Document display methods, systems, devices and media

Abstract

This application discloses a document display method, system, device, and medium, applied in the field of computer technology, to address the problem of low efficiency in compliance review of administrative normative supporting documents. The method includes: responding to a viewing operation of an administrative normative supporting document, displaying a summary map display area and a supporting document display area in a split-screen manner; presenting the administrative normative supporting document in the supporting document display area and a keyword summary map in the summary map display area; responding to a viewing operation of the original text of the keyword summary information in the keyword summary map, determining a first positioning label corresponding to the keyword summary information and a second positioning label associated with the first positioning label from a positioning label database; and, based on the second positioning label, jumping to the original text content corresponding to the keyword summary information in the supporting document display area for presentation, thereby improving the efficiency of compliance review through split-screen display and jumping display based on positioning labels.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a document display method, system, device and medium. Background Technology

[0002] With the continuous development of computer applications and electronic information technology, compliance review support systems specifically developed for administrative departments have been gradually put into use, greatly facilitating the review of administrative normative certification documents by administrative departments.

[0003] Currently, the compliance review assistance system supports the display of each administrative normative supporting document provided by citizens, legal persons or other organizations to assist administrative departments in conducting compliance reviews. This method of displaying administrative normative supporting documents requires manual identification of key content related to the compliance review before the review can be conducted, resulting in low efficiency of the compliance review. Summary of the Invention

[0004] This application provides a document display method, system, device, and medium to address the problem of low efficiency in the compliance review of administrative regulatory certification documents in existing technologies. The technical solution provided in this application is as follows:

[0005] On the one hand, this application provides a document display method, including:

[0006] In response to the viewing operation of administrative normative supporting documents, the summary map display area and the supporting documents display area are displayed in a split screen. The supporting documents display area presents the administrative normative supporting documents, while the summary map display area presents a keyword summary map of the administrative normative supporting documents. The keyword summary map is obtained by extracting the original text content related to compliance review from the administrative normative supporting documents, generating keyword summary information for each original text content, and then structuring the keyword summary information of each original text content according to the hierarchical relationship and semantic relationship between the original text content.

[0007] In response to the operation of viewing the original text of the keyword summary information in the keyword summary map currently displayed in the summary map display area, the first positioning label corresponding to the keyword summary information and the second positioning label associated with the first positioning label are determined from the positioning label database; wherein, the first positioning label is used to represent the position information of the keyword summary information in the keyword summary map; the second positioning label is used to represent the position information of the keyword summary information in the original text of the administrative normative certification document;

[0008] Based on the second location label, the user is redirected within the supporting documents display area to view the original text of the keyword summary information in the administrative normative supporting documents.

[0009] Optionally, extract the relevant sections of the original text from the administrative regulatory certification documents, including:

[0010] A hierarchical analysis model is used to identify the hierarchical structure of administrative normative certification documents to obtain hierarchical structure information;

[0011] A location-based model is used to dynamically match hierarchical structured information with a compliance review terminology database to locate the position information of each paragraph related to compliance review within the hierarchical structured information.

[0012] A paragraph extraction model is adopted to extract the initial original text content of each paragraph related to compliance review from the administrative normative certification documents based on the position information of each paragraph. Then, based on a multi-dimensional filtering strategy, the redundancy of each paragraph's initial original text content is removed to obtain the original text content of each paragraph related to compliance review.

[0013] Optionally, keyword summary information can be generated for each paragraph of the original text, including:

[0014] A keyword extraction model is used to extract keywords from each paragraph of the original text.

[0015] Based on the keywords in each paragraph of the original text, keyword summary information is generated for each paragraph of the original text.

[0016] Optionally, the keyword summary information of each paragraph is structured according to the hierarchical and semantic relationships between the paragraphs, including:

[0017] Based on the hierarchical structured information of administrative normative certification documents, the hierarchical relationship between the original text content of each paragraph is determined;

[0018] Based on the semantic similarity between keywords in each paragraph of the original text, the semantic relationships between the paragraphs of the original text are determined.

[0019] Based on the hierarchical relationship between the original text segments, a main structure with keyword summary information as nodes is constructed, and based on the semantic relationship between the original text segments, a branch structure with keyword summary information as nodes is constructed to obtain a keyword summary mind map.

[0020] Optionally, after structuring the keyword summary information of each paragraph according to the logical relationship between the paragraphs, the method further includes:

[0021] Based on the position information of each keyword summary information in the keyword summary mind map, generate the first positioning label of each keyword summary information;

[0022] Based on the location information of the original text of each keyword summary information in the administrative normative certification document, a second positioning label of the original text of each keyword summary information is generated.

[0023] Each keyword summary information and its first location label are associated with the second location label of the original text content of each keyword summary information and stored in the location label database.

[0024] Optionally, the document display method provided in this application also includes:

[0025] In response to the auxiliary review of administrative normative supporting documents, the compliance review of administrative normative supporting documents is conducted in accordance with the compliance review method to obtain the compliance review results of the original text of each paragraph in the administrative normative supporting documents that is related to the compliance review; wherein, the compliance review method includes the compliance review rules corresponding to each compliance review clause;

[0026] Within the supporting documentation display area, a label box is displayed at each section of the original text, and the corresponding compliance review result and applicable compliance review clauses are presented within the label box.

[0027] Optionally, when conducting a compliance review of administrative normative supporting documents according to the compliance review method to obtain the compliance review results of the original text content of each paragraph related to the compliance review in the administrative normative supporting documents, it also includes:

[0028] Obtain the associated administrative normative certification documents;

[0029] In accordance with the compliance review method, the compliance review of the relevant administrative normative supporting documents is carried out to obtain the compliance review results of the original text of each paragraph in the relevant administrative normative supporting documents that are related to the compliance review;

[0030] Based on the compliance review results of the original text of each paragraph related to compliance review in the relevant administrative normative certification documents, a supplementary review is conducted on the compliance review results of the original text of each paragraph related to compliance review in the administrative normative certification documents, so as to obtain the final compliance review results of the original text of each paragraph related to compliance review in the administrative normative certification documents.

[0031] On the other hand, this application provides a document display system, including:

[0032] The split-screen display unit is used to respond to the viewing operation of administrative normative certification documents. It splits the display area into a summary map display area and a certification document display area. The certification document display area displays the administrative normative certification document, and the summary map display area displays a keyword summary map of the administrative normative certification document. The keyword summary map is obtained by extracting the original text content related to compliance review from the administrative normative certification document, generating keyword summary information for each original text content, and then structuring the keyword summary information of each original text content according to the logical relationship between the original text content.

