0057] Extract the knowledge attribute q from the mathematical expression j , eliminate the elements in the set of mathematical expressions with fewer characters than ε, and form a new set of mathematical expressions , and extract the knowledge attribute q from the mathematical expression j . In the specific implementation process, those skilled in the art can set the value of ε according to the specific situation. In this embodiment, it is illustrated by taking the value of ε as 22, specifically as Figure 3 shown, for the mathematical resource L i , eliminate the elements in the set of mathematical expressions with fewer characters than 22, and extract the new set of mathematical expressions , and the knowledge attributes in it form the knowledge attribute vector Q i of the mathematical resource L i , Determine its knowledge attribute according to the properties, theorems, axioms, etc. included in the mathematical expression, and one mathematical expression corresponds to one knowledge attribute; q ij represents the knowledge attribute q i of the learning resource L j There is a situation. If the knowledge attribute q j appears k times in L i , then q ij =k; if it does not appear, then q ij =0.

[0058] The sorting of the importance of the knowledge attributes includes:

[0059] Using the TF-IDF algorithm to test and measure the importance of each knowledge attribute in the set of knowledge attributes in the learning resource; among them, the knowledge attribute q j The importance measure of the current learning resource uses the tf-idf index of this knowledge attribute to represent.

[0060] It should be noted that TF-IDF (term frequency–inverse document frequency) is a commonly used weighting technique for information retrieval and text mining. TF-IDF is a statistical method used to evaluate the importance of a word for a document set or a single document in a corpus. The importance of a word increases in direct proportion to the number of times it appears in the document, but at the same time decreases in inverse proportion to the frequency of its appearance in the corpus.

[0061] Specifically, the knowledge attribute q jThe specific representation of the tf-idf index of the current learning resources is as follows:

[0062] Among them, represents the importance measure of the j-th knowledge attribute to the current learning resource; L i represents the learning resource; j represents the order of the knowledge attribute among the knowledge attributes in the current learning resource; represents the set of mathematical resources; tf(j, L i ) is the term frequency of the knowledge attribute q j , is the inverse document frequency of the knowledge attribute of the learning resource.

[0063] tf(j, L i ) is given by the following expression:

[0064] In the formula, q ij represents the total number of occurrences of all knowledge attributes in the learning resource L i ; the value of q ij is the number of occurrences of the knowledge attribute q j in L i ; tf(j, L i ) is the proportion of the number of occurrences of the knowledge attribute q j in the total number of occurrences of all knowledge attributes, representing the term frequency of the knowledge attribute q j , characterizing the importance of the knowledge attribute q j to the current learning resource.

[0065] is given by the following expression:

[0066] In the formula, N represents the total number of the set of mathematical resources , represents the number of learning resources containing the j-th knowledge attribute; takes the reciprocal of the proportion of the number of learning resources covered by the knowledge attribute q j in the total number of learning resources N and takes the logarithm to become the inverse document frequency, representing the discrimination ability of the knowledge attribute q j in the entire set of mathematical resources for the learning resource L i .

[0067] Determining whether the number of knowledge attributes exceeds the threshold; when it exceeds the threshold, fusing the cyclic knowledge attributes among the knowledge attributes, and the cyclic knowledge attributes among the fused knowledge attributes include:

[0068] Construct a sequence of the knowledge attributes of the obtained mathematical resources in the order of precedence, perform a loop search on all subsequences of the knowledge attribute sequence, fuse the same knowledge attribute subsequences, first eliminate single-knowledge loops and then eliminate multi-knowledge loops, and first eliminate knowledge attributes with more loops and then eliminate those with fewer loops.

[0069] Traverse the learning resource L i of the knowledge attribute order sequence , where j ∈ N c , N c represents the number of knowledge attributes of the current learning resource, and perform knowledge attribute fusion and knowledge attribute elimination on the knowledge attribute sequence; specifically, construct a sequence of the knowledge attributes of the mathematical resource L i obtained in step 1.2 in the order of precedence , represents that the knowledge attribute numbered j in the total knowledge attributes is in the r-th position of the current resource; adopt the strategies of "short length first" and "long loop first", perform a loop search on all subsequences of the knowledge attribute sequence, fuse the same knowledge attribute subsequences, first eliminate single-knowledge loops and then eliminate multi-knowledge loops, and first eliminate knowledge attributes with more loops and then eliminate those with fewer loops.

