A teaching resource informationization sorting and auditing method of an online education platform

Abstract

The present application relates to the technical field of teaching resource information auditing, and particularly relates to a teaching resource information processing and auditing method for an online education platform. The method comprises the following steps: obtaining historical text and current auditing text in teaching resources; screening real knowledge points according to the reference relationship between the knowledge points in the historical text and the knowledge point vocabulary in the historical text; obtaining the traversal characteristic value of the combined vocabulary according to the traversal of the combined vocabulary formed by the knowledge point vocabulary in the historical text and the knowledge point vocabulary in the current auditing text in the knowledge point hierarchical classification tree; obtaining the correct degree of the real knowledge points in the current auditing text according to the reference text in the knowledge point hierarchical classification tree; and auditing the current auditing text according to the occurrence of the real knowledge points in the current auditing text in the knowledge point hierarchical classification tree, the correct degree and the traversal characteristic value. The present application improves the credibility of the auditing result of the teaching resource information.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of teaching resource information auditing, and particularly relates to a teaching resource information processing and auditing method for an online education platform. BACKGROUND

[0002] When the teaching resource information of the online education platform is processed and audited, the traditional method analyzes the content appropriateness by the occurrence of simple vocabulary in different reference auditing texts, which may cause confusion between the ordinary knowledge points and the application background or the relationship between the knowledge points, thus causing the error judgment of the degree of the sentence problem, and further causing the low credibility of the auditing result of the teaching resource information of the education platform. SUMMARY

[0003] In order to solve the problem of low credibility of the auditing result when the teaching resource information is audited by the existing method, the purpose of the present application is to provide a teaching resource information processing and auditing method for an online education platform, and the technical solution is as follows:

[0004] The present application provides a teaching resource information processing and auditing method for an online education platform, which comprises the following steps:

[0005] Obtaining the historical text and the current auditing text in the teaching resource of the online education platform;

[0006] According to the referential relationship of the knowledge points in the sentence and the distribution of the knowledge point vocabulary in the sentence of the historical text, the real combined knowledge points are screened; according to the traversal of the combination vocabulary composed of the knowledge point vocabulary in the historical text and the knowledge point vocabulary in the current auditing text when traversing the knowledge point hierarchical classification tree, the contrast vocabulary of each knowledge point vocabulary in the current auditing text is determined, and the traversal characteristic value of the combination vocabulary is obtained; the knowledge point hierarchical classification tree is constructed based on the knowledge points contained in the historical text; the reference text is screened by comprehensively considering the occurrence of the contrast vocabulary in the historical text and the traversal characteristic value;

[0007] According to the reference text in the knowledge point hierarchical classification tree, the combined correct degree of the real combined knowledge points in the current auditing text is obtained; according to the occurrence of the real combined knowledge points in the current auditing text in the knowledge point hierarchical classification tree, the combined correct degree and the traversal characteristic value, the problem degree of the real combined knowledge points in the current auditing text is determined;

[0008] The current auditing text is audited based on the problem degree.

[0009] Preferably, the real combined knowledge points are screened according to the referential relationship of the knowledge points in the sentence and the distribution of the knowledge point vocabulary in the sentence of the historical text, which comprises:

[0010] For any combination knowledge point:

[0011] Respectively, count the number of occurrences of the corresponding reference of the any combination knowledge point in each historical text;

[0012] The sentence containing the any combination knowledge point and the sentence containing the combination knowledge point reference are recorded as the first sentence;

[0013] Respectively, count the number of other knowledge point words in the first sentence other than the any combination knowledge point, and if the number ratio is greater than the preset ratio threshold, the corresponding other knowledge point word is taken as the associated knowledge point of the any combination knowledge point;

[0014] Respectively, count the first number of occurrences of the any combination knowledge point in the historical text, and the first number of sentences containing all knowledge point words contained in the any combination knowledge point;

[0015] According to the number ratio, the first number and the first number, the real combination knowledge point is screened.

[0016] Preferably, according to the number ratio, the first number and the first number, the real combination knowledge point is screened, including:

[0017] Respectively, calculate the sum of the number ratio of all knowledge point words in the first sentence other than the any combination knowledge point, and the sum of the first number and the first number; The normalized result of the product between the sum and the sum is taken as the possibility value of the any combination knowledge point being a real knowledge point;

[0018] If the possibility value is greater than a preset first possibility threshold, the any combination knowledge point is determined as a real combination knowledge point.

[0019] Preferably, according to the traversal of the combination of the knowledge point words in the historical text and the knowledge point words in the current audit text in the traversal of the knowledge point hierarchical classification tree, the control word of each knowledge point word in the current audit text is determined, including:

[0020] Respectively, combine each knowledge point word in the current audit text with each knowledge point word in the historical text to obtain a combination word;

[0021] For any combination word: traverse the two knowledge point words in the any combination word upwards along the knowledge point hierarchical classification tree by depth-first algorithm, and obtain the traversal characteristic value of the any combination word according to the traversal;

[0022] For any knowledge point vocabulary in the current audit text, the knowledge point vocabulary in the historical text in the combined vocabulary where the any knowledge point vocabulary in the current audit text is located is taken as the control vocabulary of the any knowledge point vocabulary in the current audit text when the minimum value of the traversal feature value of the combined vocabulary is taken.