[0033] The label query unit is used to respond to the operation of viewing the original text of the keyword summary information in the keyword summary map currently displayed in the summary map display area. It determines the first positioning label corresponding to the keyword summary information and the second positioning label associated with the first positioning label from the positioning label database. The first positioning label is used to represent the position information of the keyword summary information in the keyword summary map; the second positioning label is used to represent the position information of the keyword summary information in the original text of the administrative normative certification document.

[0034] The jump display unit is used to jump to the original text of the keyword summary information in the administrative normative certification document within the certification material display area based on the second positioning number.

[0035] On the other hand, this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described document display method.

[0036] On the other hand, this application provides a computer-readable storage medium that stores computer instructions, which, when executed by a processor, implement the above-described file display method.

[0037] The beneficial effects of this application are as follows:

[0038] This application presents administrative regulatory supporting documents in the supporting materials display area and keyword summary maps of these documents in the summary map display area. This allows for split-screen display of the administrative regulatory supporting documents and keyword summary maps, facilitating reviewers' quick understanding of the documents using the keyword summary maps. Furthermore, reviewers can view the full text of the keyword summary information in the currently displayed keyword summary map within the supporting materials display area, thus providing significant convenience for compliance review and improving its efficiency.

[0039] Other features and advantages of this application will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings. Attached Figure Description

[0040] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0041] Figure 1 This is a schematic diagram outlining the document display method in the embodiments of this application;

[0042] Figure 2 This is a schematic diagram of the split-screen display interface of the document and keyword summary mind map in the embodiments of this application;

[0043] Figure 3 This is a schematic diagram illustrating the generation process of keyword summary map and location labels in the embodiments of this application;

[0044] Figure 4 This is a functional structure diagram of the document display device in the embodiments of this application;

[0045] Figure 5 This is a schematic diagram of the hardware structure of the electronic device in the embodiments of this application. Detailed Implementation

[0046] To make the objectives, technical solutions, and beneficial effects of this application clearer, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0047] This application provides a document display method, applied to a document display system, see below. Figure 1 As shown, the general flow of the document display method provided in this application embodiment is as follows:

[0048] Step 101: In response to the viewing operation of administrative normative supporting documents, the summary map display area and the supporting documents display area are displayed in a split screen. The administrative normative supporting documents are presented in the supporting documents display area, and the keyword summary map of the administrative normative supporting documents is presented in the summary map display area. The keyword summary map is obtained by extracting the original text content related to compliance review from the administrative normative supporting documents, generating keyword summary information for each original text content, and then structuring the keyword summary information of each original text content according to the hierarchical relationship and semantic relationship between the original text content.

[0049] In this embodiment, administrative normative supporting documents are supporting documents submitted to administrative departments for review, such as tax-related supporting documents, qualification-related supporting documents, and intellectual property-related supporting documents. The document display system can be loaded into the administrative department's office system as an auxiliary tool. The document display system can detect various administrative normative supporting documents received by the office system and present them in a list format for review. When reviewers perform viewing operations on any administrative normative supporting document, such as clicking, double-clicking, or right-clicking and selecting the view option, the document display system responds to the viewing operation on the administrative normative supporting document. (See also...) Figure 2 As shown, the interface displays a split-screen summary map area and a supporting document display area. The supporting document display area shows the administrative regulatory supporting documents, while the summary map display area shows a keyword summary map of these documents. This split-screen presentation of the administrative regulatory supporting documents and their keyword summary maps allows reviewers to quickly grasp the documents and facilitates compliance review. Virtual buttons for marking, voice prompts, annotations, auxiliary review, modification, and sliders are provided above both the summary map and supporting document display areas, allowing reviewers to easily interact with the documents and keyword summary maps. Zooming and pagination are also supported within these areas, further enhancing review convenience. It's worth noting that the keyword summary map for the administrative regulatory supporting documents can be generated either when the document is detected or when a viewing operation is performed; there are no restrictions on this.

[0050] Step 102: In response to the operation of viewing the original text of the keyword summary information in the keyword summary map currently displayed in the summary map display area, determine the first positioning label corresponding to the keyword summary information and the second positioning label associated with the first positioning label from the positioning label database; wherein, the first positioning label is used to represent the position information of the keyword summary information in the keyword summary map; the second positioning label is used to represent the position information of the keyword summary information in the original text of the administrative normative certification document.

[0051] In this embodiment, each keyword summary in the keyword summary map is equipped with an event detector and node attributes (such as expand or collapse). Reviewers can perform different operations on the keyword summary information currently displayed in the summary map display area to achieve different presentations of that keyword summary information. For example, when a reviewer performs an operation such as clicking to view the original text, the document display system detects this operation through the event detector and triggers a query from the location label database for the first location label and its associated second location label. Similarly, when a reviewer performs an operation such as double-clicking to collapse or expand the keyword summary information, the document display system detects this operation through the event detector and triggers the collapse or expansion presentation of the keyword summary information. It is worth noting that the first location label and its associated second location label can be generated and stored during the generation of the keyword summary map.

[0052] Step 103: Based on the second positioning label, jump to the original text of the keyword summary information in the administrative normative certification document within the supporting materials display area.

[0053] In this embodiment, after retrieving the first location marker of the keyword summary information from the location marker database, the second location marker associated with the first location marker can be retrieved. Based on the second location marker, the user can jump to the original text of the keyword summary information in the administrative normative supporting document within the supporting document display area. This allows reviewers to view the original text of the keyword summary information in the currently displayed keyword summary map within the summary map display area, thus enabling them to jump to the original text of the keyword summary information in the administrative normative supporting document within the supporting document display area. This significantly facilitates compliance review and improves its efficiency.

[0054] Furthermore, in this embodiment, reviewers can also perform auxiliary review operations on administrative normative supporting documents, such as clicking the auxiliary review virtual button displayed above the supporting document display area. In response to these auxiliary review operations, the document display system performs a compliance review on the administrative normative supporting documents according to compliance review methods, such as the compliance review rules corresponding to each compliance review clause. This yields the compliance review results for each paragraph of the original text related to the compliance review within the supporting document display area. A label box is displayed at each paragraph of the original text, showing the corresponding compliance review result and applicable compliance review clause within the label box. In this way, by performing auxiliary review operations on the administrative normative supporting documents, reviewers can automatically conduct a compliance review of the documents and present the compliance review results and applicable compliance review clauses for each paragraph of the original text related to the compliance review. This achieves auxiliary review of the administrative normative supporting documents, providing a reference for reviewers and improving review efficiency. When conducting a compliance review of administrative normative supporting documents to obtain the compliance review results for the original text content related to the compliance review in the administrative normative supporting documents, the following methods may be used, but are not limited to:

[0055] The compliance review model, based on a pre-built compliance library of administrative normative legal documents (such as the Patent Agency Regulations and industry standards, the Enterprise Income Tax Law and supporting regulations), combined with custom review rule templates containing the corresponding compliance review rules for each compliance review clause (such as subject qualification and content compliance rules), automatically verifies the violations of each paragraph of the original text in the administrative normative supporting documents related to compliance review, and obtains the compliance review results of each paragraph of the original text in the administrative normative supporting documents related to compliance review.

