Courseware knowledge point classification method based on multi-label classification and knowledge base matching

Abstract

The present application relates to the technical field of neural network and knowledge point classification. Specifically, the present application relates to a method for classifying courseware knowledge points based on a combination of multi-label classification and knowledge base matching. The method comprises: obtaining the courseware and preprocessing it to obtain the target text; sorting the text according to the coordinate axis of the target text to obtain a paginated text sequence; assembling all text sequences into a long text; inputting the long text into a multi-label text classification model based on a neural network for multi-label classification to obtain a candidate knowledge point list; verifying the candidate knowledge point list based on knowledge base matching to obtain a target knowledge point list. The present application reduces the reliance on manual operation, reduces the work cost, realizes a more accurate classification and identification of knowledge points, and thus improves work efficiency.

Description

Technical Field

[0001] The present application relates to the field of deep learning technology. More specifically, the present application relates to a courseware knowledge point classification method based on a combination of multi-label classification and knowledge base matching. Background Art

[0002] With the rapid development of informatization today, a large number of user-generated content (UGC) platforms have been accumulated in the education industry, such as courseware resource platforms. In the courseware resource platform, the resources of the courseware need to be classified and labeled to facilitate subsequent management and use. Teaching knowledge points (test points) are relatively independent minimum units of knowledge and theory, which can be used to measure students' mastery and proficiency of the entire teaching knowledge system. Therefore, they are often labeled as electronic courseware.

[0003] However, the commonly used method in the industry currently still relies on manual classification during the user upload process. Specifically: when a user uploads a courseware to a courseware resource platform, the reviewer or uploader needs to manually enter the knowledge points that the courseware corresponds to in the knowledge point system. This approach has two shortcomings: on the one hand, a courseware usually corresponds to multiple knowledge points, and the number of knowledge points is uncertain. Therefore, the classification of courseware is essentially a multi-label classification rather than a single-label classification problem. This requires reviewers to compare and review the knowledge points one by one to obtain a more accurate result, which makes the annotation cost high and the speed slow; on the other hand, the knowledge point system is not fixed. Unlike the version chapters, the current education department has not given a unified knowledge point system. Different schools and institutions often launch their own teaching knowledge point systems. Therefore, the knowledge point system will change later, and the cost of subsequent revisions is quite high. Summary of the invention

[0004] Based on the above technical problems, the present invention aims to solve the problem that the classification of courseware knowledge points is too dependent on manual work, and provides a courseware knowledge point classification method that combines multi-label classification and knowledge base matching to achieve high efficiency and low cost classification of courseware knowledge points.

[0005] A first aspect of the present invention provides a courseware knowledge point classification method based on a combination of multi-label classification and knowledge base matching, the method comprising:

[0006] Obtain the courseware and preprocess it to obtain the target text;

[0007] Sort the text according to the coordinate axis of the target text to obtain a paginated text sequence;

[0008] Put all text sequences together into a long text;

[0009] Inputting the long text into a multi-label text classification model based on a neural network for multi-label classification to obtain a candidate knowledge point list;

[0010] The candidate knowledge point list is verified based on knowledge base matching to obtain a target knowledge point list.

[0011] Specifically, the candidate knowledge point list is verified based on knowledge base matching to obtain a target knowledge point list, including:

[0012] In the paginated text sequence, a knowledge point definition category is selected as a first category of knowledge points, and a learning goal or learning summary category is selected as a second category of knowledge points;

[0013] Matching the first category of knowledge points with the candidate knowledge point list based on the semantic similarity of word vectors;

[0014] Matching the second category of knowledge points with the knowledge point dictionary corresponding to the candidate knowledge point list;

[0015] The knowledge points in the candidate knowledge point list that successfully match the first category knowledge points and the second category knowledge points are deemed to have passed the verification, and a target knowledge point list is obtained.

[0016] Furthermore, matching the first category of knowledge points with the candidate knowledge point list based on semantic similarity of word vectors includes:

[0017] The weighted sum and average of the word vectors in the first category of knowledge points are used to obtain the corresponding first sentence vector;

[0018] The weighted sum and average of the word vectors in the candidate knowledge point list are used to obtain the corresponding second sentence vector;

[0019] Calculate the cosine similarity between the first sentence vector and the second sentence phasor;

[0020] When the cosine similarity value is greater than a preset threshold, it is recorded that the knowledge point is matched successfully.

