Rule system file processing method and device, electronic equipment and medium

Abstract

The invention provides a rule system file processing method and device, electronic equipment and a medium, and can be applied to the technical field of big data and the technical field of artificial intelligence. The method comprises the steps that supervision rule information and enterprise internal system information are acquired, and the supervision rule information and the enterprise internal system information respectively comprise a plurality of rule entries and a plurality of system entries; performing data preprocessing and text analysis on the plurality of rule entries and the plurality of system entries to obtain a plurality of analytic entries; allocating a corresponding label to each parsed entry by using a pre-constructed label system to obtain a plurality of entry label groups; traversing the entry tag group by using a pre-constructed association knowledge graph to obtain an entry association relationship; and calculating the matching between the rule entry and the system entry by using the entry association relationship, and outputting the processing level of the internal system information of the enterprise based on the matching.

Description

Technical Field

[0001] The present invention relates to the fields of big data technology and artificial intelligence technology, and more specifically to a method, device, electronic device and medium for processing rule and system documents. Background Art

[0002] Compliance issues have always been one of the key issues facing enterprises and organizations. However, due to the complexity, diversity and constant changes of regulatory rules, as well as differences in regulatory standards between different fields, many enterprises have difficulty in mastering regulatory rules and have deviations in the implementation of regulatory actions, which has led to the exposure of compliance risks and may lead to adverse consequences such as legal proceedings, fines and reputation damage.

[0003] As regulatory rules continue to escalate, companies and organizations need more efficient, intelligent and precise methods to meet compliance challenges. However, traditional compliance management methods can no longer meet the requirements of the new regulatory environment. Therefore, it is necessary to research and develop new technologies and methods to better understand, interpret and comply with regulatory rules.

[0004] Current intelligent technologies, including natural language processing, machine learning, and knowledge graphs, provide important support in helping companies understand regulatory rules, track rule changes, and improve compliance management efficiency. However, current technologies are still limited in their accuracy in identifying complex regulatory rules, which leads to problems in understanding and interpretation. Summary of the invention

[0005] In view of the above problems, according to a first aspect of the present invention, a method for processing rule and system files is provided, the method comprising: obtaining regulatory rule information and enterprise internal system information, wherein the regulatory rule information and enterprise internal system information respectively include multiple rule entries and multiple system entries; performing data preprocessing and text analysis on the multiple rule entries and the multiple system entries to obtain multiple parsed entries; using a pre-built label system to assign corresponding labels to each parsed entry to obtain multiple item label groups, wherein each item label group includes a parsed item and a label corresponding to the parsed item; using a pre-built associated knowledge graph to traverse the item label groups to obtain item association relationships, wherein the associated knowledge graph includes multiple entities and multiple edges, each of the entities represents the item label group, each of the edges represents the association relationship between the item label groups, and the item association relationship is used to indicate the association relationship between the rule item and the system item; and using the item association relationship to calculate the matching between the rule item and the system item, and outputting the processing level of the enterprise internal system information based on the matching.

[0006] According to some exemplary embodiments, pre-constructing an associated knowledge graph specifically includes: obtaining first text information of historical regulatory rules and second text information of historical enterprise internal systems, wherein the first text information and the second text information respectively include multiple historical entries; using a pre-constructed label system to assign corresponding labels to the first text information and the second text information to obtain a historical entry label group; performing knowledge extraction on the historical entry label group to obtain entities and inter-entity relationships in the historical entry label group; constructing a triple representation based on the entities and inter-entity relationships, wherein the triple representation includes the form of entities, relationships and entities; and constructing the associated knowledge graph based on the triple representation.

[0007] According to some exemplary embodiments, a label system is pre-constructed, specifically including: performing text analysis on the first text information and the second text information; setting label classifications based on the results of the text analysis and utilizing expert experience in compliance management, wherein the label classifications include major label categories that clarify the essence of the content and minor label categories that clarify the content links; setting label descriptions for labels in each category of the label classifications, wherein the label descriptions include associations between labels; and forming a label system based on the label classifications and label descriptions.

[0008] According to some exemplary embodiments, the associated knowledge graph is stored using a Neo4j graph database.

[0009] According to some exemplary embodiments, the method of calculating the matching between the rule items and the system items by utilizing the item association relationship and outputting the processing level of the enterprise internal system information based on the matching specifically includes: obtaining target regulatory rule information and corresponding enterprise internal system information based on the item association relationship; calculating the cosine similarity between the target regulatory rule information and the corresponding enterprise internal system information; and in response to the cosine similarity being less than a preset threshold, outputting the corresponding enterprise internal system information as a high processing level, wherein the high processing level indicates that the corresponding enterprise internal system information should be changed based on the target regulatory rule information.

[0010] According to some exemplary embodiments, the method also includes: in response to the number of corresponding internal enterprise system information being greater than 1, calculating weights based on corresponding label entries of the corresponding internal enterprise system information; sorting the corresponding internal enterprise system information in descending order according to the weights to obtain an adjustment order; and adjusting the internal enterprise system information according to the adjustment order.

[0011] According to some exemplary embodiments, the weight calculation based on the corresponding tag entry of the corresponding enterprise internal system information specifically includes: obtaining the tag usage frequency and the tag marking order based on the corresponding tag entry; and calculating the sum of the tag usage frequency and the tag marking order to obtain the weight.

[0012] According to some exemplary embodiments, after traversing the item tag group using a pre-built associated knowledge graph to obtain item association relationships, the method further includes: obtaining unrelated regulatory rule information based on the item association relationships, the unrelated regulatory rule information being used to represent regulatory rule information that cannot be associated with an enterprise's internal system; and outputting processing prompts based on the unrelated regulatory rule information to prompt the enterprise to reformulate its internal system.