[0056] In another embodiment, the document display system, in response to an auxiliary review operation on administrative normative supporting documents, performs a compliance review on the administrative normative supporting documents according to a compliance review method, such as according to the compliance review rules corresponding to each compliance review clause, to obtain the compliance review results of each paragraph of the original text related to the compliance review in the administrative normative supporting documents. Simultaneously, it obtains the associated administrative normative supporting documents and performs a compliance review on the associated administrative normative supporting documents according to a compliance review method, such as according to the compliance review rules corresponding to each compliance review clause, to obtain the associated... After reviewing the compliance review results of each paragraph of the original text related to compliance review in the administrative normative certification documents, a supplementary review is conducted on the compliance review results of each paragraph of the original text related to compliance review in the administrative normative certification documents, based on the compliance review results of the original text of each paragraph related to compliance review in the associated administrative normative certification documents. This results in the final compliance review results of each paragraph of the original text in the administrative normative certification documents. In the certification document display area, a label box is displayed at each paragraph of the original text, and the final compliance review results of the corresponding original text and the applicable compliance review clauses are presented in the label box. This approach not only automatically performs compliance reviews on the administrative regulatory certification document and presents the applicable compliance review clauses and corresponding results for each section of the original text related to the compliance review, providing a reference for reviewers, but also allows for supplementary reviews of the administrative regulatory certification document's compliance results using the compliance review results of related administrative regulatory certification documents, thereby improving the accuracy of the compliance review results. When obtaining related administrative regulatory certification documents, the following methods can be used, but are not limited to:

[0057] Step 1031: Using an improved TF-IDF model, employing part-of-speech weighting (e.g., nouns 1.0, verbs 0.8) and a sliding window mechanism (e.g., 5-7 keywords), identify core keywords with a co-occurrence frequency not lower than the co-occurrence frequency threshold (e.g., 3 times) in administrative normative supporting documents and each candidate administrative normative supporting document, and mark the logical relationships between each core keyword to generate a core keyword relationship graph corresponding to each administrative normative supporting document and each candidate administrative normative supporting document; wherein, candidate administrative normative supporting documents include other administrative normative supporting documents from the same reviewing entity, and / or, historical administrative normative supporting documents of the same document type from the same reviewing entity. Specifically as follows:

[0058] First, segment the administrative regulatory certification material file and each candidate administrative regulatory certification material file respectively, then remove the stop words (such as high-frequency meaningless words like "de", "le", etc.), and use a词性标注工具 (such as NLTK or Jieba) to mark the词性 of each remaining keyword (such as noun, verb). For example, a sentence "Artificial intelligence analyzes data" in the file is processed as [artificial intelligence (noun), analyze (verb), data (noun)].

[0059] Then, use the following formula to calculate the weighted term frequency of each keyword in the administrative regulatory certification material file and each candidate administrative regulatory certification material file respectively;

[0060] TFweighted(t,d)=TF(t,d)×POS_weight(t)

[0061] Among them, TFweighted(t,d) represents the weighted term frequency of keyword t in the administrative regulatory certification material file d; TF(t,d) represents the term frequency of keyword t in the administrative regulatory certification material file d, TF(t,d) = the number of occurrences of keyword t in the administrative regulatory certification material file d / the total number of keywords in the administrative regulatory certification material file d; POS_weight(t) represents the词性 weight of keyword t. For example, the noun weight is 1.0, the verb weight is 0.8, and the default weight for other词性 is 1.0, which can be customized or trained and learned according to the file type and file characteristics of the administrative regulatory certification material file.

[0062] Secondly, use the following formula to calculate the inverse document frequency of each keyword in the administrative regulatory certification material file and each candidate administrative regulatory certification material file respectively;

[0063] IDF(t)=log(total number of corpus files / (number of files containing keyword t + 1))

[0064] Among them, IDF(t) represents the inverse document frequency of keyword t; the total number of corpus files represents the total number of historical administrative regulatory certification material files of the same file type as the administrative regulatory certification material file. The corpus needs to contain a sufficient number of historical administrative regulatory certification material files to ensure statistical validity, for example, not less than 100; the number of files containing keyword t represents the total number of files containing keyword t in each historical administrative regulatory certification material file; 1 represents the smoothing parameter to avoid extreme situations such as the denominator being zero.

[0065] It should be noted that the Chinese terms "词性标注工具" and "词性" in the original text are not accurately translated as there may be more appropriate English expressions in the context of text processing, but following the rule of only translating the text while keeping the tags as they are, the above translation is provided. You may need to adjust the English expressions of relevant terms according to the actual situation.Secondly, based on the weighted word frequency and inverse document frequency of each keyword in the administrative normative certification documents and each candidate administrative normative certification documents, the following formula is used to calculate the importance value of each keyword in the administrative normative certification documents and each candidate administrative normative certification documents respectively. The keywords with the top K importance values ​​are selected and / or the keywords with importance values ​​lower than the second set threshold are filtered out to obtain all the core keywords in the administrative normative certification documents and each candidate administrative normative certification documents.

[0066] TF-IDFweighted(t,d)=TFweighted(t,d)×IDF(t)

[0067] Wherein, TF-IDFweighted(t,d) represents the importance value of keyword t; TFweighted(t,d) represents the weighted term frequency of keyword t; and IDF(t) represents the inverse document frequency of keyword t.

[0068] Subsequently, a sliding window was used to count the co-occurrence frequency of each core keyword pair in administrative normative certification documents and each candidate administrative normative certification document. The sliding window size can be set to 5-7 keywords. The administrative normative certification documents are traversed by sliding window. Within each sliding window, the co-occurrence frequency of each core keyword pair is counted. Based on the co-occurrence frequency of each core keyword and the total number of sliding windows, the co-occurrence frequency of each core keyword pair is calculated using the following formula.