[0021] Furthermore, matching the second category of knowledge points with the knowledge point dictionary corresponding to the candidate knowledge point list includes:

[0022] Obtaining a knowledge point dictionary corresponding to the candidate knowledge point list;

[0023] Matching the second category of knowledge points with the candidate knowledge point list based on a maximum backward matching method;

[0024] When the corresponding keyword after matching is greater than 1, it is recorded that the knowledge point is matched successfully.

[0025] Further preferably, the multi-label text classification model based on neural network is modeled using TextCNN, and the last layer of the model established using TextCNN is calculated using a sigmoid function.

[0026] Optionally, the method further comprises:

[0027] Using the obtained target knowledge point list as training samples to train the neural network-based multi-label text classification model;

[0028] The trained neural network-based multi-label text classification model is used for text recognition in a preset application scenario.

[0029] Preferably, the acquiring and preprocessing of the courseware to obtain the target text includes:

[0030] Get courseware;

[0031] Converting the courseware into a computer-readable format;

[0032] Extract text from the converted courseware to obtain the target text.

[0033] A second aspect of the present invention provides a courseware knowledge point classification device based on a combination of multi-label classification and knowledge base matching, the device comprising:

[0034] The acquisition module is used to acquire the courseware and preprocess it to obtain the target text;

[0035] A sorting module, used to sort the text according to the coordinate axis of the target text to obtain a paginated text sequence;

[0036] The concatenation module is used to concatenate all text sequences into a long text;

[0037] A multi-label module, used for inputting the long text into a multi-label text classification model based on a neural network for multi-label classification to obtain a candidate knowledge point list;

[0038] The matching verification module is used to verify the candidate knowledge point list based on knowledge base matching to obtain a target knowledge point list.

[0039] A third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the following steps:

[0040] Obtain the courseware and preprocess it to obtain the target text;

[0041] Sort the text according to the coordinate axis of the target text to obtain a paginated text sequence;

[0042] Put all text sequences together into a long text;

[0043] Inputting the long text into a multi-label text classification model based on a neural network for multi-label classification to obtain a candidate knowledge point list;

[0044] The candidate knowledge point list is verified based on knowledge base matching to obtain a target knowledge point list.

[0045] A fourth aspect of the present invention provides a computer program product, comprising a computer program, which, when executed by a processor, implements the following steps:

[0046] Obtain the courseware and preprocess it to obtain the target text;

[0047] Sort the text according to the coordinate axis of the target text to obtain a paginated text sequence;

[0048] Put all text sequences together into a long text;

[0049] Inputting the long text into a multi-label text classification model based on a neural network for multi-label classification to obtain a candidate knowledge point list;

[0050] The candidate knowledge point list is verified based on knowledge base matching to obtain a target knowledge point list.

[0051] The beneficial effects of the present application are as follows: the present application sorts the text according to the coordinate axis of the target text to obtain a paginated text sequence, assembles all the text sequences into a long text, inputs the long text into a multi-label text classification model based on a neural network for multi-label classification, obtains a candidate knowledge point list, verifies the candidate knowledge point list based on knowledge base matching, obtains a target knowledge point list, reduces reliance on manual operations, and reduces work costs. In particular, in the paginated text sequence, the knowledge point definition class is selected as the first category of knowledge points, and the learning goal or learning summary class is selected as the second category of knowledge points, the first category of knowledge points is matched with the candidate knowledge point list based on the semantic similarity of the word vector, the second category of knowledge points is matched with the knowledge point dictionary corresponding to the candidate knowledge point list, and the knowledge points in the candidate knowledge point list that successfully match the first category of knowledge points and the second category of knowledge points are regarded as having passed the verification, and the target knowledge point list is obtained, which realizes a more accurate classification and recognition of knowledge points and improves work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The accompanying drawings, which constitute a part of the specification, illustrate embodiments of the present application and, together with the description, serve to explain the principles of the present application.