[0013] According to some exemplary embodiments, the regulatory rule information includes image information, and the internal enterprise system information is text information; the data preprocessing and text analysis of the multiple rule entries and the multiple system entries to obtain multiple parsed entries specifically include: inputting the multiple rule entries into a layout analysis model to obtain a valid layout area; based on the valid layout area, using optical character recognition technology to extract text information of the multiple rule entries; and data cleaning and word segmentation of the text information of the multiple rule entries and the multiple system entries, and obtaining the parsed entries using a topic modeling algorithm.

[0014] According to some exemplary embodiments, based on the resolved entries, a corresponding label is assigned to each resolved entry using a pre-built label system, specifically including: using a relationship extraction algorithm to automatically identify key entities and relationships in the resolved entries; and based on the key entities and relationships, using a pre-built label system to assign a corresponding label to each entry.

[0015] According to some exemplary embodiments, the method of assigning a corresponding label to each entry using a pre-built label system specifically includes: based on the pre-built label system, assigning a corresponding label to each entry using an automatic labeling model, wherein the automatic labeling model includes a BERT feature representation layer for encoding each character, an IDCNN-BiLSTM feature learning layer for deep feature extraction, and a CRF knowledge reasoning layer for prediction.

[0016] According to a second aspect of the present invention, a device for processing internal enterprise system documents is proposed, the device comprising: an information acquisition module, used to: acquire regulatory rule information and internal enterprise system information, wherein the regulatory rule information and the internal enterprise system information respectively include a plurality of rule items and a plurality of system items; a parsed rule item acquisition module, used to: perform data preprocessing and text analysis on the plurality of rule items and the plurality of system items to obtain a plurality of parsed items; an item label group acquisition module, used to: assign a corresponding label to each parsed item using a pre-constructed label system to obtain a plurality of item label groups, wherein each item label group includes a parsed An entry and a label corresponding to the parsed entry; an entry association relationship acquisition module, used to: traverse the entry label group using a pre-built association knowledge graph to obtain the entry association relationship, wherein the association knowledge graph includes multiple entities and multiple edges, each of the entities represents the entry label group, each of the edges represents the association relationship between the entry label groups, and the entry association relationship is used to indicate the association relationship between the rule entry and the system entry; and a processing level output module, used to: calculate the matching between the rule entry and the system entry using the entry association relationship, and output the processing level of the internal system information of the enterprise based on the matching.

[0017] According to some exemplary embodiments, the layout analysis unit may be used to input the multiple rule entries into a layout analysis model to obtain a valid layout area.

[0018] According to some exemplary embodiments, the text extraction unit may be configured to extract texts of a plurality of rule entries based on the valid layout area by using an optical character recognition technology.

[0019] According to some exemplary embodiments, the parsed item acquisition unit may be used to perform data cleaning and word segmentation on the text information of the multiple rule items and the multiple system items, and obtain the parsed items using a topic modeling algorithm.

[0020] According to some exemplary embodiments, the entry tag group acquisition module may include a relationship extraction unit and a tag assignment unit.

[0021] According to some exemplary embodiments, the relationship extraction unit may be configured to automatically identify key entities and relationships in the parsed entries using a relationship extraction algorithm.

[0022] According to some exemplary embodiments, the label assigning unit may be configured to assign a corresponding label to each entry based on the key entities and relationships using a pre-built labeling system.

[0023] According to some exemplary embodiments, the device for processing internal enterprise system documents may further include a label system building module.

[0024] According to some exemplary embodiments, the tag system construction module may include a text analysis unit, a tag classification setting unit, a tag description setting unit, and a tag system forming unit.

[0025] According to some exemplary embodiments, the text analysis unit may be configured to perform text analysis on the first text information and the second text information.

[0026] According to some exemplary embodiments, the label classification setting unit can be used to set label classification based on the results of text analysis and using expert experience in compliance management, wherein the label classification includes major label categories that clarify the essence of the content and minor label categories that clarify the content links.

[0027] According to some exemplary embodiments, the tag description setting unit may be used to set tag descriptions for tags of each category in the tag classification, wherein the tag descriptions include association relationships between tags.

[0028] According to some exemplary embodiments, the label system forming unit may be configured to form a label system based on the label classification and the label description.

[0029] According to some exemplary embodiments, the entry association relationship acquisition module may include a history entry acquisition unit, a history tag assignment unit, a knowledge extraction unit, a triple representation construction unit and an associated knowledge graph acquisition unit.

[0030] According to some exemplary embodiments, the historical entry acquisition unit may be used to acquire first text information of historical regulatory rules and second text information of historical enterprise internal systems, wherein the first text information and the second text information respectively include a plurality of historical entries.

[0031] According to some exemplary embodiments, the history tag assigning unit may be configured to assign corresponding tags to the first text information and the second text information using a pre-built tag system to obtain a history entry tag group.

[0032] According to some exemplary embodiments, the knowledge extraction unit may be used to perform knowledge extraction on the historical item tag group to obtain entities and relationships between entities in the historical item tag group.

[0033] According to some exemplary embodiments, the triple representation construction unit may be used to construct a triple representation based on the entities and the relationships between entities, wherein the triple representation includes entities, relationships, and forms of entities.

[0034] According to some exemplary embodiments, the associated knowledge graph acquisition unit may be used to construct the associated knowledge graph based on the triple representation.

