System and method for automatically skipping documents based on model application

Abstract

The invention provides a system and a method for automatically skipping documents based on model application, and belongs to the field of intelligent retrieval and user login systems, the system comprises a target input medium, a target analysis model, an index model and a document storer, a user sends a target to the target analysis model through the target input medium to obtain a direction and a purpose to be indexed, and the direction and the purpose are indexed. According to the method, big data is subjected to networking indexing through an index model, an index result is obtained, sorted and then sent to a document storage, the index model inputs knowledge points into an SNMF matrix decomposition space, finds a hidden structure of a low-order representation capture word vector, improves clustering features, and optimizes a TF-IDF model through information gain; and indexing item combinations appearing in the networked database to obtain a retrieved webpage result.

Description

Technical Field

[0001] This specification relates to the field of intelligent retrieval, and particularly to a system and method for automatically jumping to documents based on model applications. Background Art

[0002] The exploration of course cases has received high attention in the academic community. However, for artificial intelligence education in colleges and universities, the teaching of artificial intelligence courses generally focuses on the imparting of artificial intelligence knowledge, often neglecting the leading role of curriculum value, resulting in the "knowledge-ization" of course cases; currently, the cases of artificial intelligence courses and other courses are all the same, and it is difficult for course cases to achieve the effect of deep integration, resulting in the "label-ization" of course cases; when carrying out artificial intelligence course case education, there is also the dilemma of "shallow integration and difficult integration" of course cases, and it is difficult to excavate the integration points between artificial intelligence knowledge and case content. Forcibly integrating some cases leads to the superficial integration of the two, which is likely to cause "course fatigue" among students.

[0003] Chinese Patent CN117493695A, an intelligent recommendation method and system for artificial intelligence course cases, the method includes: obtaining and screening M cases related to artificial intelligence courses to obtain N cases; annotating knowledge points for S cases among the N cases; training a multi-classifier according to the results; using the multi-classifier to annotate knowledge points for the remaining N - S cases among the N cases; using an open-source knowledge graph tool to construct a knowledge graph for the knowledge points of artificial intelligence courses; constructing a case model according to the N cases and the knowledge graph; training and testing the case model; and recommending cases corresponding to the target knowledge points according to the target knowledge points. This method can provide high-quality artificial intelligence course case recommendation services for teachers, with good course integration effects, solve the problems of shallow integration and difficult integration, not only reduce the workload of teachers, but also enrich the teaching content.

[0004] The above technology aims to collect cases by networking and indexing the similarity of knowledge points. The collected cases are mostly repetitive content. When retrieving through artificial intelligence, more often, the results are indexed by analyzing the retrieval requirements, which requires artificial intelligence to deeply analyze the requirements.

[0005] Therefore, it is necessary to provide a system and method for automatically jumping to documents based on model applications to achieve intelligent retrieval across the network, match teaching requirements, and obtain more accurate content. Summary of the Invention

[0006] One embodiment of this specification provides a system and method for automatically jumping to documents based on model applications. By identifying the target needs of users, combining the teaching progress and teaching purposes of users, case retrieval across the network is performed to obtain suitable cases for users to select.

[0007] In some embodiments, for a system that automatically jumps to documents based on model applications, the user logs in to the system. The system includes a target input medium, a target analysis model, an indexing model, and a document storage. The user sends a target to the target analysis model through the target input medium, obtaining the indexing direction and purpose. The indexing model connects to the network to index big data, and after obtaining the indexing results, they are sorted and sent to the document storage.

[0008] Furthermore, the target input medium includes performing teaching matching on the target content input by the user. First, it obtains the user's teaching type, determines the teaching chapters and requirements, and then performs a first-level limitation on the target content. The target analysis model performs semantic decomposition and semantic expansion on the target content. Through semantic decomposition, knowledge points are extracted, and through semantic expansion, the target content is expanded synonymously, including matching synonymous predicate objects.

[0009] Furthermore, the indexing model includes querying the input text online, summarizing the query web page results. There is a three-round screening mechanism set in the indexing model, and the screened results are archived to the document storage.