[0053] The present application can be more clearly understood from the following detailed description with reference to the accompanying drawings, in which:

[0054] Figure 1 A schematic diagram of the steps of a courseware knowledge point classification method based on a combination of multi-label classification and knowledge base matching in an exemplary embodiment of the present application is shown;

[0055] Figure 2 A flow chart of a courseware knowledge point classification method based on a combination of multi-label classification and knowledge base matching in an exemplary embodiment of the present application is shown;

[0056] Figure 3 A schematic diagram of a multi-label text classification model based on a neural network in an exemplary embodiment of the present application is shown;

[0057] Figure 4 A schematic diagram showing the combination of TextCNN and a secondary calibration model in an exemplary embodiment of the present application is shown;

[0058] Figure 5 A schematic diagram of the structure of a courseware knowledge point classification method and device based on a combination of multi-label classification and knowledge base matching according to an exemplary embodiment of the present application is shown;

[0059] Figure 6 A schematic diagram of the structure of a computer device provided by an exemplary embodiment of the present application is shown;

[0060] Figure 7 A schematic diagram of a storage medium provided by an exemplary embodiment of the present application is shown. DETAILED DESCRIPTION

[0061] Hereinafter, embodiments of the present application will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present application. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present application. It is obvious to those skilled in the art that the present application can be implemented without one or more of these details. In other examples, in order to avoid confusion with the present application, some technical features known in the art are not described.

[0062] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should also be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of the features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0063] Now, exemplary embodiments according to the present application will be described in more detail with reference to the accompanying drawings. However, these exemplary embodiments may be implemented in a variety of different forms and should not be construed as being limited to the embodiments described herein. The accompanying drawings are not drawn to scale, and certain details may be magnified and certain details may be omitted for the purpose of clear expression. The shapes of the various regions and layers shown in the figures and the relative sizes and positional relationships therebetween are merely exemplary, and may deviate in practice due to manufacturing tolerances or technical limitations, and those skilled in the art may additionally design regions / layers with different shapes, sizes, and relative positions according to actual needs.

[0064] The following is attached to the instruction manual Figure 1-7 Several embodiments are given to describe exemplary implementations according to the present application. It should be noted that the following application scenarios are only shown to facilitate understanding of the spirit and principles of the present application, and the implementations of the present application are not limited in this regard. On the contrary, the implementations of the present application can be applied to any applicable scenario.

[0065] Embodiment 1:

[0066] In this embodiment, a courseware knowledge point classification method based on multi-label classification and knowledge base matching is provided. Figure 1 As shown, including:

[0067] S1, obtain the courseware and preprocess it to obtain the target text;

[0068] S2, sorting the text according to the coordinate axis of the target text to obtain a paginated text sequence;

[0069] S3, assemble all text sequences into a long text;

[0070] S4, inputting the long text into a multi-label text classification model based on a neural network for multi-label classification to obtain a candidate knowledge point list;

[0071] S5. Verify the candidate knowledge point list based on knowledge base matching to obtain a target knowledge point list.

[0072] Specifically, the candidate knowledge point list is verified based on knowledge base matching to obtain a target knowledge point list, including: screening knowledge point definition classes in the paginated text sequence as the first category of knowledge points, and screening learning objectives or learning summary classes as the second category of knowledge points; matching the first category of knowledge points with the candidate knowledge point list based on semantic similarity of word vectors; matching the second category of knowledge points with the knowledge point dictionary corresponding to the candidate knowledge point list; and considering the knowledge points in the candidate knowledge point list that successfully match the first category of knowledge points and the second category of knowledge points as having passed the verification to obtain a target knowledge point list.

[0073] Furthermore, the first category of knowledge points and the candidate knowledge point list are matched based on the semantic similarity of word vectors, including: weighted summing and averaging the word vectors in the first category of knowledge points to obtain the corresponding first sentence vector; weighted summing and averaging the word vectors in the candidate knowledge point list to obtain the corresponding second sentence vector; calculating the cosine similarity between the first sentence vector and the second sentence vector; when the cosine similarity is greater than a preset threshold, it is recorded that the knowledge point is matched successfully. This means that each short sentence in the first category of recognition points is traversed and extracted in turn for similarity recognition.

[0074] Furthermore, the second category of knowledge points is matched with the knowledge point dictionary corresponding to the candidate knowledge point list, including: obtaining the knowledge point dictionary corresponding to the candidate knowledge point list; matching the second category of knowledge points with the candidate knowledge point list based on the maximum backward matching method; when the corresponding keyword after matching is greater than 1, it is recorded that the knowledge point is matched successfully.