[0023] Preferably, the traversal feature value of the any combined vocabulary is obtained according to the traversal situation, comprising: taking the sum of the traversal layer numbers of two knowledge points in the any combined vocabulary when the traversal positions coincide as the traversal feature value of the any combined vocabulary.

[0024] Preferably, the comprehensive control vocabulary appearance in the historical text and the traversal feature value are used to filter the reference text, comprising:

[0025] For any knowledge point vocabulary in the current audit text:

[0026] If the traversal feature value of the combined vocabulary composed of the any knowledge point vocabulary in the current audit text and its control vocabulary is less than or equal to a preset first threshold value, the first number of times of the appearance of the any knowledge point vocabulary in the historical text where the control vocabulary is located is counted;

[0027] The second number of combined vocabularies whose traversal feature values are less than a preset second threshold value is counted for all combined vocabularies of knowledge point vocabularies in the current audit text;

[0028] For any historical text, the reference value of the any historical text to the current audit text is obtained according to the sum of the second number, the first number of times of the appearance of the knowledge point vocabulary in the current audit text in the any historical text, and the average value of the traversal feature values of all combined vocabularies composed of the knowledge point vocabulary in the current audit text and the knowledge point vocabulary in the any historical text; if the reference value is greater than a preset reference value threshold value, the any historical text is taken as the reference text.

[0029] Preferably, the combined correct degree of the real combined knowledge point in the current audit text is obtained according to the reference text in the knowledge point hierarchical classification tree, comprising:

[0030] The triple corresponding to each sentence in the current audit text is obtained.

[0031] The common traversal position of all knowledge point vocabularies in each sentence in the current audit text is obtained, and the knowledge point classification in the next level of the common traversal position is taken as the classification alias of each knowledge point vocabulary.

[0032] The knowledge point vocabulary in the triple is replaced by the corresponding classification alias to obtain the replaced triple corresponding to each sentence in the current audit text.

[0033] The hierarchical alias of the vocabulary of the general knowledge point contained in the real combination knowledge point in all texts is regarded as the feature representation of each real combination knowledge point;

[0034] The second number of the feature representation of each real combination knowledge point existing in each sentence in the current audit text in the feature representation of the sentence in all reference texts, and the frequency of the replaced triple corresponding to each sentence in the current audit text in the replaced triple corresponding to the sentence in all reference texts are counted;

[0035] According to the second number and the frequency, the combination correctness degree of the real combination knowledge point in the current audit text is obtained.

[0036] Preferably, according to the occurrence of the real combination knowledge point in the current audit text in the knowledge point hierarchical classification tree, the combination correctness degree and the traversal feature value, the problem degree described by the real combination knowledge point in the current audit text is determined, comprising:

[0037] Respectively, another knowledge point vocabulary in the triple in which the real combination knowledge point appears in the current audit text except the vocabulary of the real combination knowledge point is recorded as a candidate vocabulary;

[0038] All candidate vocabularies are placed in the knowledge point hierarchical classification tree for traversal, and the two candidate vocabularies corresponding to the maximum value of the traversal feature value are obtained, and the first difference between the combination correctness degrees of the real combination knowledge points in which the two candidate vocabularies are located is obtained.

[0039] For any real combination knowledge point in the current audit text, according to the sum value of the traversal feature value corresponding to the two candidate vocabularies and the combination correctness degree corresponding to the triple in which the any real combination knowledge point is located and the corresponding first difference, the problem degree described by the any real combination knowledge point in the current audit text is obtained.

[0040] Preferably, according to the sum value of the traversal feature value corresponding to the two candidate vocabularies and the combination correctness degree corresponding to the triple in which the any real combination knowledge point is located and the corresponding first difference, the problem degree described by the any real combination knowledge point in the current audit text is obtained, comprising:

[0041] The first product of the first difference between the traversal feature value corresponding to the two candidate vocabularies and the combination correctness degree of the real combination knowledge point in which the two candidate vocabularies are located is calculated;

[0042] According to the first product and the sum value of the combination correctness degree corresponding to the any real combination knowledge point, the problem degree of the any real combination knowledge point in the current auditing text is obtained, the first product is positively correlated with the problem degree, and the sum value of the combination correctness degree is negatively correlated with the problem degree.

[0043] Preferably, the auditing of the current auditing text based on the problem degree comprises:

[0044] If the problem degree is greater than a preset problem degree threshold, the corresponding real combination knowledge point is determined as an error combination knowledge point.