[0010] Furthermore, the document storage saves the web page as a document and saves the web page title and retrieval logic as the label of this document.

[0011] Furthermore, the document storage divides the storage space according to the target content, and the document labels in each storage space are automatically updated and overwritten.

[0012] A method for automatically jumping to documents based on model applications includes: S1: Combining the user's teaching stage, analyzing the target content input by the user, including matching the target content with the teaching chapters, expanding the knowledge points of the target content according to the teaching chapters, and expanding the knowledge points based on the semantic knowledge graph; S2: The target analysis model includes adjusting the target content retrieval formula. By limiting the retrieval range of the target content within this teaching chapter, the retrieval target is a set of expanded knowledge points, and the retrieval suffix is a set of synonymous expansions of the target requirements, obtaining the retrieval web page results; S3: The indexing model follows a three-layer screening logic to sequentially eliminate duplicate web pages, Trojan horse web pages, and secondary login web pages; S4: Convert the eliminated web pages into a document format and store them in the same space, and save the web page title and retrieval formula as the label of this document.

[0013] Furthermore, the Trojan horse web page is an advertisement web page that jumps to a third party, and the secondary login web page is a web page that requires login to open the full text.

[0014] Further, the indexing model inputs knowledge points into the SNMF matrix factorization space, finds the hidden structure of the low-order representation to capture the word vectors, enhances the clustering features, and then optimizes the TF-IDF model through information gain to index the frequently occurring item combinations in the networked database, obtaining the retrieved web page results.

[0015] The beneficial effects of the present invention are as follows: 1. It can automatically combine the teaching progress of users, process the target requirements of users and then retrieve, obtaining more accurate teaching cases; 2. Index the target content through the network, construct the knowledge point features through the model, and obtain highly utilized teaching cases. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] This specification will be further described in the form of exemplary embodiments, and these exemplary embodiments will be described in detail through the drawings. These embodiments are not restrictive. In these embodiments, the same numbers represent the same structures, where: Figure 1 is a schematic diagram of the working principle shown in some embodiments of this specification; Figure 2 is a schematic diagram of the indexing model shown in some embodiments of this specification. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] To more clearly illustrate the technical solutions of the embodiments of this specification, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some examples or embodiments of this specification. For those of ordinary skill in the art, without creative efforts, this specification can also be applied to other similar scenarios based on these drawings. Unless obvious from the language context or otherwise stated, the same reference numerals in the drawings represent the same structure or operation.

[0018] It should be understood that the "system", "device", "unit" and / or "module" used herein is a method for distinguishing different components, elements, parts, portions or assemblies at different levels. However, if other words can achieve the same purpose, the said words can be replaced by other expressions.

[0019] As shown in this specification and the claims, unless the context clearly indicates an exception, words such as "a", "an", "one" and / or "the" are not specifically singular and may also include the plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of the clearly identified steps and elements, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0020] Flowcharts are used in this specification to illustrate the operations performed by the systems according to the embodiments of this specification. It should be understood that the preceding or subsequent operations do not necessarily have to be executed precisely in sequence. Instead, the steps can be processed in reverse order or simultaneously. Also, other operations can be added to these processes, or one or several steps can be removed from these processes.

[0021] Embodiment: Please refer to Figure 1 , where the user logs in to the system. The system includes a target input medium, a target analysis model, an index model, and a document storage. The user sends the target to the target analysis model through the target input medium, obtaining the indexing direction and purpose. The index model indexes big data through the network, obtains the indexed results, and sends them to the document storage after sorting. Among them, the index model inputs knowledge points into the SNMF matrix decomposition space, finds the hidden structure of the low-order representation to capture word vectors, enhances the clustering characteristics, and then optimizes the TF-IDF model through information gain to index the frequently occurring item combinations in the networked database, obtaining the retrieved web page results.

[0022] The target input medium includes performing teaching matching on the target content input by the user. First, it obtains the teaching type of the user, determines the teaching chapter and requirements, and then performs a first limitation on the target content. The target analysis model performs semantic decomposition and semantic extension on the target content. Knowledge points are extracted through semantic decomposition, and the target content is synonymously extended through semantic extension, including matching synonymous predicate objects.