[0075] Further preferably, the multi-label text classification model based on neural network is modeled using TextCNN, and the last layer of the model established using TextCNN is calculated using a sigmoid function.

[0076] Optionally, the method further includes: using the obtained target knowledge point list as training samples to train a neural network-based multi-label text classification model; and using the trained neural network-based multi-label text classification model for text recognition in a preset application scenario.

[0077] Preferably, obtaining the courseware and preprocessing it to obtain the target text includes: obtaining the courseware; converting the courseware into a computer-readable format; and extracting text from the converted courseware to obtain the target text.

[0078] Embodiment 2:

[0079] In this embodiment, a courseware knowledge point classification method based on multi-label classification and knowledge base matching is provided. The overall working process is as follows: Figure 2 , the specific steps are described as follows.

[0080] The first step is to obtain the courseware and preprocess it to get the target text.

[0081] There are many user-generated contents in the education industry, so you can get courseware on relevant courseware resource platforms. This application prefers the electronic courseware resource of Seewo Whiteboard. Obtaining courseware and preprocessing it to obtain the target text includes: obtaining courseware; converting the courseware into a computer-readable format; extracting text from the converted courseware to obtain the target text. The preprocessing here also includes removing invalid text in the electronic courseware, such as removing stop words, removing invalid web page information, etc.

[0082] The second step is to sort the text according to the coordinate axis of the target text to obtain a paged text sequence.

[0083] In a preferred embodiment, in order to obtain ordered information, it is necessary to sort the text elements in the electronic courseware. The specific method is to extract and sort from top to bottom and from left to right according to the coordinate (X, Y) value corresponding to each element in the entire area, and then obtain a paged text sequence.

[0084] The third step is to piece together all the text sequences into a long text.

[0085] The fourth step is to input the long text into a multi-label text classification model based on a neural network for multi-label classification to obtain a list of candidate knowledge points.

[0086] The "candidate knowledge point list" here is Figure 2 The biggest difference between the neural network multi-label classification model and the multi-classification model is the design of the output layer and the choice of the final loss function. The same thing is that the number of nodes in the last layer of the neural network is equal to the number of categories for both multi-label classification and multi-classification. The difference is that Figure 3 In the upper half, the last layer of the traditional multi-classification method uses the softmax activation function to extract the maximum value of all categories. The formula is:

[0087]

[0088] Among them, i represents the i-th category, x i represents the input of the i-th category in the last layer, and n represents the number of all categories. At this time, the sum of the probabilities of all categories is 1. However, for multi-label classification, since the number of outputs is not fixed, it is necessary to calculate the probability of each category, which is essentially to calculate n binary classifications. Therefore, this application has modified the neural network multi-label classification model and multi-classification model, such as Figure 3 As shown, the sigmoid function is used for calculation in the last layer, and the calculation formula is:

[0089]

[0090] Among them, i represents the i-th category, x i Represents the input of the i-th category in the last layer, and n represents the number of all categories.

[0091] In addition, the loss function for multi-classification usually uses a BCELoss. Multi-label classification requires calculating n BCELoss and then taking the average. Although using intermediate labels can greatly simplify the problem, direct use also has certain shortcomings. The most obvious shortcoming is that although in most cases, a courseware only corresponds to one chapter, there is not a strict one-to-one match between the courseware and the chapter.

[0092] Therefore, as a convertible implementation method, when there is a one-to-many relationship between courseware and chapters, we only need to summarize the knowledge points corresponding to multiple chapters as the knowledge point labels of the current courseware. However, when there is a many-to-one relationship between courseware and chapters, there will be many invalid knowledge points matched. In order to reduce the probability of this happening, in the process of extracting text, a model of secondary verification by page extraction is used, such as Figure 4 As shown in the figure, after dividing the text into a long text list by page, we first need to screen the content of each page. Although the content of the electronic courseware is very rich, in order to simplify the model, after a lot of data exploration, we divide the content into three categories: definition, learning objectives / summary, and invalid information. For the content division of the electronic courseware, keywords are used here. For example, if "key points of this chapter" or "summary of this chapter" appear, it can be considered that the current page is related to the summary / learning objectives; if words such as "definition" appear, it can be determined that the page is related to the definition of the knowledge point; pages that are neither summaries / learning objectives nor definitions are considered invalid pages and are not analyzed. After determining the approximate classification of the content of the page, it needs to be matched.