[0035] According to some exemplary embodiments, the processing level output module may include a first processing level output module and a second processing level output module.

[0036] According to some exemplary embodiments, the first processing level output module may include a target information acquisition unit, a cosine similarity calculation unit, and an enterprise internal system change unit.

[0037] According to some exemplary embodiments, the target information acquisition unit may be used to acquire target regulatory rule information and corresponding enterprise internal system information based on the item association relationship.

[0038] According to some exemplary embodiments, the cosine similarity calculation unit may be used to calculate the cosine similarity between the target regulatory rule information and the corresponding enterprise internal system information.

[0039] According to some exemplary embodiments, the enterprise internal system change unit can be used to output the corresponding enterprise internal system information as a high processing level in response to the cosine similarity being less than a preset threshold, wherein the high processing level indicates that the corresponding enterprise internal system information should be changed based on the target regulatory rule information.

[0040] According to some exemplary embodiments, the second processing level output module may include a weight calculation module, an adjustment sequence acquisition unit, and a sequential adjustment unit.

[0041] According to some exemplary embodiments, the weight calculation module may be configured to calculate a weight based on a corresponding tag entry of the corresponding enterprise internal system information in response to the number of the corresponding enterprise internal system information being greater than 1.

[0042] According to some exemplary embodiments, the adjustment order acquisition unit may be used to sort the corresponding internal enterprise system information in descending order according to the weight to acquire the adjustment order.

[0043] According to some exemplary embodiments, the sequential adjustment may be used to adjust the internal system information of the enterprise according to the adjustment sequence.

[0044] According to some exemplary embodiments, the weight calculation module may include a weight acquisition unit and a sorting unit.

[0045] According to some exemplary embodiments, the weight acquisition unit may be used to acquire the tag usage frequency and the tag marking order based on the corresponding tag entry.

[0046] According to some exemplary embodiments, the sorting unit may be used to calculate the sum of the tag usage frequency and the tag marking order to obtain the weight.

[0047] According to some exemplary embodiments, the device for processing the internal enterprise system documents may further include an unrelated regulatory rule information processing module.

[0048] According to some exemplary embodiments, the unrelated regulatory rule information processing module may include an unrelated regulatory rule information acquisition unit and an enterprise internal system reformulation unit.

[0049] According to some exemplary embodiments, the unrelated regulatory rule information acquisition unit may be used to acquire unrelated regulatory rule information based on the item association relationship, where the unrelated regulatory rule information is used to represent regulatory rule information that cannot be associated with an internal system of an enterprise.

[0050] According to some exemplary embodiments, the enterprise internal system reformulation unit may be used to output a processing prompt based on the unrelated regulatory rule information to prompt the enterprise to reformulate its internal system.

[0051] According to a third aspect of the present invention, there is provided an electronic device, comprising: one or more processors; a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors execute the method as described above.

[0052] According to a fourth aspect of the present invention, there is provided a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, causes the processor to execute the method as described above.

[0053] According to a fifth aspect of the present invention, there is provided a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the method as described above is implemented.

[0054] One or more of the above embodiments have the following advantages or beneficial effects: by pre-assigning a label to each entry through a label system, the computer can more accurately analyze the association relationship of the knowledge graph, which means that the computer can obtain the association relationship between entries more quickly and accurately, thereby improving the calculation speed and reducing the calculation time and resource consumption; the establishment of a knowledge graph can help users better understand the content relevance between regulatory rules and internal corporate systems; at the same time, by analyzing the massive data of regulatory rules information and internal corporate system information through automated methods, the tediousness of manual analysis is reduced, thereby improving customer satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] The above contents and other objects, features and advantages of the present invention will become more apparent through the following description of the embodiments of the present invention with reference to the accompanying drawings, in which:

[0056] Figure 1 The application scenario diagram of the method, device, equipment, and medium for processing rule and system files according to an embodiment of the present invention is schematically shown.

[0057] Figure 2 The flowchart of the method for processing a rule file according to an embodiment of the present invention is schematically shown.

[0058] Figure 3 The flowchart of the method of preprocessing and text analysis according to an embodiment of the present invention is schematically shown.

[0059] Figure 4 The flowchart of the method for allocating tags using a tag system according to an embodiment of the present invention is schematically shown.

[0060] Figure 5 The flowchart of the method for pre-building a label system according to an embodiment of the present invention is schematically shown.

[0061] Figure 6 A flowchart of a method for pre-constructing an associated knowledge graph according to an embodiment of the present invention is schematically shown.

[0062] Figure 7 The flowchart of the method for adjusting the internal system information of an enterprise according to the calculation results according to an embodiment of the present invention is schematically shown.

[0063] Figure 8 The flowchart of the method for adjusting and processing a plurality of corresponding internal system information of an enterprise according to an embodiment of the present invention is schematically shown.

[0064] Fig. 9 The flowchart of the method for calculating weight according to an embodiment of the present invention is schematically shown.

[0065] Fig.10 The flowchart of the method for adjusting and processing internal system information of an enterprise according to another embodiment of the present invention is schematically shown.

[0066] Fig.11 The structure block diagram of the device for processing internal system documents of an enterprise according to an embodiment of the present invention is schematically shown.

[0067] Fig.12 The block diagram schematically shows an electronic device suitable for a method for processing a rule file according to an embodiment of the present invention. DETAILED DESCRIPTION

[0068] Below, embodiments of the present invention will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present invention. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of embodiments of the present invention. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of concepts of the present invention.