[0023] The index model includes querying the input text through the network, summarizing the query web page results. There are three rounds of screening mechanisms set in the index model, and the screened results are archived to the document storage.

[0024] The document storage saves the web page as a document and saves the web page title and retrieval logic as the document label. The document storage divides the storage space according to the target content, and the document labels in each storage space are automatically updated and overwritten. Including, S1: Combining the teaching stage of the user, analyzing the target content input by the user, including matching the target content with the teaching chapter, expanding the knowledge points of the target content according to the teaching chapter, and expanding the knowledge points based on the semantic knowledge graph; S2: The target analysis model includes adjusting the target content retrieval formula, limiting the retrieval range of the target content within this teaching chapter, retrieving the set of knowledge points after expansion as the target, and retrieving the synonymous expansion set with the target requirements as the suffix, obtaining the retrieved web page results; S3: The index model follows a three-layer screening logic to sequentially eliminate duplicate web pages, Trojan horse web pages, and secondary login web pages; S4: Convert the filtered web pages into document format and store them in the same space. Save the web page titles and search expressions as labels for these documents.

[0025] In S3, the Trojan web pages are advertising web pages that redirect to third parties, and the secondary login web pages are web pages that require login to view the full text. The indexing model inputs knowledge points into the SNMF matrix factorization space to find the hidden structure that captures word vectors with low order, enhancing the clustering features. Then, it optimizes the TF-IDF model through information gain, indexes the combination of items that frequently appear in the online database, and obtains the retrieved web page results. The information gain optimized TF-IDF model is used to calculate the keyword weights in the category distribution. It constructs a word similarity matrix through frequent word sets and applies the symmetric non-negative matrix factorization technique to expand the feature space. By calculating the support and confidence of each word in different categories, it identifies the frequent word sets with the same category trend. Through the above processing, not only the word frequency is considered, but also the information of the knowledge point category distribution can be better reflected, enhancing the keyword weights of each category.

[0026] The usage steps of the AI large model for generating ideological and political education cases in college courses are as follows: Step 1: User registration and login Register an account: Users access the platform and create a personal or institutional account.

[0027] Log in to the system: Use the registered account information to log in to the system.

[0028] Step 2: Define ideological and political teaching objectives Select a course: Select or enter relevant course information.

[0029] Determine teaching objectives: Determine the specific objectives of ideological and political education based on the course content.

[0030] Step 3: Customize case requirements Enter keywords: Enter keywords or topics related to ideological and political issues.

[0031] Set filtering conditions: Set filtering conditions such as time, location, type, etc. for the cases according to needs.

[0032] Step 4: AI-driven case retrieval Start the search: The system starts the case search according to the objectives and conditions defined by the user.

[0033] Case recommendation: The AI model analyzes the database and recommends relevant cases.

[0034] Step 5: Case content review and selection Review cases: Users browse the cases recommended by AI for a preliminary review.

[0035] Select cases: Select cases that meet the teaching needs.

[0036] Step 6: Case Document Generation Generate Document: The system generates a detailed document based on the selected case.

[0037] Edit and Customize: Users can edit and customize the case document to meet specific needs.

[0038] Step 7: Case and Knowledge Base Association Knowledge Base Matching: The system automatically associates the case with the knowledge points and skills in the knowledge base.

[0039] Manual Association: Users can manually adjust or add associations to ensure the fit between the case and the teaching content.

[0040] Step 8: Case Application and Teaching Integrate into the Course: Integrate the case document into the course teaching plan.

[0041] Classroom Teaching: Use the case for teaching in the classroom to guide students to discuss and think.

[0042] Step 9: Collect Feedback and Evaluation Teaching Feedback: Collect feedback from students and teachers on the use of the case.

[0043] Effect Evaluation: Evaluate the effect and impact of the case in teaching.

[0044] Step 10: Case and Knowledge Base Update Update Case: Update the case content or replace the case according to the feedback.

[0045] Knowledge Base Maintenance: Update the knowledge base to ensure that the knowledge points and skills are in line with the latest teaching requirements.