[0045] The present application has at least the following beneficial effects:

[0046] The present application first screens a plurality of real combination knowledge points according to the corresponding results of the anaphora resolution of the combination knowledge points in the historical text of the teaching resources of the online education platform and the distribution of the knowledge point vocabulary in the sentence, then constructs a knowledge point hierarchical classification tree based on the knowledge points contained in the historical text, and screens reference texts according to the traversal of the combination vocabulary formed by the knowledge point vocabulary in the historical text and the knowledge point vocabulary in the current auditing text in the traversal of the knowledge point hierarchical classification tree, so that the reference value of the current auditing text is greater, and the subsequent analysis of the current auditing text is based on the similarity between the reference texts and the current auditing text, which reduces the calculation amount, and then evaluates the combination correctness degree of the real combination knowledge points, and comprehensively considers the occurrence of the real combination knowledge points in the current auditing text in the knowledge point hierarchical classification tree and the traversal characteristic value, and audits the current auditing text, thereby improving the credibility of the auditing result of the information of the teaching resources of the education platform. BRIEF DESCRIPTION OF DRAWINGS

[0047] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0048] Figure 1 The flow chart of the online education platform teaching resource informationization arrangement and auditing method provided by the embodiments of the present application;

[0049] Figure 2 The structural block diagram of the online education platform teaching resource informationization arrangement and auditing system provided by the embodiments of the present application. DETAILED DESCRIPTION

[0050] In order to further clarify the technical means and effects taken by the present application to achieve the predetermined inventive purpose, the following will be described in detail below in combination with the drawings and preferred embodiments. A teaching resource informationization sorting and auditing method of an online education platform according to the present application is described as follows.

[0051] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0052] The specific scheme of the teaching resource informationization sorting and auditing method of an online education platform provided by the present application will be described in detail below in combination with the drawings.

[0053] Embodiment of a teaching resource informationization sorting and auditing method of an online education platform:

[0054] The specific scenario to which the present embodiment is directed is that teaching resource informationization sorting and auditing is the core mechanism of education platform quality assurance. By establishing a strict auditing system, it can effectively filter out errors, outdated and pseudo-scientific content, and ensure the accuracy and reliability of knowledge dissemination. The present embodiment will judge the combination of knowledge points according to the lexical relationship in the text, judge the degree of combination of knowledge points in the current auditing text according to the similarity of the upper classification of knowledge point lexical in the historical text and the current auditing text, and then realize the auditing and sorting of the current auditing text.

[0055] The present embodiment proposes a teaching resource informationization sorting and auditing method of an online education platform, as shown in Figure 1 The teaching resource informationization sorting and auditing method of an online education platform of the present embodiment includes the following steps:

[0056] Step S1, obtaining historical text and current auditing text in the teaching resources of an online education platform.

[0057] The present embodiment first obtains all texts from existing education corpus, and records the obtained texts as historical texts, and obtains the current auditing text.

[0058] Step S2, according to the reference relationship of the combination of knowledge points in the historical text and the distribution of the knowledge point lexical in the sentence, filtering the real combination of knowledge points; according to the traversal of the combination of knowledge point lexical formed by the knowledge point lexical in the historical text and the knowledge point lexical in the current auditing text when traversing the knowledge point hierarchical classification tree, determining the contrast lexical of each knowledge point lexical in the current auditing text, and obtaining the traversal characteristic value of the combination of lexical; the knowledge point hierarchical classification tree is constructed based on the knowledge points contained in the historical text; the reference text is filtered by comprehensively considering the occurrence of the contrast lexical in the historical text and the traversal characteristic value.

[0059] There may be a situation in which two different knowledge points are combined in the teaching resources, and the new knowledge point vocabulary generated after the combination may have different knowledge fields and application backgrounds than the original knowledge points. Therefore, when a simple vocabulary matching method is used to analyze the content of a combined knowledge point in new teaching resources, the relationship between ordinary knowledge points and the application background or related knowledge points of the combined knowledge point may be confused, resulting in errors in judging the teaching content of new teaching resources.

[0060] To accurately judge the combined knowledge points in the teaching resources, the referents in the vocabulary of a single resource text can be resolved. The more the combination of a knowledge point is referred to in the resource text, the more the combination of the knowledge point directly appears in the sentences contained in the resource text, and the more consistent other related knowledge points appear, the more the combined knowledge point is established.

[0061] The KMP algorithm is used to match and identify the vocabulary of ordinary knowledge points in these historical texts. The pre-trained language model BERT is used to obtain the referential relationship in the text of the teaching resources. For example, the combination of atrioventricular node ablation and pacemaker implantation in the sentence "Atrioventricular node ablation and pacemaker implantation as an alternative therapy for reentrant tachycardia" is referred to as this alternative therapy in other sentences.

[0062] Next, this embodiment takes any combined knowledge point in the historical text as an example to illustrate that the method provided in this embodiment can be used to process other combined knowledge points in the historical text.