[0069] The terms used herein are only for describing specific embodiments and are not intended to limit the present invention. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0070] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0071] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0072] In the technical solution of the present invention, the acquisition, storage and application of user personal information involved all comply with the provisions of relevant laws and regulations, take necessary confidentiality measures, and do not violate public order and good morals.

[0073] First, the technical terms described in this article are explained and described as follows.

[0074] Knowledge graph is a structured and semantic knowledge representation method used to describe and organize entities, concepts, relationships and attributes in the real world. It aims to simulate and utilize the structure of human knowledge so that computers can better understand and reason about knowledge.

[0075] Neo4j graph database is a popular graph database management system that is specifically designed to store, manage, and query graph data. It is a highly flexible and high-performance database system that is particularly suitable for processing relational data and graph data.

[0076] Optical Character Recognition (OCR) is a technology that converts printed or handwritten text into editable text. It is mainly used to recognize and extract text from paper documents, pictures or scanned documents for further processing, search or archiving by computers.

[0077] In recent years, the global regulatory environment has undergone significant changes, with regulatory standards continuously upgraded and regulatory rules changing frequently, which has brought severe compliance challenges to various enterprises and organizations. This trend reflects the regulators' stricter supervision of market activities to ensure compliance in finance, healthcare, technology and other industries, as well as the emergence of new challenges and threats, such as privacy, data protection, and cybersecurity. In this context, compliance risk events have exposed the problems of unclear understanding of regulatory rules and deviations in the implementation of regulatory actions, which have attracted widespread attention in various industries.

[0078] Frequent changes and upgrades in regulatory rules, as well as their complexity across different areas, make it difficult for companies to track, understand and comply with relevant rules. This not only increases compliance risks, but can also lead to regulatory penalties and reputational damage. Regulators are increasingly demanding compliance, placing higher demands on companies, including timely reporting, data transparency and risk management. Failure to comply with regulatory rules can cause serious damage to a company's business, so accurate understanding and effective implementation of regulatory rules are crucial.

[0079] In addition, with the continued growth of international trade and globalization, enterprises and financial institutions are facing diversified and complex regulatory challenges. Compliance risks are no longer limited to domestic regulation, but also include cross-border compliance issues. Enterprises need to comply with the regulations of multiple countries and regions at the same time, which makes compliance more complex and challenging.

[0080] Based on this, an embodiment of the present invention provides a method for processing a rule and system file, characterized in that the method includes: obtaining regulatory rule information and enterprise internal system information, wherein the regulatory rule information and enterprise internal system information respectively include multiple rule entries and multiple system entries; performing data preprocessing and text analysis on the multiple rule entries and the multiple system entries to obtain multiple parsed entries; using a pre-constructed label system to assign a corresponding label to each parsed entry to obtain multiple item label groups, wherein each item label group includes a parsed item and a label corresponding to the parsed item; using a pre-constructed associated knowledge graph to traverse the item label groups to obtain item association relationships, wherein the associated knowledge graph includes multiple entities and multiple edges, each of the entities represents the item label group, each of the edges represents the association relationship between the item label groups, and the item association relationship is used to indicate the association relationship between the rule item and the system item; and using the item association relationship to calculate the matching between the rule item and the system item, and outputting the processing level of the enterprise internal system information based on the matching. According to the method for processing rule and system files provided by the present invention, a label is pre-assigned to each entry through a label system, and the computer can analyze the association relationship of the knowledge graph more accurately, which means that the computer can obtain the association relationship between entries more quickly and accurately, thereby improving the calculation speed, and then reducing the calculation time and resource consumption; the establishment of a knowledge graph can help users better understand the content relevance between regulatory rules and internal corporate systems; at the same time, the massive data of regulatory rule information and internal corporate system information is analyzed by automated methods, which reduces the tediousness of manual analysis, thereby improving customer satisfaction.

[0081] It should be noted that the processing method, device, equipment and medium of the rule and system file determined by the present invention can be used in the fields of big data technology and artificial intelligence technology, can also be used in the financial field, and can also be used in various fields other than the fields of big data technology and artificial intelligence technology and the financial field. The application field of the processing method, device, equipment and medium of the rule and system file provided by the embodiments of the present invention is not limited.

[0082] In the technical solution of the present invention, the user information (including but not limited to user personal information, user image information, user device information, such as location information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved are all information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of relevant data comply with the relevant laws, regulations and standards of relevant countries and regions, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0083] Figure 1 The application scenario diagram of the method, device, equipment, and medium for processing rule and system files according to an embodiment of the present invention is schematically shown.

[0084] like Figure 1 As shown, the application scenario 100 according to this embodiment may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for a communication link between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0085] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only examples).

[0086] The terminal devices 101 , 102 , and 103 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers.

[0087] The server 105 may be a server that provides various services, such as a background management server (only an example) that provides support for websites browsed by users using the terminal devices 101, 102, and 103. The background management server may analyze and process the received data such as user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device.

[0088] It should be noted that the method for processing the rule and system file provided in the embodiment of the present invention can generally be executed by the server 105. Accordingly, the device for processing the internal system file of an enterprise provided in the embodiment of the present invention can generally be set in the server 105. The method for processing the rule and system file provided in the embodiment of the present invention can also be executed by a server or server cluster that is different from the server 105 and can communicate with the terminal devices 101, 102, 103 and / or the server 105. Accordingly, the device for processing the internal system file of an enterprise provided in the embodiment of the present invention can also be set in a server or server cluster that is different from the server 105 and can communicate with the terminal devices 101, 102, 103 and / or the server 105.