[0093] The fifth step is to verify the candidate knowledge point list based on knowledge base matching to obtain a target knowledge point list.

[0094] In a preferred embodiment, the candidate knowledge point list is verified based on knowledge base matching to obtain a target knowledge point list, including: screening knowledge point definition classes in the paginated text sequence as the first category of knowledge points, and screening learning objectives or learning summaries as the second category of knowledge points; matching the first category of knowledge points with the candidate knowledge point list based on the semantic similarity of word vectors; matching the second category of knowledge points with the knowledge point dictionary corresponding to the candidate knowledge point list; and considering the knowledge points in the candidate knowledge point list that successfully match the first category of knowledge points and the second category of knowledge points as having passed the verification to obtain a target knowledge point list. The "target knowledge point list" here refers to Figure 2 The final knowledge point labels shown in .

[0095] Furthermore, the first category of knowledge points and the candidate knowledge point list are matched based on the semantic similarity of word vectors, including: weighted summing and averaging the word vectors in the first category of knowledge points to obtain the corresponding first sentence vector; weighted summing and averaging the word vectors in the candidate knowledge point list to obtain the corresponding second sentence vector; calculating the cosine similarity between the first sentence vector and the second sentence vector; when the cosine similarity is greater than a preset threshold, it is recorded that the knowledge point is matched successfully. This means that each short sentence in the first category of recognition points is traversed and extracted in turn for similarity recognition.

[0096] Furthermore, the second category of knowledge points is matched with the knowledge point dictionary corresponding to the candidate knowledge point list, including: obtaining the knowledge point dictionary corresponding to the candidate knowledge point list; matching the second category of knowledge points with the candidate knowledge point list based on the maximum backward matching method; when the corresponding keyword after matching is greater than 1, it is recorded that the knowledge point is matched successfully. It is worth noting that because this step is not used for global matching, but for further screening of the candidate label set obtained by the existing multi-label classification method, in the matching process, only the labels on the candidate label set are matched.

[0097] It is worth noting that in natural language tasks, manual resource annotation of data sets is often required. If electronic courseware is directly annotated manually, the cost will be very high. Unlike ordinary short texts, the average number of pages of an electronic courseware is more than 20 pages. The annotator needs to read 20 pages of electronic courseware before annotating the labels. This process will consume huge human resources. At the same time, once the knowledge point system is replaced in this way, all annotations need to be re-annotated, which is not conducive to product iteration and updating. Therefore, in this application, the labels of electronic courseware->knowledge points will not be directly annotated, but the electronic courseware->chapter system will be annotated first, and then the final annotation data will be constructed through the chapter system->knowledge points.

[0098] Embodiment 3:

[0099] This embodiment provides a courseware knowledge point classification device based on the combination of multi-label classification and knowledge base matching. Figure 5 As shown, the device comprises:

[0100] The acquisition module 501 is used to acquire the courseware and pre-process it to obtain the target text;

[0101] A sorting module 502, used to sort the text according to the coordinate axis of the target text to obtain a paginated text sequence;

[0102] A concatenation module 503 is used to concatenate all text sequences into a long text;

[0103] A multi-label module 504 is used to input the long text into a multi-label text classification model based on a neural network for multi-label classification to obtain a candidate knowledge point list;

[0104] The matching verification module 505 is used to verify the candidate knowledge point list based on knowledge base matching to obtain a target knowledge point list.

[0105] Among them, the multi-label text classification model based on neural network is modeled using the TextCNN method, and the last layer of the model built using the TextCNN method is calculated using the sigmoid function.

[0106] Please refer to the following Figure 6 , which shows a schematic diagram of a computer device provided by some embodiments of the present application. Figure 6 As shown, the computer device 2 includes: a processor 200, a memory 201, a bus 202 and a communication interface 203, and the processor 200, the communication interface 203 and the memory 201 are connected via the bus 202; the memory 201 stores a computer program that can be run on the processor 200, and when the processor 200 runs the computer program, it executes the courseware knowledge point classification method based on the combination of multi-label classification and knowledge base matching provided in any of the aforementioned embodiments of the present application, and the computer device can be a computer device with a touch-sensitive display.