[0063] Specifically, for any combined knowledge point:

[0064] The number of occurrences of the corresponding reference of the combined knowledge point in each historical text is counted; the sentence containing the combined knowledge point and the sentence containing the reference of the combined knowledge point are recorded as the first sentence; the number of each kind of knowledge point vocabulary other than the combined knowledge point in all first sentences is counted, and the larger the number ratio, the more relevant the corresponding other knowledge point vocabulary is to the combined knowledge point. Therefore, if the number ratio is greater than a preset ratio threshold, the corresponding other knowledge point vocabulary is used as the associated knowledge point of the combined knowledge point; in this embodiment, the preset ratio threshold is 0.5, and in specific applications, the implementer can set it according to the specific circumstances.

[0065] The more the number of times the combination knowledge point appears in the historical text, the more the number of sentences containing all the knowledge point vocabularies in the combination knowledge point, and the greater the proportion of the appearance of each larger associated knowledge point, the more the combination knowledge point is established. Therefore, the number of times the combination knowledge point appears in the historical text is counted, and the number is denoted as a first number; the number of sentences containing all the knowledge point vocabularies in the combination knowledge point is counted, and the number is denoted as a first number; the sum of the number of all knowledge point vocabularies in the first sentence except the combination knowledge point is calculated, and the sum of the first number and the first number is calculated; the normalized result of the product between the sum and the sum is taken as the possibility value of the combination knowledge point being a real knowledge point; if the possibility value is greater than a preset first possibility threshold, the combination knowledge point is determined as a real combination knowledge point. In this embodiment, there are many methods of data normalization, and the existing linear normalization method can be used to normalize the data, so that the value range of the result obtained after normalization is (0, 1). In this embodiment, the preset first possibility threshold is 0.7, and in specific applications, the implementer can set it according to the specific circumstances.

[0066] By using the above method, a plurality of real knowledge points can be screened from the historical text.

[0067] In order to reduce the excessive running consumption when searching for reference text, it is necessary to roughly screen the historical text, and to perform simple screening on the historical text according to the interval level size of the vocabularies contained in the historical text and the vocabularies in the current audit text in the knowledge point hierarchical classification tree and the existence of similar classification vocabularies in the historical text.

[0068] All knowledge point vocabularies contained in the historical text are classified by a classification and regression (C&R) tree node algorithm to establish a knowledge point hierarchical classification tree. The establishment process of the knowledge point hierarchical classification tree is prior art, which will not be described in detail here.

[0069] Each knowledge point vocabulary in the current audit text is combined with each knowledge point vocabulary in the historical text to obtain a plurality of combined vocabularies, each of which is composed of one knowledge point vocabulary in the current audit text and one knowledge point vocabulary in the historical text.

[0070] For any combined vocabulary: the two knowledge point vocabularies in the combined vocabulary are traversed upwards along the knowledge point hierarchical classification tree by a depth-first search (DFS) algorithm, and when the traversal positions coincide, the sum of the traversal levels of the two knowledge points in the combined vocabulary is taken as the traversal characteristic value of the combined vocabulary. By using this method, the traversal characteristic value of each combined vocabulary can be obtained.

[0071] For any knowledge point vocabulary in the current audit text, since there are multiple historical texts, the knowledge point vocabulary may exist in multiple combination vocabularies at the same time, and the knowledge point vocabulary in the historical text in the combination vocabulary corresponding to the minimum value of the feature value of the combination vocabulary in which the knowledge point vocabulary is located is taken as the control vocabulary of the knowledge point vocabulary in the current audit text. By using this method, the control vocabulary of each knowledge point vocabulary in the current audit text can be obtained.

[0072] For any knowledge point vocabulary in the current audit text: if the traversal feature value of the combination vocabulary composed of the knowledge point vocabulary and its control vocabulary in the current audit text is less than or equal to a preset first threshold value, the proportion of the number of times that the knowledge point vocabulary appears in the historical text in which the control vocabulary is located is calculated. By using the above method, the proportion of the number of times that each knowledge point vocabulary in the current audit text appears in the historical text in which the control vocabulary is located is calculated. In this embodiment, the preset first threshold value is 1 / 2 of the number of layers of the knowledge point hierarchical classification tree, and in specific applications, the implementer can set it according to the specific situation.

[0073] The number of combination vocabularies in which the traversal feature value of the combination vocabulary of all knowledge point vocabularies in the current audit text is less than a preset second threshold value is calculated, and the number is recorded as a second number; in this embodiment, the preset second threshold value is 5, and in specific applications, the implementer can set it according to the specific situation.

[0074] For any historical text, the reference value of the historical text to the current audit text is obtained according to the second number, the sum of the proportions of the number of times that the knowledge point vocabularies in the current audit text appear in the historical text, and the average value of the traversal feature values of all combination vocabularies composed of the knowledge point vocabularies in the current audit text and the knowledge point vocabularies in the historical text. The calculation formula of the reference value of the historical text to the current audit text is specifically:

[0075]

[0076] wherein, represents the reference value of the jth historical text to the current audit text; represents the number of combination vocabularies in which the traversal feature value of the combination vocabulary of all knowledge point vocabularies in the current audit text is less than a preset second threshold value, that is, the second number; represents the number of knowledge point vocabularies in the current audit text, represents the proportion of the number of times that the ith knowledge point vocabulary in the current audit text appears in the jth historical text, represents the average value of the traversal feature values of all combination vocabularies composed of the knowledge point vocabularies in the current audit text and the knowledge point vocabularies in the jth historical text, is a normalization function.​