[0089] It should be understood that Figure 1The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0090] Figure 2 The flowchart of the method for processing a rule file according to an embodiment of the present invention is schematically shown.

[0091] like Figure 2 As shown, the method 200 for processing a rule file of this embodiment may include operations S210 to S250.

[0092] In operation S210, regulatory rule information and enterprise internal system information are obtained, wherein the regulatory rule information and enterprise internal system information include a plurality of rule items and a plurality of system items, respectively.

[0093] In the embodiments of the present invention, regulatory rule information and enterprise internal system information may be collected from different sources, including regulatory documents, government agency announcements, company internal documents, contracts, policy documents, website content, etc.

[0094] In operation S220, data preprocessing and text analysis are performed on the multiple rule entries and the multiple policy entries to obtain multiple parsed entries.

[0095] In an embodiment of the present invention, the regulatory rule information that can be directly obtained is usually a PDF file, which can be converted into a multi-page picture including image information to facilitate further analysis; the internal system information of the enterprise is usually text data.

[0096] Figure 3 The flowchart of the method of preprocessing and text analysis according to an embodiment of the present invention is schematically shown.

[0097] like Figure 3 As shown, the method for preprocessing and text analysis of this embodiment may include operations S310 to S330 , and operations S310 to S330 may at least partially execute operation S220 .

[0098] In operation S310, the plurality of rule entries are input into a layout analysis model to obtain a valid layout area.

[0099] In an embodiment of the present invention, the task of the layout analysis model is to analyze the structure of the document, determine the location of text and images, and identify the effective layout area. This helps to reduce the scope of text processing to improve efficiency. The effective layout area can include different layout areas, each area represents a picture, a paragraph or a text line, so as to perform subsequent text extraction and processing.

[0100] In operation S320, based on the valid layout area, texts of a plurality of rule entries are extracted using optical character recognition technology.

[0101] In an embodiment of the present invention, optical character recognition (OCR) technology can recognize text characters in an image and convert them into editable text, which makes the text content become digital data that can be understood and processed by a computer. Specifically, OCR technology can extract the text form of multiple rule entries in the effective layout area for subsequent analysis.

[0102] In operation S330, data cleaning and word segmentation are performed on the text information of the plurality of rule entries and the plurality of system entries, and the parsed entries are obtained by using a topic modeling algorithm.

[0103] In an embodiment of the present invention, data cleaning may include removing unnecessary characters, formats, punctuation marks and special characters; word segmentation includes splitting continuous text into separate words or phrases. A topic modeling algorithm is used to identify topics, keywords and context information from the text to obtain parsed entries of multiple rule entries and multiple institutional entries.

[0104] Return to reference Figure 2 In operation S230, a corresponding label is assigned to each resolved item using a pre-built label system to obtain a plurality of item label groups, wherein each item label group includes a resolved item and a label corresponding to the resolved item.

[0105] Figure 4 The flowchart of the method for allocating tags using a tag system according to an embodiment of the present invention is schematically shown.

[0106] like Figure 4 As shown, the method for assigning tags using a tag system in this embodiment may include operations S410 to S420, and operations S410 to S420 may at least partially execute operation S230.

[0107] In operation S410, a relationship extraction algorithm is used to automatically identify key entities and relationships in the parsed entries.

[0108] In an embodiment of the present invention, the key entities and relations in the parsed entries are automatically identified by using a relation extraction algorithm, which may include the following operations: the dependency relationship between each word in the text may be determined by using a dependency syntactic analysis technique, including identifying subject-predicate-object structures, modification relationships, etc., to understand the structure and grammar of the text; the named entity recognition (NER) technique may be used to identify named entities in the text, such as names of people, places, and organizations, etc. These entities are usually candidates for key entities, wherein HanLP may be selected as a Chinese word segmentation toolkit, which supports functions such as Chinese word segmentation, part-of-speech tagging, keyword extraction, and text classification; the relation extraction algorithm may be used to identify entities from the text and extract semantic relationships between entities based on the results of dependency syntactic analysis and the output of entity recognition, including screening out key entities in the text, and classifying the extracted relationships to determine their properties. This may include labeling the relationships as "working for", "belonging to", "responsible for", etc., and the results may be converted into triples representing "'entity 1', 'relationship', and 'entity 2'".

[0109] In operation S420, a corresponding tag is assigned to each entry using a pre-built tag system based on the key entities and relationships.

[0110] In an embodiment of the present invention, the assignment of labels can be based on the following considerations: entity attributes, the attributes and features of key entities in an entry can be used to determine labels; relationship types, parsed relationship types can help determine appropriate labels; context information, the context and context of the text can also affect the assignment of labels; multi-label classification, each entry can be assigned one or more labels, which means that an entry can belong to multiple categories, reflecting the diversity of text information. Among them, the multi-label classification rule can decide whether to assign a label based on the relevance of each label.

[0111] In an embodiment of the present invention, labeling can be performed based on a pre-built labeling system using an automatic labeling model to assign a corresponding label to each entry.

[0112] Figure 5 The flowchart of the method for pre-building a label system according to an embodiment of the present invention is schematically shown.

[0113] like Figure 5 As shown, the method for pre-building a label system in this embodiment may include operations S510 to S540.

[0114] In operation S510, text analysis is performed on the first text information of the historical regulatory rules and the second text information of the historical enterprise internal system.

[0115] In operation S520, based on the result of the text analysis and using the expert experience of compliance management, a tag classification is set, wherein the tag classification includes a tag category that clarifies the essence of the content and a tag subcategory that clarifies the content links.