[0069] Co-occurrence frequency = Number of times core keywords co-occur / Total number of sliding windows

[0070] Finally, based on the co-occurrence frequency of each core keyword pair in the administrative normative certification documents and each candidate administrative normative certification documents, the relationships between each core keyword in the administrative normative certification documents and each candidate administrative normative certification documents are marked. Based on the relationships between each core keyword in the administrative normative certification documents and each candidate administrative normative certification documents, a core keyword relationship graph (nodes are core keywords, and edge weights are co-occurrence frequencies) is generated for each administrative normative certification document and each candidate administrative normative certification document. Among them, core keyword pairs with high co-occurrence frequencies (e.g., > threshold 0.5) and no logical connectors are parallel relationships; core keyword pairs containing explicit logical words (e.g., "lead to" or "therefore") or implicit derivation chains in the co-occurrence window are causal relationships; and core keyword pairs with sequential co-occurrence (e.g., A→B→C appearing in consecutive windows) are progressive relationships.

[0071] Step 1032: Convert the administrative normative certification documents and each candidate administrative normative certification document into document-term frequency matrices (generated by CountVectorizer). Input the document-term frequency matrices corresponding to each administrative normative certification document and each candidate administrative normative certification document into the LDA model. Through the LDA model, perform topic distribution analysis on each administrative normative certification document and each candidate administrative normative certification document to obtain the potential topic distribution corresponding to each administrative normative certification document and each candidate administrative normative certification document. Among them, the potential topic distribution includes: topic-word distribution, that is, the top N keywords with the highest output probability for each topic (such as topic 1: "blockchain, encryption, distributed", etc.), and document-topic distribution, that is, the probability of a document belonging to each topic (such as 0.2 topic A, 0.8 topic B, etc.).

[0072] Step 1033: Based on the core keyword relationship graph and potential topic distribution of the administrative normative certification documents and each candidate administrative normative certification document, after obtaining the text feature data of the administrative normative certification documents and each candidate administrative normative certification document, calculate the text similarity (such as cosine similarity) between the administrative normative certification documents and each candidate administrative normative certification document based on the text feature data of the administrative normative certification documents and each candidate administrative normative certification document, and determine the candidate administrative normative certification documents with a text similarity not lower than the similarity threshold as the associated administrative normative certification documents of the administrative normative certification documents.

[0073] Furthermore, when conducting a compliance review of related administrative normative supporting documents in accordance with the compliance review method, and obtaining the compliance review results of the original text content of each paragraph related to the compliance review in the related administrative normative supporting documents, and when conducting a supplementary review of the compliance review results of the original text content of each paragraph related to the compliance review in the related administrative normative supporting documents, the following methods may be adopted, but are not limited to:

[0074] First, a compliance review model is adopted. Based on a pre-built compliance database of administrative normative legal documents (such as the Patent Agency Regulations and industry standards, the Enterprise Income Tax Law and supporting regulations), combined with custom rule templates (such as subject qualification and content compliance rules), the model automatically verifies the violations of related administrative normative supporting documents and whether these violations affect the compliance review results of the administrative normative supporting documents (such as the lack of subject qualification triggering the original text risk tracing), thus obtaining the first supplementary review result. Then, based on logical screening (such as consistency of time sequence) and contradiction analysis (such as the semantic opposition between "must" and "prohibited") rule templates, the model automatically identifies the conflict points between related administrative normative supporting documents (such as the contradiction between inspection batch data and construction logs), and automatically verifies whether these conflict points affect the compliance review results of the administrative normative supporting documents, thus obtaining the second supplementary review result.

[0075] Then, based on the first and second supplementary review results of the related administrative normative supporting documents output by the compliance review model, the compliance review knowledge graph is updated in real time using knowledge graph storage rules (such as entity relationship mapping). For example, when there is a violation in the related administrative normative supporting documents, the compliance review results of the original text content corresponding to the violation in the administrative normative supporting documents are automatically traced (such as the subsidiary's violation tracing back to the parent company's liability clause). This continues until the compliance review knowledge graph is updated based on the first and second supplementary review results of all related administrative normative supporting documents. Based on the updated compliance review knowledge graph, the final compliance review results of each section of the original text content related to compliance review in the administrative normative supporting documents are determined. The compliance review knowledge graph is constructed based on the compliance review results of each section of the original text content in the administrative normative supporting documents, as well as the first and second supplementary review results of each section of the original text content in each related administrative normative supporting documents.

[0076] In this embodiment, explicit local associations are captured through a co-occurrence matrix, global latent themes are revealed through LDA, and gaps in the compliance review of the original text are filled by associating the original text content with relevant information. This achieves collaborative review driven by AI and a rule base, automating and enhancing the review results and implementing closed-loop risk management, thereby further improving the accuracy of compliance review. This provides reviewers with valuable reference for the compliance review of administrative normative supporting documents, thus improving the efficiency of compliance review. Furthermore, by integrating and presenting the compliance review results, a feedback loop for reviewers is supported to optimize the compliance review model (e.g., marking misjudgment cases), thereby improving the accuracy of the compliance review model.

[0077] The following is a brief explanation of the method for generating keyword summary maps and location labels in the embodiments of this application. (See attached document for details.) Figure 3 As shown, the general process for generating keyword summary mind maps and location labels is as follows:

[0078] Step 301: Extract the original text of each paragraph related to compliance review from the administrative normative certification documents.

[0079] In this embodiment of the application, when extracting the original text content related to compliance review from administrative normative certification documents, a deeply optimized Transformer model can be used. A compliance review terminology database is introduced, and the textual features of administrative normative certification documents are learned through pre-training to accurately identify compliance review-related terms and chapter logical relationships, thereby extracting the original text content related to compliance review. The Transformer model includes a hierarchical parsing model, a location-based model, and a paragraph extraction model. When extracting the original text content related to compliance review from administrative normative certification documents, the following methods can be used, but are not limited to:

[0080] First, a hierarchical parsing model is employed to identify the hierarchical structure of administrative normative certification documents, obtaining hierarchical structured information. Specifically, the hierarchical parsing model includes a multimodal feature extraction module, a vertical paragraph reorganization module, and a document tree construction module connected in sequence. The multimodal feature extraction module extracts textual and graphical semantic features from the administrative normative certification documents. Based on these features, the CRAFT algorithm is used to parse visual attributes such as characters, fonts, font sizes, and colors, as well as related attributes, and detects the position coordinates and bounding boxes of text blocks. It also establishes the mapping relationship between the position coordinates and bounding boxes of each text block and the visual and related attributes. The vertical paragraph reorganization module uses the DBSCAN clustering algorithm to merge adjacent text blocks based on their position coordinates and bounding boxes, obtaining target text blocks while retaining the visual and related attributes of adjacent text blocks within the target text blocks. Finally, the document tree construction module infers the hierarchical relationships based on the visual and related attributes of adjacent text blocks within each target text block and converts these relationships into a tree structure to obtain hierarchical structured information.