[0046] Step 11: System Usage Feedback Feedback System Issues: Users feedback problems or improvement suggestions encountered during system usage.

[0047] System Optimization: The development team conducts system optimization and upgrades based on the feedback.

[0048] Step 12: Continuous Iteration and Improvement Continuous Iteration: Regularly update the system functions to improve the user experience.

[0049] Technology Upgrade: Keep up with the latest AI technologies and continuously upgrade the case generation tool.

[0050] Through these steps, college teachers can efficiently use this tool to generate and apply ideological and political cases, enhancing the ideological and political education effect of the course. At the same time, the continuous iteration and optimization of the system will ensure that the tool always meets the latest requirements of educational practice.

[0051] Step 1: Integrate Ideological and Political Direction and Issues 1. Enumerate ideological and political education goals: Enumerate the directions and purposes of education, such as "cultivating social responsibility", "awareness of the rule of law", etc., which can be customized and used as the concept set and labels for model training.

[0052] 2. Collect ideological and political issues: Collect issues and topics related to ideological and political education, such as historical events, social phenomena, moral dilemmas, etc.

[0053] 3. Define case criteria: Establish the selection criteria for cases to ensure that the cases match the ideological and political education goals and issues.

[0054] 4. Professional setting: Conduct professional setting to facilitate the professional classification and screening of the found ideological and political cases.

[0055] Step 2: Search for Ideological and Political Cases through the AI Large Model 1. Build an AI model: Train the existing AI large model to understand the main idea and connotation of ideological and political cases.

[0056] 2. Data collection: Collect a large number of materials related to ideology and politics, such as texts, videos, audios, etc.

[0057] 3. Case screening: Use the AI model to analyze the materials and screen out the cases that meet the criteria.

[0058] 4. Case verification: Ensure the quality of the cases screened by AI through manual review.

[0059] 5. Output document: Organize the screened cases into a document, including case descriptions, analyses, and discussions, etc.

[0060] Step 3: Associate Case Documents with Knowledge and Skills 1. Knowledge base construction: Establish a knowledge base containing course knowledge and skill requirements.

[0061] 2. Case annotation: Annotate the case documents to clarify the association between the cases and the knowledge points and skills in the knowledge base.

[0062] 3. Intelligent matching: Develop an intelligent matching system to automatically associate the cases with the relevant content in the knowledge base.

[0063] 4. Knowledge base update: Continuously update and enrich the knowledge base according to new cases.

[0064] Step 4: Integrate into the Knowledge Base and Apply 1. Knowledge base integration: Integrate ideological and political cases into the knowledge base to form structured data.

[0065] 2. Case Application: During the teaching process, relevant cases are extracted from the knowledge base according to the course content and students' needs.

[0066] 3. Teaching Feedback: Collect feedback during the teaching process and evaluate the application effect of the cases.

[0067] In summary, through the above-mentioned goal sorting and matching, and the extended indexing of knowledge points, the teaching cases that best meet the user's needs are obtained.

[0068] The basic concepts have been described above. Obviously, for those skilled in the art, the above detailed disclosure is only an example and does not constitute a limitation of this specification. Although not explicitly stated here, those skilled in the art may make various modifications, improvements, and corrections to this specification. Such modifications, improvements, and corrections are suggested in this specification, so such modifications, improvements, and corrections still belong to the spirit and scope of the exemplary embodiments of this specification.

[0069] At the same time, this specification uses specific terms to describe the embodiments of this specification. Such as "one embodiment", "an embodiment", and / or "some embodiments" mean a certain feature, structure, or characteristic related to at least one embodiment of this specification. Therefore, it should be emphasized and noted that the "one embodiment" or "an embodiment" or "an alternative embodiment" mentioned twice or more at different positions in this specification does not necessarily refer to the same embodiment. In addition, certain features, structures, or characteristics in one or more embodiments of this specification can be appropriately combined.