[0116] In operation S530, a tag description is set for each tag in the tag classification, wherein the tag description includes an association relationship between tags.

[0117] In operation S540, a tag system is formed based on the tag classification and the tag description.

[0118] In the embodiment of the present invention, Table 1 gives an example of a label system. A multi-level, multi-dimensional intelligent label system can be constructed based on compliance management needs and expectations. By pre-defining the specific content of the label, accurate matching and guidance can be achieved. The label classification should include a major label category that clarifies the essence of the content and a minor label category that clarifies the content link, so that the entities in the system clauses correspond to the four label relationships of field, matter, region, and industry.

[0119] In an embodiment of the present invention, the corpus for data processing can be manually annotated according to the established label system. The language sequence can be annotated using the BIO rule, where B (Begin) describes the starting position of each named entity in the sentence; I (Internal) describes the other parts of the named entity except the starting position, and O (Other) is used to describe other non-predefined entities in the sentence.

[0120] Furthermore, the BERT-IDCNN-BiLSTM-CRF model can be trained based on the labeled data to obtain an automatic labeling model capable of label assignment. Specifically, the BERT-IDCNN-BiLSTM-CRF model can be used, which mainly consists of three parts: BERT feature representation layer, IDCNN-BiLSTM feature learning layer, and CRF knowledge reasoning layer. First, the text character sequence to be predicted is input into the BERT feature representation layer, which encodes each character and obtains the corresponding word vector; then the word vector feature is input into the IDCNN model for preliminary feature extraction; then the features extracted by IDCNN are input into the recurrent neural network BiLSTM to extract deep features; finally, CRF is used to predict the label sequence with the highest probability, thereby completing the labeling.

[0121] Specifically, for a given sequence X = [x 1 , x 2 ...x n ] and the corresponding label sequence Y = [y 1 ,y 2 ...y n], and its scoring formula is:

[0122]

[0123] Where W is the transfer matrix, represents the label transfer score.

[0124] Then the Softmax function is used to get the maximum probability of the predicted sequence, and the maximum probability is calculated to determine whether it matches:

[0125]

[0126] Table 1 Examples of labeling systems

[0127]

[0128] According to an embodiment of the present invention, BERT adopts a multi-layer bidirectional Transformer structure, each unit of which is composed of a self-attention mechanism. The model uses two unsupervised pre-training tasks, namely, a Masked language model and a next sentence prediction (NSP), and is trained on a large-scale corpus, so as to fully learn the meaning of each word. The IDCNN-BiLSTM model uses convolutional neural networks and recurrent neural networks to learn the entire semantic information of the text, and predicts the label corresponding to each word in the sequence according to the contextual features of the current word and the word vector information. CRF can consider the adjacent relationship between labels, and learn the constraint information between labels during the training process, such as the constraint that the head label of the entity is "B" and the "I" label cannot follow the "O" label.

[0129] Return to reference Figure 2 In operation S240, the item label group is traversed using a pre-built associated knowledge graph to obtain item association relationships, wherein the associated knowledge graph includes multiple entities and multiple edges, each of the entities represents the item label group, and each of the edges represents the association relationship between the item label groups, and the item association relationship is used to indicate the association relationship between the rule item and the system item.

[0130] Figure 6 A flowchart of a method for pre-constructing an associated knowledge graph according to an embodiment of the present invention is schematically shown.

[0131] like Figure 6 As shown, the method for pre-constructing an associated knowledge graph in this embodiment may include operations S610 to S650, and operations S610 to S650 may at least partially execute operation S240.

[0132] In operation S610, first text information of historical regulatory rules and second text information of historical enterprise internal systems are obtained, wherein the first text information and the second text information respectively include a plurality of historical entries.

[0133] It should be noted that the text tags in the history entry tag group need to go through some text preprocessing steps, such as word segmentation, stemming or lemmatization, to prepare the text for subsequent processing.

[0134] In operation S620, the pre-built tag system is used to assign corresponding tags to the first text information and the second text information to obtain a history entry tag group.

[0135] In operation S630, knowledge extraction is performed on the historical item tag group to obtain entities and relationships between entities in the historical item tag group.

[0136] In an embodiment of the present invention, the purpose of the knowledge extraction process is to convert unstructured text information into a structured knowledge representation in order to better understand the information, support tasks such as search, reasoning and decision making, which is very valuable for building knowledge graphs, automated knowledge management and information retrieval.

[0137] In operation S640, a triple representation is constructed based on the entities and the relationships between the entities, wherein the triple representation includes the form of the entity, the relationship, and the entity.

[0138] In operation S650, the associated knowledge graph is constructed based on the triple representation.

[0139] In an embodiment of the present invention, through the identification of entities and relationships, the system can construct entity-relationship-entity (ERE) triples to represent this information. Each triple consists of three main components: entity 1, relationship, and entity 2. By combining multiple triples together, an associative knowledge graph can be constructed in which the relationship between entities is explicitly represented to organize and understand the relevance of information.

[0140] It should be noted that in the process of building a knowledge graph, quality control is also required to ensure the accuracy of entities and relationships, and may involve dealing with ambiguity or polysemy issues.

[0141] In an embodiment of the present invention, knowledge storage is a very important part of the knowledge graph construction process. After the data from different sources and in different forms are integrated, these data need to be stored in a database. Among them, the associated knowledge graph can be stored using the Neo4j graph database. The Neo4j graph database enables users to effectively manage and query data in order to perform various analysis and information retrieval tasks. The Neo4j graph database stores information in the form of a graph structure, and the graph database can directly display the characteristics of associated data and the relationship between data.