[0077] The more traversal layers and smaller collocation times the current review text and the knowledge point vocabulary in the j-th historical text have on the knowledge point hierarchical classification tree, the smaller the average traversal feature values ​​of all combined vocabulary consisting of the knowledge point vocabulary in the corresponding current review text and the knowledge point vocabulary in the j-th historical text, and the greater the proportion of the number of times the i-th knowledge point vocabulary in the current review text appears in the j-th historical text, it means that the content in the j-th historical text is similar to the content knowledge points in the current review text, and the proportion of existence in the text is large, so the j-th historical text has a higher reference value for the current review resource text.

[0078] Using the above method, the reference value of each historical text to the current review text can be obtained. If the reference value is greater than a preset reference value threshold, the corresponding historical text is used as the reference text. In this embodiment, the preset reference value threshold is 0.7. In specific applications, the implementer can set it according to the specific situation.

[0079] So far, this embodiment has screened out multiple reference texts from all historical texts using the above method.

[0080] Step S3, based on the reference text in the knowledge point hierarchical classification tree, obtain the degree of combination accuracy of the real combined knowledge points in the current review text; based on the appearance, combination accuracy and traversal characteristic value of the real combined knowledge points in the current review text in the knowledge point hierarchical classification tree, determine the degree of problem described by the real combined knowledge points in the current review text.

[0081] There may be some combined knowledge points in the newly added resource text that do not appear in the reference text, but knowledge points with the same combination based on their superordinate classification may exist in the reference text. The more such knowledge points exist, the more likely it is that the combined knowledge points are correct.

[0082] For each common knowledge point vocabulary in each sentence of the current review resource text, multiple knowledge point vocabulary items are placed into a knowledge point hierarchical classification tree. A depth-first-search (DFS) algorithm is used to traverse the tree to obtain the common traversal position of all knowledge point vocabulary items. The knowledge point classification in the next level below this position is used as the hierarchical alias for each knowledge point vocabulary item. For example, in the sentence "AV node ablation and pacemaker implantation as alternative therapies for reentrant tachycardia," AV node ablation, pacemaker implantation, and alternative therapies are classified as medical procedures, while reentrant tachycardia is classified as a disease.

[0083] The triples are obtained for each sentence in each text by the GenerateIE algorithm. The triples are composed of three elements, represented as (subject / predicate, object / object), for describing the association between entities. For example, "William Shakespeare wrote Hamlet", the subject is William Shakespeare, the predicate is writing, and the object is Hamlet, so the corresponding triple is (William Shakespeare, writing, Hamlet). By using this method, the triple corresponding to each sentence in the current review text and the triple corresponding to each sentence in the reference review text can be obtained.

[0084] The common traversal position of all knowledge point vocabularies in each sentence in the current review text is obtained, and the knowledge point classification in the next level of the common traversal position is used as the classification substitute of each knowledge point vocabulary. The knowledge point vocabulary in the triple is replaced by the corresponding classification substitute to obtain the replaced triple corresponding to each sentence in the current review text.

[0085] The classification substitute of the vocabulary contained in the real combination knowledge point in all texts is used as the feature representation of each real combination knowledge point.

[0086] The number of times that the feature representation of each real combination knowledge point existing in each sentence in the current review text appears in the feature representation of the sentence in all reference texts is counted, and this number is recorded as the second number. Each real combination knowledge point in the current review text has a corresponding second number. The frequency of the replaced triple corresponding to each sentence in the current review text appearing in the replaced triple corresponding to the sentence in all reference texts is calculated. Next, according to the second number and the frequency, the combination correctness of the real combination knowledge point in the current review text is obtained. Specifically, the maximum value of all second numbers is obtained. For any real combination knowledge point in the current review text, the ratio between the second number corresponding to the real combination knowledge point and the maximum value of all second numbers is calculated. The product of this ratio and the frequency of the replaced triple corresponding to each sentence in the current review text appearing in the replaced triple corresponding to the sentence in all reference texts is used as the combination correctness of the real combination knowledge point. The greater the combination correctness, the more likely it is that the upper classification composition form of the real combination knowledge point meets the requirements and the content is correct. By using the above method, the combination correctness of each real combination knowledge point in the current review text can be obtained.