[0070] The overall process of the adopted knowledge attribute fusion strategy is specifically as Figure 4 shown, and the specific implementation method is as follows:

[0071] Let k = 1, where k represents the length of the loop for searching the knowledge attribute sequence;

[0072] used to determine whether the sequence has the possibility of having a knowledge attribute subsequence loop of length k;

[0073] When n > 1, it indicates the possibility of having a knowledge attribute subsequence loop of length k, and then determine whether there is a knowledge attribute subsequence loop of length k in the sequence , if a loop exists, perform fusion and only retain one such subsequence, that is , update the value of N , N c = N c -(n - 1); if c there is no subsequence loop, n = n - 1, if n ≤ 1 at this time, then k = k + 1, if k ≤ N return to the previous step until k > N c or N c ≤ λ, indicating that the knowledge attribute fusion is completed; c

[0074] When n ≤ 1, it indicates that the sequence​ There is no possibility of a cycle of knowledge attribute subsequences of length k. Let k = k + 1. If k ≤ N c Return to the previous step; if k > N c , then the knowledge attribute fusion ends;

[0075] The knowledge attribute cycle starts from position (m + 1). The length of the knowledge sequence cycle is k, and it cycles n times;

[0076] Furthermore, judge the sequence to see if there is a knowledge attribute subsequence of length k The specific implementation process of cycling and cycling n times is as follows Figure 5 shown, the method is as follows:

[0077] 1) When k = 1, let h = 1,

[0078] Under the condition of n + h - 1 ≤ N c , judge whether the subsequence in the sequence repeatedly appears n times continuously starting from r = h + 1. If it repeatedly appears n times continuously, then fuse the same sequences, that is, eliminate the subsequence Update the value of N c , N c = N c -(n - 1); if the above situation does not exist, then terminate the comparison with and make h = h + 1;

[0079] Traverse the sequence sequentially until n + h - 1 > N c .

[0080] 2) When k = 2, let h = 1,

[0081] Under the condition of h + 2n - 1 ≤ N c , judge whether the subsequence in the sequence repeatedly appears n times continuously starting from r = h + 2. If it repeatedly appears n times continuously, then fuse the same sequences, that is, eliminate the subsequence Update the value of N c , N c = N c -2(n - 1); if the above situation does not exist, then terminate the comparison with and make h = h + 1;

[0082] Traverse the sequence sequentially until h + 2n - 1 > N c .

[0083] 3) When k = k, let h = 1,

[0084] Under the condition that h + k(n - 1) + 1 ≤ N c , determine whether the subsequence in the sequence starts to continuously repeat n times from r = h + k. If it continuously repeats n times, fuse the same sequences, that is, eliminate subsequence Update the value of N c , N c = N c - k(n - 1); if the above situation does not exist, terminate the comparison with and set h = h + 1;

[0085] Traverse the sequence in order until h + k(n - 1) + 1 > N c .

[0086] Determine again whether the number of knowledge attributes exceeds the threshold λ; when it exceeds the threshold λ, eliminate the knowledge attributes with low redundancy and importance; when it does not exceed the threshold λ, return to the upper-level judgment until the number of knowledge attributes does not exceed the threshold λ, and end the judgment.

[0087] In this embodiment, eliminating the knowledge attributes with low redundancy and importance includes:

[0088] If the number of knowledge attributes corresponding to the learning resource L i is less than or equal to the threshold, that is, N c ≤ λ, then the extraction process of the knowledge attributes of this learning resource ends; if the number of knowledge attributes N i corresponding to the learning resource L c exceeds the threshold λ, then use the tf-idf index corresponding to the knowledge attributes in the obtained learning resource L i to sequentially delete the knowledge attributes with the lowest tf-idf index, that is, delete the knowledge attributes with low importance, so that the number of knowledge attributes N i corresponding to the learning resource L c = λ.

[0089] In the embodiment of the present invention, the threshold λ of N c is taken as 4, that is, 4 knowledge attribute numbers are used to represent a learning resource. Those skilled in the art of this research can change the value of the threshold according to the actual situation; if N c ≤ 4, the extraction process of the knowledge attributes of the learning resource L i ends; if N c > 4, then perform a multi-knowledge attribute importance measure test based on information gain and delete the knowledge attributes with low importance.

[0090] When λ takes the value of 4, it specifically includes: according to the knowledge attribute importance reduction index attribute, for the learning resource L i the number of knowledge attributes N c in the order sequence exceeding the threshold λ in accordance with its values, reorder them from largest to smallest into , and successively eliminate the knowledge attributes at the end of the sorting and update the value of N c until N c = λ;

[0091] According to the knowledge attribute importance reduction index attribute, for the learning resource L i the number of knowledge attributes N c in the order sequence greater than 4 in accordance with its values, reorder them from largest to smallest into , and successively eliminate the knowledge attributes at the end of the sorting and update the value of N c until N c = 4.