[0142] Return to reference Figure 2 In operation S250, the matching between the rule item and the system item is calculated using the item association relationship, and the processing level of the internal system information of the enterprise is output based on the matching.

[0143] In an embodiment of the present invention, the processing level of the internal system information of an enterprise can be divided into a high processing level and a low processing level. Among them, when the rule item and the system item are highly matched, a low processing level can be output, indicating that the relevance of the internal system information of the enterprise is high and does not need to be adjusted or modified; when the rule item and the system item are less matched, a high processing level can be output, indicating that the relevance of the internal system information of the enterprise is low and needs to be adjusted or modified.

[0144] Figure 7 The flowchart of the method for adjusting the internal system information of an enterprise according to the calculation results according to an embodiment of the present invention is schematically shown.

[0145] like Figure 7 As shown, the method for adjusting the internal system information of an enterprise according to the calculation result in this embodiment may include operations S710 to S730, and operations S710 to S730 may at least partially execute operation S250.

[0146] In operation S710, target regulatory rule information and corresponding enterprise internal system information are obtained based on the item association relationship.

[0147] In an embodiment of the present invention, based on the established item association relationships, the system can find matching regulatory rules and internal systems according to the association relationships.

[0148] In operation S720, the cosine similarity between the target regulatory rule information and the corresponding enterprise internal system information is calculated.

[0149] In an embodiment of the present invention, for each pair of matched target regulatory rule information and corresponding enterprise internal system information, the cosine similarity between them is calculated. Cosine similarity is a commonly used text similarity measurement method used to compare the similarity between texts. A high cosine similarity indicates a higher similarity between text information, and a low cosine similarity indicates a lower similarity between text information. For example, the target regulatory rule A and the corresponding enterprise internal system information B can be represented in the form of a vector model a=[a 1 , a 2 , a 3 , ..., a n ],b=[b 1 , b 2 , b 3 , ..., b n ], then the cosine similarity can be expressed as:

[0150]

[0151] In operation S730, in response to the cosine similarity being less than a preset threshold, the corresponding enterprise internal system information is output as a high processing level, wherein the high processing level indicates that the corresponding enterprise internal system information should be changed based on the target regulatory rule information.

[0152] In the embodiment of the present invention, the range of cosine similarity is from -1 to 1. -1 means that the two vectors point in opposite directions, 1 means that the two vectors point in the same direction, 0 generally means that the two vectors are independent, and other values ​​between -1 and 1 represent intermediate similarities or dissimilarities.

[0153] In an embodiment of the present invention, the system can set a preset cosine similarity threshold, for example, if the similarity is greater than 90%, it is considered similar, which is used to determine when the internal system of the enterprise needs to be changed. If the cosine similarity is less than this threshold, it means that the similarity between the regulatory rule information and the internal system information is low. In response to the high processing level output due to the low cosine similarity, the system can choose to change the internal system information of the enterprise to make it more compatible or compliant with the regulatory rule information.

[0154] According to an embodiment of the present invention, the above process can compare regulatory rule information with the enterprise's internal system information to detect potential inconsistencies or mismatches, and adjust the internal system based on the similarity threshold to ensure compliance, which helps to ensure the consistency of the enterprise's internal system with external regulatory rules, reduce compliance risks, and promote compliance management, thereby automating the inspection and adjustment of regulatory compliance and improving efficiency.

[0155] In addition, one regulatory rule may correspond to multiple internal systems, but the degree of similarity is not the same. In order to put important information in the forefront and further improve the efficiency of compliance management and information retrieval, an embodiment of the present invention also provides a method for adjusting and processing multiple corresponding internal system information of enterprises.

[0156] Figure 8 The flowchart of the method for adjusting and processing a plurality of corresponding internal system information of an enterprise according to an embodiment of the present invention is schematically shown.

[0157] like Figure 8 As shown, the method for adjusting and processing a plurality of corresponding internal enterprise system information in this embodiment may include operations S810 to S830.

[0158] In operation S810 , in response to the number of the corresponding internal enterprise system information being greater than 1, a weight is calculated based on a corresponding tag entry of the corresponding internal enterprise system information.

[0159] Fig. 9 The flowchart of the method for calculating weight according to an embodiment of the present invention is schematically shown.

[0160] like Fig. 9 As shown, the method for calculating weights in this embodiment may include operations S910 to S920.

[0161] In operation S910, based on the corresponding tag entry, the tag usage frequency and the tag marking order are acquired.

[0162] In an embodiment of the present invention, the system can collect the usage frequency of tags and the tagging order of tags according to the existing corresponding tag entries, which includes analyzing the existing text or data to determine which tags are frequently used and their tagging order.

[0163] In operation S920, the sum of the tag usage frequency and the tag marking order is calculated to obtain the weight.

[0164] In an embodiment of the present invention, the system can consider the importance of different tags and assign different weights to the tag usage frequency and the tag annotation order, for example, 1.0, 0.8, 0.6, 0.4, 0.2, and 0.1, respectively, and calculate the sum of the two to obtain the weight. Specifically, the weights corresponding to the tag usage frequency and the tag annotation order can be summed separately, and then combined into a single weight value, and the calculation formula is as follows:

[0165] Weight(t i )=Freq(t i )+Seq(t i ) (4)

[0166] Among them, Freq(t i ) indicates the number of times the tag is used, Seq(t i ) indicates the order of labeling. The labels are sorted in descending order according to their total weight values ​​and recommended to users to facilitate internal system updates.

[0167] Return to reference Figure 8 In operation S820, the corresponding internal enterprise system information is sorted in descending order according to the weight to obtain an adjustment order.