[0070] In addition, unless clearly stated in the claims, the order of the processing elements and sequences, the use of numbers and letters, or the use of other names in this specification is not used to limit the order of the processes and methods of this specification. Although various examples are discussed in the above disclosure for some currently considered useful invention embodiments, it should be understood that such details only serve the purpose of illustration, and the appended claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that conform to the essence and scope of the embodiments of this specification. For example, although the system components described above can be implemented by hardware devices, they can also be implemented only through software solutions, such as installing the described system on existing servers or mobile devices.

[0071] Similarly, it should be noted that in order to simplify the expression of the disclosure of this specification and thus help the understanding of one or more invention embodiments, in the previous description of the embodiments of this specification, sometimes multiple features are merged into one embodiment, drawing, or description thereof. However, this disclosure method does not mean that the features required by the object of this specification are more than those mentioned in the claims. In fact, the features of the embodiment are less than all the features of the single embodiment disclosed above.

[0072] Finally, it should be understood that the embodiments described in this specification are only used to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, by way of example and not limitation, alternative configurations of the embodiments of this specification may be regarded as consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly presented and described in this specification.

Claims

1. A system for automatically jumping to a document based on model application, characterized in that, The user logs in to the system. The system includes a target input medium, a target analysis model, an indexing model, and a document storage. The user sends the target to the target analysis model through the target input medium, obtaining the indexing direction and purpose. The indexing model connects to the network to index big data, and after organizing the indexing results, sends them to the document storage. Among them, the indexing model inputs knowledge points into the SNMF matrix decomposition space, finds the hidden structure of the low-order representation to capture the word vector, enhances the clustering characteristics, and then optimizes the TF-IDF model through information gain to index the item combinations that appear in the networked database, obtaining the retrieved web page results.

2. The system for automatically jumping to a document based on model application according to claim 1, wherein The target input medium includes performing teaching matching on the target content input by the user. First, it obtains the teaching type of the user, determines the teaching chapter and requirements, and then performs a primary limitation on the target content. The target analysis model performs semantic decomposition and semantic expansion on the target content, extracts knowledge points through semantic decomposition, and performs synonymous expansion on the target content through semantic expansion, including matching synonymous predicate objects.

3. The system for automatically jumping to a document based on model application according to claim 2, wherein, The indexing model includes querying the input text through the network, summarizing the query web page results. There are three rounds of screening mechanisms set in the indexing model, and the screened results are archived to the document storage.

4. The system for automatically jumping to a document based on model application according to claim 3, wherein The document storage saves the web page as a document and stores the web page title and retrieval logic as the label of this document.

5. The system for automatically jumping to a document based on model application according to claim 4, wherein The document storage divides the storage space according to the target content, and the document labels in each storage space are automatically updated and overwritten.

6. A method for automatically jumping to a document based on a model application, applied to the system for automatically jumping to a document based on a model application according to claim 5, characterized in that, Including, S1: Combining the teaching stage of the user, analyzing the target content input by the user, including matching the target content with the teaching chapter, expanding the knowledge points of the target content according to the teaching chapter, and expanding the knowledge points based on the semantic knowledge graph; S2: The target analysis model includes adjusting the target content retrieval formula, restricting the retrieval range of the target content within this teaching chapter, retrieving the set of expanded knowledge points as the target, and retrieving the suffix as the synonymous expansion set of the target requirements, obtaining the retrieved web page results; S3: The indexing model follows a three-layer screening logic to sequentially eliminate duplicate web pages, Trojan horse web pages, and secondary login web pages; S4: Convert the eliminated web pages into document format and store them in the same space, and store the web page title and retrieval formula as the label of this document.

7. The method for automatically jumping to a document based on model application according to claim 6, wherein In S3, the Trojan horse web page is an advertising web page that jumps to a third party, and the secondary login web page is a web page that requires login to open the full text.

8. The method for automatically jumping to a document based on model application according to claim 7, wherein The indexing model inputs knowledge points into the SNMF matrix decomposition space, finds the hidden structure of the low-order representation to capture the word vector, enhances the clustering characteristics, and then optimizes the TF-IDF model through information gain to index the item combinations that appear in the networked database, obtaining the retrieved web page results.

Images