[0107] The memory 201 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 203 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used.

[0108] The bus 202 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The memory 201 is used to store programs, and the processor 200 executes the programs after receiving the execution instruction. The courseware knowledge point classification method based on the combination of multi-label classification and knowledge base matching disclosed in any implementation method of the aforementioned embodiment of the present application may be applied to the processor 200, or implemented by the processor 200.

[0109] The processor 200 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 200. The above processor 200 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a readily available programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as a hardware decoding processor to be executed, or the hardware and software modules in the decoding processor can be executed. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 201, and the processor 200 reads the information in the memory 201 and completes the steps of the above method in combination with its hardware.

[0110] The computer device provided in the embodiment of the present application and the courseware knowledge point classification method based on the combination of multi-label classification and knowledge base matching provided in the embodiment of the present application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented therein.

[0111] The present application also provides a computer-readable storage medium corresponding to the courseware knowledge point classification method based on multi-label classification and knowledge base matching provided in the above embodiment. Please refer to Figure 7 , Figure 7 The computer-readable storage medium shown is a CD 30 on which a computer program (ie, a program product) is stored. When the computer program is executed by a processor, the courseware knowledge point classification method based on the combination of multi-label classification and knowledge base matching provided in any of the aforementioned embodiments will be executed.

[0112] In addition, examples of the computer-readable storage medium may also include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other optical or magnetic storage media, which are not listed here one by one.

[0113] The computer-readable storage medium provided in the above-mentioned embodiments of the present application and the method for allocating channels for quantum key distribution in a space-division multiplexed optical network provided in the embodiments of the present application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the application programs stored therein.

[0114] The present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of a courseware knowledge point classification method based on a combination of multi-label classification and knowledge base matching provided in any of the aforementioned embodiments, the steps of the method including: obtaining the courseware and preprocessing it to obtain a target text; sorting the text according to the coordinate axis of the target text to obtain a paged text sequence; concatenating all text sequences into a long text; inputting the long text into a multi-label text classification model based on a neural network for multi-label classification to obtain a candidate knowledge point list; verifying the candidate knowledge point list based on knowledge base matching to obtain a target knowledge point list.

[0115] It should be noted that the algorithms and displays provided herein are not inherently related to any particular computer, virtual device or other device. Various general devices can also be used together with the teachings based thereon. According to the above description, it is obvious that the structure required for constructing such devices is constructed. In addition, the present application is not directed to any specific programming language. It should be understood that the content of the present application described herein can be implemented using various programming languages, and the above description of specific languages ​​is to disclose the best implementation mode of the present application. In the specification provided herein, a large number of specific details are described. However, it is understood that the embodiments of the present application can be practiced without these specific details. In some instances, known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this specification.

[0116] Similarly, it should be understood that in order to streamline the present application and aid in understanding one or more of the various inventive aspects, in the above description of the exemplary embodiments of the present application, the various features of the present application are sometimes grouped together into a single embodiment, figure, or description thereof. However, this disclosed method should not be interpreted as reflecting the following intention: the claimed application requires more features than the features explicitly recited in each claim. More specifically, as reflected in the claims below, the inventive aspects lie in less than all the features of the single embodiment disclosed above. Therefore, the claims that follow the specific embodiment are hereby expressly incorporated into the specific embodiment, with each claim itself serving as a separate embodiment of the present application.

[0117] Those skilled in the art will appreciate that the modules in the devices in the embodiments can be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments can be combined into one module or unit or component, and in addition they can be divided into multiple submodules or subunits or subassemblies. Except that at least some of such features and / or processes or units are mutually exclusive, all features disclosed in this specification and all processes or units of any method or device disclosed in this manner can be combined in any combination. Unless otherwise clearly stated, each feature disclosed in this specification can be replaced by an alternative feature providing the same, equivalent or similar purpose.