[0092] Specifically as Figure 6 shown, for the learning resource L with the number of knowledge attributes N c greater than 4 i the order sequence of knowledge attributes in in accordance with its values, reorder them from largest to smallest into , and successively eliminate the knowledge attributes at the end of the sorting and update the value of N c until N c = 4. At this time, the set of knowledge attributes retained in the learning resource L i will be used as the representative knowledge attributes of this learning resource. Among them, for the extraction of knowledge attributes for mathematical resources, the method of reducing knowledge attributes according to the measure of knowledge attribute importance is as follows. If the number of knowledge attributes corresponding to the learning resource L

[0093] is less than or equal to 4, then the extraction process of the knowledge attributes of this learning resource ends; if the number of knowledge attributes N i corresponding to the learning resource L i exceeds the threshold of 4, then use the tf-idf index corresponding to the knowledge attributes in the learning resource L c to successively delete the knowledge attributes with the lowest tf-idf index, that is, delete the knowledge attributes with low importance, so that the number of knowledge attributes corresponding to the learning resource L i is less than or equal to 4. i

[0094] ​In summary, in view of the problems existing in the current online learning and natural language processing fields, the present invention will adopt a method for extracting knowledge attributes from mathematical expressions in mathematical learning resources, sorting the importance of knowledge attributes, and fusing and eliminating knowledge attributes, and propose a method for extracting knowledge attributes of mathematical resources. On the basis of effectively extracting mathematical knowledge attributes, this method also sorts and filters the importance of knowledge attributes of specific learning resources, facilitating the platform to quickly locate the knowledge attributes contained in the mathematical learning resources and enabling users to quickly search for relevant mathematical learning resources through knowledge attributes in a vast amount of learning resources.

[0095] According to the second aspect of the present invention, the present invention also provides a system for extracting knowledge attributes of mathematical resources, including: a data processing module and a logical judgment module, wherein,

[0096] The data processing module is used to extract knowledge attributes from mathematical expressions in mathematical learning resources and sort the importance of the knowledge attributes;

[0097] The logical judgment module is used to judge whether the number of knowledge attributes exceeds a given threshold; when it exceeds the threshold, fuse the cyclic knowledge attributes in the knowledge attributes, and then judge whether the number of knowledge attributes exceeds the threshold again; when it exceeds the threshold, eliminate the redundant knowledge attributes with low importance; when it is less than the threshold, return to the upper-level judgment until the number of knowledge attributes does not exceed the threshold, and end the judgment.

[0098] It can be understood that a system for extracting knowledge attributes of mathematical resources provided by the present invention corresponds to a method for extracting knowledge attributes of mathematical resources provided in the foregoing embodiments. The relevant technical features of a system for extracting knowledge attributes of mathematical resources can refer to the relevant technical features of a method for extracting knowledge attributes of mathematical resources, which will not be elaborated here.

[0099] According to the third aspect of the present invention, an electronic device includes a memory and a processor. When the processor executes a computer program stored in the memory, it realizes the steps of a method for extracting knowledge attributes of mathematical resources as described above. The steps include:

[0100] Extract knowledge attributes from mathematical expressions in mathematical learning resources;

[0101] Sort the importance of the knowledge attributes;

[0102] Judge whether the number of knowledge attributes exceeds a given threshold; when it exceeds the threshold, fuse the cyclic knowledge attributes in the knowledge attributes, and then judge whether the number of knowledge attributes exceeds the threshold again; when it exceeds the threshold, eliminate the redundant knowledge attributes with low importance; when it is less than the threshold, return to the upper-level judgment until the number of knowledge attributes does not exceed the threshold, and end the judgment.

[0103] According to a third aspect of the present invention, a computer-readable storage medium, characterized in that a computer management program is stored thereon, and when the computer program is executed by a processor, the steps of a method for extracting knowledge attributes of mathematical resources as described above are implemented. The steps include:

[0104] Extract knowledge attributes from mathematical expressions in mathematical learning resources;

[0105] Sort the knowledge attributes according to the degree of importance;

[0106] Judge whether the number of knowledge attributes exceeds a given threshold; when it exceeds the threshold, fuse the cyclic knowledge attributes among the knowledge attributes, and then judge whether the number of knowledge attributes exceeds the threshold again; when it exceeds the threshold, eliminate the redundant knowledge attributes with low importance; when it is less than the threshold, return to the upper-level judgment until the number of knowledge attributes does not exceed the threshold, and end the judgment.

[0107] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications once they learn the basic creative concept. Therefore, the appended claims are intended to be construed to include the preferred embodiments and all changes and modifications falling within the scope of the present invention.