[0081] Then, a location-based model is employed to dynamically match the hierarchical structured information with a compliance review terminology database to pinpoint the location of each paragraph within the hierarchical structured information relevant to compliance review. Specifically, the location-based model includes a triple-module approach: a rule-driven module, a semantic role analysis module, and a structural feature matching module. The rule-driven module uses the compliance review terminology database to locate the first paragraph within the hierarchical structured information relevant to compliance review. For example, if a mandatory verb (must / prohibited) is detected in the hierarchical structured information, the text range from that verb to the next heading is extracted as the paragraph location. The semantic role analysis module uses the ERNIE-4.0 algorithm to parse the semantic roles in the hierarchical structured information and, based on these semantic roles, locates the paragraphs within the hierarchical structured information relevant to compliance review. The system retrieves relevant paragraph location information, such as "[If Party A fails to make payment on time] <condition>", "[Party B has the right to terminate the contract] <behavior>", and automatically marks the paragraphs containing the "<condition>" and "<behavior>" as the second paragraph location information. It also identifies paragraphs containing implicit constraints that do not contain explicit keywords as the second paragraph location information. Through the structural feature matching module, after filtering the chapter titles in the hierarchical structured information, it detects paragraphs containing text with specific keywords (such as "responsibility", "confidentiality", "breach of contract") and paragraphs with specific identifiers (such as independent clause numbers, bold / underline, etc.) as paragraph location information.

[0082] Finally, a paragraph extraction model is used to extract the initial original text content of each paragraph related to compliance review from the administrative normative certification documents based on the position information of each paragraph. Then, based on a multi-dimensional filtering strategy, redundancy removal is performed on the initial original text content of each paragraph to obtain the original text content of each paragraph related to compliance review. Specifically, the paragraph extraction model includes a boundary determination module and a redundancy filtering module. The boundary determination module, based on the position information of each paragraph, uses punctuation rules (semicolon / period) and format features (indentation, line breaks) to divide the paragraphs and obtain the initial original text content of each paragraph. The redundancy filtering module uses a multi-dimensional filtering strategy to eliminate redundant content from the initial original text content of each paragraph, obtaining the original text content of each paragraph relevant to the compliance review. The multi-dimensional filtering strategy for eliminating redundant content from the initial original text content of each paragraph includes: calculating similarity using a sentence vector model (such as BERT-wwm) and merging repeated expressions; based on a compliance review terminology database, removing modifying content (such as "in principle" and "under normal circumstances") and removing explanatory clauses without legal effect (such as "the right to interpret these clauses belongs to..."); verifying the logical connection between paragraph titles and body text, and excluding non-core content through title hierarchy. In this way, the original text content extracted by the paragraph extraction model can maintain a high degree of integrity while achieving a high degree of simplification, thus making the extracted original text content related to compliance review more concise and accurate.

[0083] Step 302: Generate keyword summary information for each paragraph of the original text.

[0084] In this embodiment of the application, when generating keyword summary information for each segment of the original text, the following methods may be used, but are not limited to:

[0085] First, a keyword extraction model is used to extract candidate keywords from each paragraph of the original text. Specifically, the keyword extraction model is a dual-channel keyword extraction model based on the BERT semantic channel and the rule filtering channel. Through the BERT semantic channel, the semantic vectors of each paragraph of the original text are generated using bert-base-chinese, and an attention weighting mechanism is used to analyze and extract semantically significant words (such as high-frequency entity words after TF-IDF weighting) to obtain candidate keywords and confidence scores for each paragraph of the original text. From the candidate keywords of each paragraph of the original text, the candidate keywords with confidence scores not lower than the score threshold and / or the top K are selected as the first candidate keywords for each paragraph of the original text. Through the rule filtering channel, keywords related to compliance review are extracted from each paragraph of the original text using a compliance review terminology library, and invalid words such as modifier adverbs and stop words are filtered out from each paragraph of the original text using a syntactic filtering rule library and an invalid filtering rule library to obtain the second candidate keywords for each paragraph of the original text.

[0086] Then, a summary generation model is used to generate keyword summary information for each paragraph of the original text based on the candidate keywords of each paragraph. Specifically, the summary generation model includes a dual-channel fusion module and a summary generation module. Through the dual-channel fusion module, the union of the first and second candidate keywords corresponding to each paragraph of the original text is taken, and synonyms are merged based on word vector similarity to obtain the core keywords of each paragraph of the original text. The core keywords of each paragraph of the original text are then combined into coherent phrases according to semantic relevance to obtain the keyword summary information of each paragraph of the original text. When combining the core keywords of each paragraph into coherent phrases based on semantic relevance, the semantic similarity between the core keywords can be calculated first, and the core keywords can be mapped as nodes with semantic similarity as edge weights to generate a core keyword relationship graph. Then, graph structure feature analysis is performed on the core keyword relationship graph to identify the topological relationship between the headword (such as a verb) and modifiers (such as nouns / adjectives). After that, with the headword (such as a verb) as the core, the action objects in the modifiers (such as "traffic" and "ranking") and the limiting words (such as "natural search" and "SEO") are associated to form basic sentence structures. Finally, the rationality of the phrase structure is verified through dependency parsing, and prepositions / conjunctions (such as "based on" and "through") are automatically completed to obtain keyword summary information.

[0087] Step 303: Based on the hierarchical and semantic relationships between the original texts of each paragraph, the keyword summary information of each paragraph is structured to obtain a keyword summary mind map.

[0088] In this embodiment of the application, when the keyword summary information of each paragraph is structurally processed to obtain the keyword summary mind map according to the hierarchical relationship and semantic association between the paragraphs, the following methods may be used, but are not limited to:

[0089] First, based on the hierarchical structure information of the administrative normative supporting documents, the hierarchical relationship between the original text content of each paragraph is determined. Specifically, the hierarchical relationship between the original text content of each paragraph is directly extracted from the hierarchical structure information of the administrative normative supporting documents.

[0090] Then, based on the semantic similarity between keywords in each segment of the original text, the semantic relationships between the segments are determined. Specifically, based on the keywords in each segment, the cosine similarity between the segments is calculated. The semantic relationship strength between the segments is determined based on the cosine similarity. Based on the semantic relationship strength, hierarchical clustering, DBSCAN, or community detection algorithms (such as Louvain) are used to group semantically similar segments to obtain the semantic relationships between them.