[0087] For any real combination knowledge point in the current audit text, another knowledge point vocabulary except the real combination knowledge point in the triple in which the real combination knowledge point is located is recorded as a candidate vocabulary; all candidate vocabularies are placed into the knowledge point hierarchical classification tree for traversal, the maximum value of the traversal feature value is obtained, and two candidate vocabularies corresponding to the maximum value are recorded as two candidate vocabularies corresponding to the real combination knowledge point, a difference between the combination correctness degrees of the real combination knowledge points in which the two candidate vocabularies are located is obtained, and the difference is recorded as a first difference. The specific obtaining process of the difference between the combination correctness degrees of the real combination knowledge points in which the two candidate vocabularies are located is that: the average value of the combination correctness degrees of the real combination knowledge points in which each of the two candidate vocabularies is located is calculated, and the absolute value of the difference between the average values of the combination correctness degrees of the real combination knowledge points in which the two candidate vocabularies are located is taken as the difference between the combination correctness degrees of the real combination knowledge points in which the two candidate vocabularies are located. Then, the product of the traversal feature value corresponding to the two candidate vocabularies corresponding to the real combination knowledge point and the first difference between the combination correctness degrees of the real combination knowledge points in which the two candidate vocabularies are located is calculated, and the product is recorded as a first product. Next, according to the first product and the sum of the combination correctness degrees of the triple corresponding to the real combination knowledge point, the problem degree described by the real combination knowledge point in the current audit text is obtained, the first product is positively correlated with the problem degree, and the sum of the combination correctness degrees is negatively correlated with the problem degree.

[0088] The positive correlation means that the dependent variable increases with the increase of the independent variable, and the dependent variable decreases with the decrease of the independent variable, which can be an addition relationship, a multiplication relationship, etc., and is determined by actual application. The negative correlation means that the dependent variable decreases with the increase of the independent variable, and the dependent variable increases with the decrease of the independent variable, which can be a subtraction relationship, a division relationship, etc., and is determined by actual application.

[0089] In the embodiment, a specific calculation formula of the problem degree is given, and the problem degree described by the kth real combination knowledge point in the current audit text can be represented as:

[0090]

[0091] wherein, represents the problem degree described by the kth real combination knowledge point in the current audit text, represents the traversal feature value corresponding to the two candidate vocabularies corresponding to the kth real combination knowledge point, represents the first difference between the combination correctness degrees of the real combination knowledge points in which the two candidate vocabularies corresponding to the kth real combination knowledge point are located, denotes the combination correctness degree corresponding to the i-th triple in which the k-th real combination knowledge point is located, denotes the number of triples in which the k-th real combination knowledge point is located, denotes a normalization function, so that the value range of the problem degree is (0, 1).

[0092] denotes the first product, denotes the sum of the combination correctness degrees corresponding to the triples in which the k-th real combination knowledge point is located. When the traversal feature values corresponding to the two candidate words corresponding to the k-th real combination knowledge point are larger, the difference between the combination correctness degrees of the real combination knowledge points in which the two candidate words corresponding to the k-th real combination knowledge point are located is larger, and the sum of the combination correctness degrees corresponding to the triples in which the k-th real combination knowledge point is located is smaller, it indicates that the k-th real combination knowledge point involves a wider field in the current audit text, and the difference in the combination correctness degree is larger, and the possibility of the problem described in the text is larger, that is, the problem degree of the k-th real combination knowledge point described in the current audit text is larger.

[0093] By using the above method, the problem degree of each real combination knowledge point described in the current audit text can be obtained.

[0094] In step S4, the current audit text is audited based on the problem degree.

[0095] In the embodiment, the problem degree of each real combination knowledge point described in the current audit text is obtained in step S3, and then the current audit text is audited based on the problem degree.

[0096] Specifically, if the problem degree is greater than a preset problem degree threshold, the corresponding real combination knowledge point is determined as an error combination knowledge point, which can be directly marked and fed back to the submitter; if the problem degree is less than or equal to the preset problem degree threshold and cannot be matched in the historical text, the corresponding real combination knowledge point is transferred to manual review for further determination. The sentence position with the problem and the corresponding problem degree are stored and transmitted to the database. The problem sentences are searched by using the SQL query statement, and the sentences with problems are fed back to the uploading person for modification. In the embodiment, the preset problem degree threshold is 0.7, which can be set according to the specific situation in the specific application.

[0097] Thus, the audit of the current audit text is completed by using the method provided in the embodiment.

[0098] The embodiment first screens a plurality of real knowledge points according to the corresponding results of the anaphora resolution of the knowledge points in the sentences of the historical texts in the teaching resources of the online education platform and the distribution of the knowledge point vocabularies in the sentences, then constructs a knowledge point hierarchical classification tree based on the knowledge points contained in the historical texts, screens reference texts according to the traversal of the combination vocabularies formed by the knowledge point vocabularies in the historical texts and the knowledge point vocabularies in the current audit text in the traversal of the knowledge point hierarchical classification tree, the reference texts have greater reference value for the current audit text, the subsequent analysis of the current audit text is performed according to the similarity between the reference texts and the current audit text, the calculation amount is reduced, the combination correctness of the real knowledge points is evaluated, and the current audit text is audited according to the occurrence of the real knowledge points in the current audit text in the knowledge point hierarchical classification tree and the traversal characteristic value, thereby improving the credibility of the audit result of the information of the teaching resources of the education platform.