[0108] Finally, it should be noted that the above is only the preferred embodiments of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for extracting knowledge attributes of mathematical resources, characterized in that, It includes the following steps: Extract knowledge attributes from the mathematical expressions in the mathematical learning resources; Sort the knowledge attributes according to their importance levels, including: using the TF-IDF algorithm to measure the importance of each knowledge attribute in the set of knowledge attributes in the learning resources; Judge whether the number of knowledge attributes is greater than a given threshold; when it is greater than the threshold, form a sequence of the knowledge attributes of the obtained mathematical resources in the order of priority, perform a cyclic search on all subsequences of the knowledge attribute sequence, fuse the same knowledge attribute subsequences, first eliminate single-knowledge loops and then eliminate multi-knowledge loops, first eliminate the knowledge attributes with more loops and then eliminate the knowledge attributes with fewer loops, and then judge again whether the number of knowledge attributes is greater than the threshold; when it is greater than the threshold, eliminate the knowledge attributes with relatively lower importance levels in the learning resources; when it is less than the threshold, return to the upper-level judgment until the number of knowledge attributes is less than or equal to the threshold, and end the judgment.

2. The method for extracting knowledge attributes of mathematical resources according to claim 1, wherein The extraction of knowledge attributes from the mathematical expressions in the mathematical learning resources includes: Perform mathematical expression extraction to obtain the mathematical expressions in the mathematical resources, segment the mathematical expressions, and extract the knowledge attributes in the mathematical expressions.

3. A method for extracting knowledge attributes of mathematical resources according to claim 1, characterized in that The tf-idf exponential expression of the knowledge attributes for the learning resources is: In the formula, represents the importance measure of the knowledge attribute to the current learning resource; represents the learning resource; j represents the order of the knowledge attribute among the knowledge attributes in the current learning resource, represents the j-th knowledge attribute; represents the set of mathematical resources; is the knowledge attribute is the word frequency of is the inverse document frequency of the knowledge attribute of the learning resource.

4. A method for extracting knowledge attributes of mathematical resources according to claim 1, characterized in that, The elimination of the knowledge attributes with relatively lower importance levels in the learning resources includes: If the number of knowledge attributes corresponding to the learning resources is less than or equal to the threshold, the extraction process of the knowledge attributes of this learning resource ends; if the number of knowledge attributes corresponding to the learning resources exceeds the given threshold, delete the knowledge attributes with lower importance levels in turn to make the number of knowledge attributes corresponding to the learning resources less than or equal to the threshold.

5. A method for extracting knowledge attributes of mathematical resources according to claim 4, characterized in that, The elimination of the knowledge attributes with relatively lower importance levels in the learning resources also includes: Re-sort the sequential sequence with the number of knowledge attributes in the learning resources greater than the threshold in descending order according to the importance test metric values, and sequentially eliminate the knowledge attributes at the end of the sorting and update the value of the number of knowledge attributes until the number of knowledge attributes is equal to the threshold.

6. A system for extracting knowledge attributes of mathematical resources, characterized in that, It includes: A data processing module and a logical judgment module, where, The data processing module is used to extract knowledge attributes from the mathematical expressions in the mathematical learning resources and sort the knowledge attributes according to their importance levels; it includes: using the TF-IDF algorithm to measure the importance of each knowledge attribute in the set of knowledge attributes in the learning resources; The logical judgment module is used to judge whether the number of knowledge attributes is greater than a given threshold; when it is greater than the threshold, form a sequence of the knowledge attributes of the obtained mathematical resources in the order of priority, perform a cyclic search on all subsequences of the knowledge attribute sequence, fuse the same knowledge attribute subsequences, first eliminate single-knowledge loops and then eliminate multi-knowledge loops, first eliminate the knowledge attributes with more loops and then eliminate the knowledge attributes with fewer loops, and then judge again whether the number of knowledge attributes is greater than the threshold; when it is greater than the threshold, eliminate the knowledge attributes with relatively lower importance levels in the learning resources; when it is less than the threshold, return to the upper-level judgment until the number of knowledge attributes is less than or equal to the threshold, and end the judgment.

7. An electronic device, characterized in that, It includes a memory and a processor, and the processor is used to implement the steps of a method for extracting knowledge attributes of mathematical resources as described in any one of claims 1-5 when executing the computer program stored in the memory.

8. A computer-readable storage medium, characterized in that, A computer management program is stored thereon, and when the computer program is executed by a processor, the steps of a method for extracting mathematical resource knowledge attributes as described in any one of claims 1-5 are implemented.

Images