[0091] Finally, based on the hierarchical relationships between the original text sections, a main structure with keyword summary information as nodes is constructed, and a branch structure with keyword summary information as nodes is constructed based on the semantic relationships between the original text sections to obtain a keyword summary mind map. For example, if there is a clear hierarchical relationship between the original text sections (such as chapter level, title level, clause level, etc.), the main structure between the original text sections is constructed based on the hierarchical relationship; if there is no clear hierarchical relationship between the original text sections but there is a semantic relationship, the branch structure between the original text sections is constructed based on the clustering results or similarity graph.

[0092] Step 304: Based on the position information of each keyword summary in the keyword summary map, generate the first positioning label for each keyword summary. The first positioning label, also known as the summary map position code, is formatted as K-level-coordinate; K is the keyword summary identifier, for example, the ordered word vector composed of the keywords in the keyword summary; level is the node depth in the keyword summary map, for example, 1 is the root node, 1.2 is the second-level node; coordinates are the (x,y) position of the keyword summary in the two-dimensional coordinate system of the keyword summary map, with the origin of the two-dimensional coordinate system being the upper left corner. For example, if a keyword summary is located in the second level, row 3, column 2 of the keyword summary map, then the first positioning label for this keyword summary is K-2.1-(3,2).

[0093] Step 305: Based on the location information of the original text content of each keyword summary information within the administrative normative certification document, generate a second location label for the original text content of each keyword summary information. The second location label, also known as the original text location code, is formatted as D-Document Type-Chapter-Paragraph-Line Number; D represents the original text content identifier, such as an ordered word vector composed of the keywords in the original text content; the document type refers to the type of administrative normative certification document, such as GF (normative document), XZ (administrative document), etc.; the chapter / section number is the chapter / section number, such as 3.2 indicating Chapter 3, Section 2; the paragraph number is the paragraph sequence number, such as starting from 1; and the line number is the starting line number, such as 5-7 indicating a span of 3 lines. Example: If a piece of original text content is located in Chapter 4, Section 1, Paragraph 3 of an administrative normative certification document, then the second location label for this original text content is D-XZ-4.1-3-9. When a keyword summary corresponds to multiple original text contents, the second positioning label of the original text contents is appended with / n. For example, if a keyword summary corresponds to two original text contents, the second positioning labels of these two original text contents are merged into D-XZ-4.1-3-9 / 1, / 2.

[0094] Step 306: Associate and store the keyword summary information and the first positioning index of each keyword summary information with the second positioning index of the original text content of each keyword summary information in the positioning index database. Specifically, the positioning index database establishes a first index relationship between each keyword summary information and its first positioning index, as well as an index relationship between the first positioning index and the second positioning index, thereby achieving associated storage.

[0095] In this embodiment, the keyword summary mind map generation method in steps 301-303 transforms the structured hierarchy and semantic relationships of administrative normative supporting documents into an intuitive relationship mind map. This helps reviewers quickly understand the logic of the administrative normative supporting documents, extract key information, and thus improve the efficiency of compliance review. The location label generation method in steps 304-306 obtains unique location codes and associations between the keyword summary information and the original text content. This allows for quick retrieval of the corresponding location labels during bidirectional navigation between the keyword summary information and the original text content, thereby facilitating compliance review for reviewers.

[0096] Based on the above embodiments, this application provides a document display system, see below. Figure 4 As shown, the document display system 400 provided in this application embodiment includes at least:

[0097] The split-screen display unit 401 is used to respond to the viewing operation of administrative normative certification documents by splitting the screen to display a summary map display area and a certification document display area. The administrative normative certification document is displayed in the certification document display area, and a keyword summary map of the administrative normative certification document is displayed in the summary map display area. The keyword summary map is obtained by extracting the original text content related to compliance review from the administrative normative certification document, generating keyword summary information for each original text content, and then structuring the keyword summary information of each original text content according to the logical relationship between the original text content.

[0098] The label query unit 402 is used to respond to the operation of viewing the original text of the keyword summary information in the keyword summary map currently displayed in the summary map display area, and to determine the first positioning label corresponding to the keyword summary information and the second positioning label associated with the first positioning label from the positioning label database; wherein, the first positioning label is used to represent the position information of the keyword summary information in the keyword summary map; and the second positioning label is used to represent the position information of the keyword summary information in the original text of the administrative normative certification document;

[0099] The jump display unit 403 is used to jump to the original text of the keyword summary information in the administrative normative certification document within the certification material display area based on the second positioning number.

[0100] In one possible implementation, the document display system 400 provided in this application embodiment further includes:

[0101] The structural parsing unit 404 is used to identify the hierarchical structure of administrative normative certification documents using a hierarchical parsing model to obtain hierarchical structure information.

[0102] Paragraph positioning unit 405 is used to dynamically match the hierarchical structured information with the compliance review terminology database using a position positioning model in order to locate the position information of each paragraph related to compliance review in the hierarchical structured information.

[0103] Paragraph extraction unit 406 is used to extract the initial original text content of each paragraph related to compliance review from the administrative normative certification document based on the position information of each paragraph using a paragraph extraction model, and to perform redundancy removal processing on the initial original text content of each paragraph based on a multi-dimensional filtering strategy to obtain the original text content of each paragraph related to compliance review.

[0104] In one possible implementation, the document display system 400 provided in this application embodiment further includes:

[0105] The abstract generation unit 407 is used to extract keywords from each paragraph of the original text using a keyword extraction model; and to generate keyword abstract information for each paragraph of the original text based on the keywords.

[0106] In one possible implementation, the document display system 400 provided in this application embodiment further includes:

[0107] The mind map generation unit 408 is used to determine the hierarchical relationship between the original text content of each paragraph based on the hierarchical structured information of the administrative normative certification materials; to determine the semantic association between the original text content of each paragraph based on the semantic similarity between the keywords of the original text content of each paragraph; to construct a main structure with keyword summary information as nodes based on the hierarchical relationship between the original text content of each paragraph; and to construct a branch structure with keyword summary information as nodes based on the semantic association between the original text content of each paragraph to obtain a keyword summary mind map.

[0108] In one possible implementation, the document display system 400 provided in this application embodiment further includes:

[0109] The labeling generation unit 409 is used to generate a first positioning label for each keyword summary based on the position information of each keyword summary in the keyword summary map; generate a second positioning label for the original content of each keyword summary based on the position information of the original content of each keyword summary in the administrative normative certification document; and associate and store each keyword summary and its first positioning label with the second positioning label of the original content of each keyword summary in the positioning label database.