[0099] An online education platform teaching resource informationization sorting and auditing system embodiment:

[0100] Referring to Figure 2 , a structural block diagram of an online education platform teaching resource informationization sorting and auditing system provided by an embodiment of the present application is shown, and the system comprises a data acquisition module, a screening module, an evaluation module and an auditing module;

[0101] The data acquisition module is configured to acquire historical texts and a current audit text in the teaching resources of the online education platform.

[0102] The screening module is configured to screen real knowledge points according to the anaphoric relationship of the knowledge points in the sentences of the historical texts and the distribution of the knowledge point vocabularies in the sentences, determine the contrast vocabularies of each knowledge point vocabulary in the current audit text according to the traversal of the combination vocabularies formed by the knowledge point vocabularies in the historical texts and the knowledge point vocabularies in the current audit text in the traversal of the knowledge point hierarchical classification tree, and obtain the traversal characteristic value of the combination vocabularies; the knowledge point hierarchical classification tree is constructed based on the knowledge points contained in the historical texts; and reference texts are screened according to the occurrence of the contrast vocabularies in the historical texts and the traversal characteristic value.

[0103] The evaluation module is configured to obtain the combination correctness of the real knowledge points in the current audit text according to the reference texts in the knowledge point hierarchical classification tree, and determine the problem degree of the real knowledge points in the current audit text according to the occurrence of the real knowledge points in the current audit text in the knowledge point hierarchical classification tree, the combination correctness and the traversal characteristic value.

[0104] The auditing module is configured to audit the current audit text based on the problem degree.

[0105] It should be understood, Figure 2 The structure diagram and modules of the teaching resource informationization arrangement and auditing system of an online education platform shown can be implemented in various ways. For example, in some embodiments, the system and its modules can be implemented by hardware, software, or a combination of software and hardware. The hardware part can be implemented by using special logic; the software part can be stored in a memory and executed by a suitable instruction execution system, such as a microprocessor or a specially designed hardware. Those skilled in the art can understand that the above method and system can be implemented by using computer executable instructions and / or included in processor control code, such as provided on a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The system and its modules of the present specification can not only have hardware circuit implementation such as very large scale integrated circuit or gate array, semiconductor such as logic chip, transistor, or programmable hardware device such as field programmable gate array, programmable logic device, etc., but also can be implemented by software executed by various types of processors, and also can be implemented by a combination of the above hardware circuit and software (for example, firmware).

[0106] More details about each of the above modules can be referred to other places in the present specification, and will not be described here.

[0107] In other embodiments, a medium is also provided, which stores at least one computer executable program, and the at least one program is executed by a computer to make the computer execute the steps in the above-mentioned teaching resource informationization arrangement and auditing method of an online education platform in the embodiments. The medium can be a computer readable storage medium.

[0108] The provided system and medium are used to execute the corresponding method provided above, so the beneficial effects that can be achieved can refer to the beneficial effects of the corresponding method provided above, and will not be described here.

[0109] It should be noted that: the above only describes the preferred embodiments of the present application, and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the principles of the present application shall be included in the protection scope of the present application.

Claims

1. An informationization sorting and auditing method for teaching resources of an online education platform, characterized in that, The method comprises the following steps: Obtain historical texts and current review texts in teaching resources of an online education platform; According to the distribution of the knowledge point vocabulary in the historical texts and the distribution of the knowledge point vocabulary in the current review texts, determine the corresponding knowledge point vocabulary of each knowledge point vocabulary in the current review texts, and obtain the traversal feature value of the combination vocabulary; the knowledge point hierarchical classification tree is constructed based on the knowledge points contained in the historical texts; According to the traversal of the combination vocabulary, a reference text is obtained; Obtain the corresponding triple of each sentence in the current review text; obtain the common traversal position of all knowledge point vocabularies in the knowledge point hierarchical classification tree in each sentence in the current review text, and classify the knowledge points in the next level of the common traversal position as the hierarchical representation of each knowledge point vocabulary; replace the knowledge point vocabulary in the triple with the corresponding hierarchical representation to obtain the replaced triple corresponding to each sentence in the current review text; the hierarchical representation of the vocabulary of the general knowledge point contained in the real combination knowledge point in all texts is used as the feature representation of each real combination knowledge point; the second number of the feature representation of each real combination knowledge point existing in each sentence in the current review text appearing in the feature representation of the sentence in all reference texts, and the frequency of the replaced triple corresponding to each sentence in the current review text appearing in the replaced triple corresponding to the sentence in all reference texts; according to the second number and the frequency, the combination correctness degree of the real combination knowledge point in the current review text is obtained; Respectively, record the other knowledge point vocabulary in the triple in which the real combination knowledge point appears as a candidate vocabulary except the vocabulary of the real combination knowledge point; place all candidate vocabularies in the knowledge point hierarchical classification tree for traversal, obtain the two candidate vocabularies corresponding to the maximum value of the traversal feature value, and obtain the first difference between the combination correctness degrees of the real combination knowledge points in which the two candidate vocabularies are located; for any real combination knowledge point in the current review text, according to the sum value of the traversal feature values corresponding to the two candidate vocabularies, the combination correctness degree corresponding to the triple in which the any real combination knowledge point is located, and the first difference corresponding to the any real combination knowledge point, the problem degree of the any real combination knowledge point in the current review text is obtained; Based on the problem degree, the current review text is reviewed.