[0118] The various component embodiments of the present application can be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It should be understood by those skilled in the art that a microprocessor or digital signal processor (DSP) can be used in practice to implement some or all functions of some or all components in the creation device of the virtual machine according to the embodiment of the present application. The present application can also be implemented as a device or device program for executing part or all of the methods described herein. The program implementing the present application can be stored on a computer-readable medium, or can have the form of one or more signals. Such a signal can be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0119] The above is only a preferred specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A courseware knowledge point classification method based on the combination of multi-label classification and knowledge base matching, characterized in that: The method comprises: Obtain the courseware and preprocess it to obtain the target text; Sort the text according to the coordinate axis of the target text to obtain a paginated text sequence; Put all text sequences together into a long text; Inputting the long text into a multi-label text classification model based on a neural network for multi-label classification to obtain a candidate knowledge point list; Verifying the candidate knowledge point list based on knowledge base matching to obtain a target knowledge point list; The candidate knowledge point list is verified based on knowledge base matching to obtain a target knowledge point list, including: In the paginated text sequence, a knowledge point definition category is selected as a first category of knowledge points, and a learning goal or learning summary category is selected as a second category of knowledge points; Matching the first category of knowledge points with the candidate knowledge point list based on the semantic similarity of word vectors; Matching the second category of knowledge points with the knowledge point dictionary corresponding to the candidate knowledge point list; The knowledge points in the candidate knowledge point list that successfully match the first category knowledge points and the second category knowledge points are deemed to have passed the verification, and a target knowledge point list is obtained.

2. The courseware knowledge point classification method based on the combination of multi-label classification and knowledge base matching according to claim 1 is characterized in that: The matching of the first category of knowledge points with the candidate knowledge point list based on the semantic similarity of word vectors includes: The weighted sum and average of the word vectors in the first category of knowledge points are used to obtain the corresponding first sentence vector; The weighted sum and average of the word vectors in the candidate knowledge point list are used to obtain the corresponding second sentence vector; Calculate the cosine similarity between the first sentence vector and the second sentence phasor; When the cosine similarity value is greater than a preset threshold, it is recorded that the knowledge point is matched successfully.

3. The courseware knowledge point classification method based on the combination of multi-label classification and knowledge base matching according to claim 1 is characterized in that: The matching of the second category of knowledge points with the knowledge point dictionary corresponding to the candidate knowledge point list includes: Obtaining a knowledge point dictionary corresponding to the candidate knowledge point list; Matching the second category of knowledge points with the candidate knowledge point list based on a maximum backward matching method; When the corresponding keyword after matching is greater than 1, it is recorded that the knowledge point is matched successfully.

4. The courseware knowledge point classification method based on the combination of multi-label classification and knowledge base matching according to claim 1 is characterized in that: The multi-label text classification model based on neural network is modeled using TextCNN, and the last layer of the model established using TextCNN is calculated using the sigmoid function.

5. The courseware knowledge point classification method based on the combination of multi-label classification and knowledge base matching according to claim 4 is characterized in that: The method further comprises: Using the obtained target knowledge point list as training samples to train the neural network-based multi-label text classification model; The trained neural network-based multi-label text classification model is used for text recognition in a preset application scenario.

6. The courseware knowledge point classification method based on the combination of multi-label classification and knowledge base matching according to claim 1 is characterized in that: The step of obtaining the courseware and preprocessing it to obtain the target text includes: Get courseware; Converting the courseware into a computer-readable format; Extract text from the converted courseware to obtain the target text.

7. A courseware knowledge point classification device based on the combination of multi-label classification and knowledge base matching, characterized in that: The device comprises: The acquisition module is used to acquire the courseware and preprocess it to obtain the target text; A sorting module, used to sort the text according to the coordinate axis of the target text to obtain a paginated text sequence; The concatenation module is used to concatenate all text sequences into a long text; A multi-label module, used for inputting the long text into a multi-label text classification model based on a neural network for multi-label classification to obtain a candidate knowledge point list; A matching verification module, used to verify the candidate knowledge point list based on knowledge base matching to obtain a target knowledge point list; The candidate knowledge point list is verified based on knowledge base matching to obtain a target knowledge point list, including: In the paginated text sequence, a knowledge point definition category is selected as a first category of knowledge points, and a learning goal or learning summary category is selected as a second category of knowledge points; Matching the first category of knowledge points with the candidate knowledge point list based on the semantic similarity of word vectors; Matching the second category of knowledge points with the knowledge point dictionary corresponding to the candidate knowledge point list; The knowledge points in the candidate knowledge point list that successfully match the first category knowledge points and the second category knowledge points are deemed to have passed the verification, and a target knowledge point list is obtained.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images