[0110] In one possible implementation, the document display system 400 provided in this application embodiment further includes:

[0111] The auxiliary triggering unit 410 is used to respond to the auxiliary review operation of the administrative normative certification document, and to conduct a compliance review of the administrative normative certification document according to the compliance review method to obtain the compliance review results of each paragraph of the original text in the administrative normative certification document related to the compliance review; wherein, the compliance review method includes the compliance review rules corresponding to each compliance review clause;

[0112] The auxiliary display unit 411 is used to display a label box at each paragraph of the original text within the supporting document display area, and to present the compliance review results and applicable compliance review clauses corresponding to the original text within the label box.

[0113] In one possible implementation, the auxiliary display unit 411 is further configured to acquire the associated original text content; conduct a compliance review on the associated original text content according to the compliance review method to obtain the compliance review result of the associated original text content; and conduct a supplementary review on the compliance review result of the original text content based on the compliance review result of the associated original text content.

[0114] It should be noted that the principle of the document display system provided in the embodiments of this application to solve the technical problem is similar to that of the document display method provided in the embodiments of this application. Therefore, the implementation of the document display system provided in the embodiments of this application can refer to the implementation of the document display method provided in the embodiments of this application, and the repeated parts will not be described again.

[0115] After introducing the document display method and system provided in the embodiments of this application, the electronic device provided in the embodiments of this application will be briefly introduced next.

[0116] The electronic devices provided in this application embodiment may be, but are not limited to, computers, mobile phones, tablet computers, etc., see reference. Figure 5As shown, the electronic device 500 provided in this application embodiment includes at least a processor 501, a memory 502, and a computer program stored on the memory 502 and executable on the processor 501. When the processor 501 executes the computer program, it implements the document display method provided in this application embodiment.

[0117] The electronic device 500 provided in this application embodiment may further include a bus 503 connecting different components (including processor 501 and memory 502). The bus 503 represents one or more types of bus structures, including memory bus, peripheral bus, local area bus, etc.

[0118] Memory 502 may include readable media in the form of volatile memory, such as random access memory (RAM) 5021 and / or cache memory 5022, and may further include read-only memory (ROM) 5023. Memory 502 may also include a program tool 5025 having a set (at least one) of program modules 5024, including but not limited to an operating subsystem, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.

[0119] Processor 501 can be a single processing element or a collective term for multiple processing elements. For example, processor 501 can be a microcontroller unit (MCU), a central processing unit (CPU), or one or more integrated circuits configured to implement the methods shown in the above-described embodiments of this application. Specifically, processor 501 can be a general-purpose processor, including but not limited to CPUs, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0120] Electronic device 500 can also communicate with one or more devices that enable users to interact with electronic device 500 (e.g., mobile phones, computers, etc.), and / or with various external devices 504 such as devices that enable electronic device 500 to communicate with one or more other electronic devices (e.g., routers, modems, etc.). This communication can be performed through input / output (I / O) interface 505. Furthermore, electronic device 500 can also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) through network adapter 506. Figure 5 As shown, network adapter 506 communicates with other modules of electronic device 500 via bus 503. It should be understood that, although... Figure 5 As not shown, other hardware and / or software modules may be used in conjunction with the electronic device 500, including but not limited to microcode, device drivers, redundant processors, external disk drive arrays, Redundant Arrays of Independent Disks (RAID) subsystems, tape drives, and data backup storage subsystems.

[0121] It should be noted that, Figure 5 The electronic device 500 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0122] In addition, this application embodiment also provides a computer-readable storage medium storing computer instructions. When these computer instructions are executed by a processor, they implement the file display method provided in this application embodiment. Specifically, the computer instructions can be built into or installed in the processor, so that the processor can implement the file display method provided in this application embodiment by executing the built-in or installed computer instructions.

[0123] Furthermore, the document display method provided in this application embodiment can also be implemented as a program product, which includes program code. When the program code is executed by a processor, it implements the above-mentioned document display method provided in this application embodiment.

[0124] The program product provided in this application embodiment can be any combination of one or more readable media, wherein the readable media can be a readable signal medium or a readable storage medium, and the readable storage medium can be, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or apparatus, or any combination thereof. Specifically, more specific examples of readable storage media (a non-exhaustive list) include electrical connections with one or more wires, portable disks, hard disks, RAM, ROM, erasable programmable read-only memory (EPROM), optical fibers, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0125] The program product provided in this application embodiment can be a CD-ROM and include program code, and can also run on an electronic device. However, the program product provided in this application embodiment is not limited thereto. In this application embodiment, the readable storage medium can be any tangible medium that contains or stores a program, which can be used by or in conjunction with an instruction execution system, device, or apparatus.

[0126] It should be noted that although several units or sub-units of the device have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of this application, the features and functions of two or more units described above can be embodied in one unit. Conversely, the features and functions of one unit described above can be further divided and embodied by multiple units.

[0127] Furthermore, although the operations of the method of this application are described in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0128] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0129] Obviously, those skilled in the art can make various modifications and variations to the embodiments of this application without departing from the spirit and scope of the embodiments of this application. Therefore, if these modifications and variations to the embodiments of this application fall within the scope of the claims of this application and their equivalents, this application also intends to include these modifications and variations.

Claims

1. A method for displaying documents, characterized in that, include: In response to the viewing operation of administrative normative supporting documents, a split-screen display area for the summary map and a display area for the supporting documents are presented. The administrative normative supporting documents are displayed in the supporting documents display area, and a keyword summary map of the administrative normative supporting documents is displayed in the summary map display area. The keyword summary map is obtained by extracting the original text content related to compliance review from the administrative normative supporting documents, generating keyword summary information for each original text content, and then structuring the keyword summary information of each original text content according to the hierarchical and semantic relationships between the original text content. In response to an operation to view the original text of the keyword summary information in the keyword summary map currently displayed in the summary map display area, a first location label corresponding to the keyword summary information and a second location label associated with the first location label are determined from the location label database; wherein, the first location label is used to characterize the location information of the keyword summary information in the keyword summary map; and the second location label is used to characterize the location information of the keyword summary information in the original text of the administrative normative certification document; Based on the second positioning number, the display area of ​​the supporting materials jumps to the original text of the administrative normative supporting materials document to present the keyword summary information; In response to the auxiliary review operation of the administrative normative supporting documents, a compliance review is conducted on the administrative normative supporting documents according to the compliance review method to obtain the compliance review results of each paragraph of the original text in the administrative normative supporting documents related to the compliance review; wherein, the compliance review method includes the compliance review rules for each compliance review clause; Using an improved TF-IDF model, and employing part-of-speech weighting and a sliding window mechanism, a core keyword relationship graph is generated for the aforementioned administrative normative supporting documents and each candidate administrative normative supporting document; wherein, the candidate administrative normative supporting documents include other administrative normative supporting documents from the same reviewing entity; Using the LDA model, based on the document-term frequency matrix of the administrative normative certification documents and each of the candidate administrative normative certification documents, the topic distribution analysis is performed on the administrative normative certification documents and each of the candidate administrative normative certification documents to obtain the latent topic distribution of the administrative normative certification documents and each of the candidate administrative normative certification documents. Based on the core keyword relationship graph and potential topic distribution of the administrative normative certification documents and each of the candidate administrative normative certification documents, the text similarity between the administrative normative certification documents and each of the candidate administrative normative certification documents is calculated, and the candidate administrative normative certification documents with a text similarity not lower than the similarity threshold are identified as the associated administrative normative certification documents of the administrative normative certification documents. According to the compliance review method, the relevant administrative normative supporting documents are reviewed for compliance, and the compliance review results of each paragraph of the original text in the relevant administrative normative supporting documents related to the compliance review are obtained. Based on the compliance review results of the original text of each paragraph related to compliance review in the aforementioned related administrative normative certification documents, a supplementary review is conducted on the compliance review results of the original text of each paragraph related to compliance review in the aforementioned administrative normative certification documents, to obtain the final compliance review results of the original text of each paragraph related to compliance review in the aforementioned administrative normative certification documents; Within the area displaying the supporting documents, a label box is displayed at each section of the original text, and the corresponding final compliance review result and applicable compliance review clauses are presented within the label box.