2. The method for informationization arrangement and auditing of teaching resources of online education platform according to claim 1, characterized in that, The combination correctness degree of the real combination knowledge point in the current review text is obtained according to the traversal of the combination vocabulary, comprising: For any combination knowledge point: Respectively, count the number of appearances of the corresponding reference of the any combination knowledge point in each historical text; Record the sentence containing the any combination knowledge point and the sentence containing the reference of the any combination knowledge point as the first sentence; respectively, if the number ratio is greater than a preset ratio threshold, the corresponding other knowledge point vocabulary is taken as the related knowledge point of the any combined knowledge point; respectively, the first number of the any combined knowledge point in the historical text, and the first number of the sentence containing all knowledge point vocabularies contained in the any combined knowledge point; According to the number ratio, the first number and the first number, the real combined knowledge point is screened.

3. The method for informationization arrangement and auditing of teaching resources of online education platform according to claim 2, characterized in that, According to the number ratio, the first number and the first number, the real combined knowledge point is screened, including: respectively, the sum of the number ratio of all knowledge point vocabularies in the first sentence except the any combined knowledge point, and the sum of the first number and the first number; The normalized result of the product between the sum and the sum is taken as the possibility value of the any combined knowledge point being a real knowledge point; If the possibility value is greater than a preset first possibility threshold, the any combined knowledge point is determined as a real combined knowledge point.

4. The method for informationization arrangement and auditing of teaching resources of online education platform according to claim 1, characterized in that, The determination of the control vocabulary of each knowledge point vocabulary in the current audit text according to the traversal of the combination of knowledge point vocabularies in the historical text and the knowledge point vocabularies in the current audit text when traversing the knowledge point hierarchical classification tree, including: respectively, each knowledge point vocabulary in the current audit text is combined with each knowledge point vocabulary in the historical text to obtain a combination vocabulary; For any combination vocabulary: two knowledge point vocabularies in the any combination vocabulary are traversed upwards along the knowledge point hierarchical classification tree by depth first algorithm, and the traversal characteristic value of the any combination vocabulary is obtained according to the traversal; For any knowledge point vocabulary in the current audit text, the knowledge point vocabulary in the historical text in the combination vocabulary corresponding to the minimum value of the traversal characteristic value in the combination vocabulary of the any knowledge point vocabulary in the current audit text is taken as the control vocabulary of the any knowledge point vocabulary in the current audit text.

5. The method for informationization arrangement and auditing of teaching resources of online education platform according to claim 4, characterized in that, The determination of the control vocabulary of each knowledge point vocabulary in the current audit text according to the traversal of the combination of knowledge point vocabularies in the historical text and the knowledge point vocabularies in the current audit text when traversing the knowledge point hierarchical classification tree, including:

6. The method for informationization arrangement and auditing of teaching resources of online education platform according to claim 1, characterized in that, For any knowledge point vocabulary in the current audit text: If the traversal characteristic value of the combination vocabulary composed of the any knowledge point vocabulary in the current audit text and its control vocabulary is less than or equal to a preset first threshold, the first number ratio of the any knowledge point vocabulary appearing in the historical text of the control vocabulary is counted; The second number of the combination vocabulary whose traversal characteristic value is less than a preset second threshold is counted. ​ For any historical text, according to the second number, the sum of the first number of the knowledge point vocabulary in the current audit text in any historical text in which the contrast vocabulary appears, and the average value of the traversal feature value of all combination vocabularies composed of the knowledge point vocabulary in the current audit text and the knowledge point vocabulary in the any historical text, the reference value of the any historical text to the current audit text is obtained; if the reference value is greater than a preset reference value threshold, the any historical text is taken as a reference text.

7. The method for informationization arrangement and auditing of teaching resources of online education platform according to claim 1, characterized in that, The problem degree of the any real combination knowledge point in the current audit text is obtained according to the sum of the traversal feature values corresponding to the two candidate vocabularies, the combination correct degree corresponding to the triple in which the any real combination knowledge point is located, and the corresponding first difference, including: A first product of the first difference between the traversal feature values corresponding to the two candidate vocabularies and the combination correct degree of the real combination knowledge point in which the two candidate vocabularies are located is calculated; According to the first product and the sum of the combination correct degree corresponding to the triple in which the any real combination knowledge point is located, the problem degree of the any real combination knowledge point in the current audit text is obtained, the first product and the problem degree are in a positive correlation relationship, and the sum of the combination correct degree and the problem degree are in a negative correlation relationship.

8. The method for informationization arrangement and auditing of teaching resources of online education platform according to claim 1, characterized in that, The current audit text is audited based on the problem degree, including: If the problem degree is greater than a preset problem degree threshold, the corresponding real combination knowledge point is determined as an error combination knowledge point.

Images