2. The document display method as described in claim 1, characterized in that, Extract the original text content related to compliance review from the aforementioned administrative normative certification documents, including: A hierarchical parsing model is used to identify the hierarchical structure of the administrative normative certification documents to obtain hierarchical structure information; A location-based model is used to dynamically match the hierarchical structured information with a compliance review terminology database to locate the position information of each paragraph related to compliance review in the hierarchical structured information. A paragraph extraction model is used to extract the initial original text content of each paragraph related to compliance review from the administrative normative certification materials based on the position information of each paragraph. Then, based on a multi-dimensional filtering strategy, the initial original text content of each paragraph is redundantly removed to obtain the original text content of each paragraph related to compliance review.

3. The document display method as described in claim 1, characterized in that, Generate keyword summary information for each of the original text segments, including: A keyword extraction model is used to extract keywords from the original text of each paragraph. Based on the keywords of each paragraph of the original text, keyword summary information is generated for each paragraph of the original text.

4. The document display method as described in claim 1, characterized in that, Based on the hierarchical and semantic relationships between the original text segments, the keyword summary information of each segment is structured, including: Based on the hierarchical structured information of the aforementioned administrative normative certification documents, the hierarchical relationship between the original text content of each paragraph is determined. Based on the semantic similarity between the keywords in each paragraph of the original text, the semantic association between the paragraphs of the original text is determined. Based on the hierarchical relationship between the original text segments, a main structure with keyword summary information as nodes is constructed, and based on the semantic relationship between the original text segments, a branch structure with keyword summary information as nodes is constructed to obtain the keyword summary mind map.

5. The document display method as described in claim 1, characterized in that, After structuring the keyword summary information of each paragraph according to the hierarchical and semantic relationships between them, the process further includes: Based on the position information of each keyword summary information in the keyword summary map, a first positioning label is generated for each keyword summary information. Based on the position information of the original text of each keyword summary information in the administrative normative certification document, a second positioning label of the original text of each keyword summary information is generated; The keyword summary information and the first positioning label of each keyword summary information are associated with the second positioning label of the original text content of each keyword summary information and stored in the positioning label database.

6. A document display system, characterized in that, include: The split-screen display unit is used to respond to the viewing operation of administrative normative certification documents by splitting the screen to display a summary map display area and a certification document display area. The administrative normative certification document is displayed in the certification document display area, and a keyword summary map of the administrative normative certification document is displayed in the summary map display area. The keyword summary map is obtained by extracting the original text content related to compliance review from the administrative normative certification document, generating keyword summary information for each original text content, and then structuring the keyword summary information of each original text content according to the hierarchical and semantic relationships between the original text content. The label query unit is used to respond to an operation of viewing the original text of keyword summary information in the keyword summary map currently displayed in the summary map display area, and to determine, from the location label database, a first location label corresponding to the keyword summary information and a second location label associated with the first location label; wherein, the first location label is used to represent the location information of the keyword summary information in the keyword summary map; and the second location label is used to represent the location information of the keyword summary information in the original text of the administrative normative certification document; The jump display unit is used to jump to the original text of the administrative normative certification document based on the second positioning number within the certification material display area and present the keyword summary information in the original text of the administrative normative certification material document. An auxiliary triggering unit is used to respond to the auxiliary review operation of the administrative normative supporting document. Following a compliance review method, it performs a compliance review on the administrative normative supporting document to obtain the compliance review results of each paragraph of the original text related to the compliance review. The compliance review method includes the compliance review rules for each compliance review clause. Using an improved TF-IDF model, employing part-of-speech weighting and a sliding window mechanism, it generates a core keyword relationship graph for the administrative normative supporting document and each candidate administrative normative supporting document. The candidate administrative normative supporting document includes other administrative normative supporting documents from the same reviewing entity. Using an LDA model, based on the document-word frequency matrix of the administrative normative supporting document and each candidate administrative normative supporting document, it performs topic distribution analysis on the administrative normative supporting document and each candidate administrative normative supporting document to obtain the results of the administrative normative supporting document and each candidate administrative normative supporting document. The potential topic distribution of each administrative normative supporting document; based on the core keyword relationship graph and potential topic distribution of each of the administrative normative supporting documents and each of the candidate administrative normative supporting documents, the text similarity between the administrative normative supporting documents and each of the candidate administrative normative supporting documents is calculated, and the candidate administrative normative supporting documents with a text similarity not lower than the similarity threshold are identified as the associated administrative normative supporting documents of the administrative normative supporting documents; according to the compliance review method, the associated administrative normative supporting documents are subject to compliance review to obtain the compliance review results of each paragraph of the original text content related to the compliance review in the associated administrative normative supporting documents; based on the compliance review results of each paragraph of the original text content related to the compliance review in the associated administrative normative supporting documents, the compliance review results of each paragraph of the original text content related to the compliance review in the administrative normative supporting documents are subject to supplementary review to obtain the final compliance review results of each paragraph of the original text content related to the compliance review in the administrative normative supporting documents; An auxiliary display unit is used to display a label box at each section of the original text within the display area of ​​the supporting materials, and to present the final compliance review result and applicable compliance review clauses corresponding to the original text within the label box.

7. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the file display method as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the file display method as described in any one of claims 1-5.

Images