[0168] In an embodiment of the present invention, multiple corresponding internal enterprise system information are arranged from high to low according to weight values, so as to determine which information has higher importance or influence, so as to improve the accessibility and comprehensibility of information and ensure that important information is easier to find and use.

[0169] In operation S830, the internal system information of the enterprise is adjusted according to the adjustment sequence.

[0170] In addition, the embodiments of the present invention also provide for the detection of regulatory rule information that cannot be associated with the internal system of the enterprise, so as to achieve comprehensive compliance management and reduce the workload of manual information processing.

[0171] Fig.10 The flowchart of the method for adjusting and processing internal system information of an enterprise according to another embodiment of the present invention is schematically shown.

[0172] like Fig.10 As shown, the method for adjusting and processing internal system information of an enterprise in this embodiment may include operations S1010 to S1020.

[0173] In operation S1010, based on the item association relationship, unrelated regulatory rule information is obtained, where the unrelated regulatory rule information is used to represent regulatory rule information that cannot be associated with the internal system of the enterprise.

[0174] In an embodiment of the present invention, the system can detect and mark regulatory rule information that cannot be associated with the enterprise's internal system based on the item association relationship for further processing. This information can be used to re-formulate the enterprise's internal system to ensure that it is consistent with the applicable regulatory rules.

[0175] In operation S1020, based on the unrelated regulatory rule information, a processing prompt is output to prompt the enterprise to reformulate its internal system.

[0176] In an embodiment of the present invention, with respect to the acquired irrelevant regulatory rule information, an enterprise may be prompted to redefine its internal system to ensure that it better complies with applicable regulatory rules.

[0177] In an embodiment of the present invention, after the internal system of an enterprise is adjusted and updated (including modification or re-enactment), the previous users can be found based on the tracking points to achieve accurate push of system information.

[0178] According to the processing method of the rule and system file provided by the present invention, through customer value classification, the computer can more accurately distinguish users of different value levels and allocate resources more targetedly, thereby improving the calculation speed, thereby improving the prediction efficiency of the pass rate prediction model, and reducing the calculation time and resource consumption; at the same time, through the screening of high-value users and the prediction of the pass rate, it is ensured that high-value customers obtain approval results faster, thereby improving customer satisfaction. Specifically, it brings the following beneficial effects:

[0179] 1. Improve compliance management efficiency: The efficiency of compliance management can be significantly improved through automated text processing, label assignment, and the use of associated knowledge graphs, which helps companies adapt to changing regulatory rules more quickly;

[0180] 2. Improve information manageability: The use of label assignment and associated knowledge graphs makes information easier to manage, organize and retrieve. By pre-assigning labels to each item through the label system, the computer can analyze the association relationship of the knowledge graph more accurately, which means that the computer can obtain the association relationship between items more quickly and accurately, thereby improving the calculation speed;

[0181] 3. Automatically detect inconsistencies: By calculating the matching, the system can automatically detect inconsistencies between regulatory rules and internal corporate systems, thereby reducing compliance risks;

[0182] 4. Improve compliance: Automated matching and adjustment processes help ensure that the company's internal systems are consistent with applicable regulatory rules. This helps improve compliance levels;

[0183] 5. Better understanding of regulatory rules: The use of associative knowledge graphs enables enterprises to better understand the relationship and impact between regulatory rules. This can help enterprises better respond to regulatory changes.

[0184] Based on the above-mentioned method for processing rule and system documents, the present invention also provides a device for processing enterprise internal system documents. Fig.11 The device is described in detail.

[0185] Fig.11 The structure block diagram of the device for processing internal system documents of an enterprise according to an embodiment of the present invention is schematically shown.

[0186] like Fig.11 As shown, according to this embodiment, the processing device 1100 for enterprise internal system documents includes an information acquisition module 1110, a parsing rule item acquisition module 1120, an item tag group acquisition module 1130, an item association relationship acquisition module 1140 and a processing level output module 1150.

[0187] The information acquisition module 1110 can be used to acquire regulatory rule information and enterprise internal system information, wherein the regulatory rule information and enterprise internal system information include multiple rule items and multiple system items, respectively. In one embodiment, the information acquisition module 1110 can be used to perform the operation S210 described above, which will not be repeated here.

[0188] The parsed rule entry acquisition module 1120 may be used to perform data preprocessing and text analysis on the multiple rule entries and the multiple policy entries to obtain multiple parsed entries. In one embodiment, the parsed rule entry acquisition module 1120 may be used to perform the operation S220 described above, which will not be described in detail here.

[0189] The item tag group acquisition module 1130 can be used to assign a corresponding tag to each parsed item using a pre-built tag system to obtain multiple item tag groups, wherein each item tag group includes a parsed item and a tag corresponding to the parsed item. In one embodiment, the item tag group acquisition module 1130 can be used to perform the operation S230 described above, which will not be repeated here.

[0190] The item association relationship acquisition module 1140 can be used to traverse the item tag group using a pre-built association knowledge graph to obtain item association relationships, wherein the association knowledge graph includes multiple entities and multiple edges, each of the entities represents the item tag group, each of the edges represents the association relationship between the item tag groups, and the item association relationship is used to indicate the association relationship between the rule item and the system item. In one embodiment, the item association relationship acquisition module 1140 can be used to perform the operation S240 described above, which will not be repeated here.

[0191] The processing level output module 1150 can be used to calculate the matching between the rule items and the system items using the item association relationship, and output the processing level of the enterprise internal system information based on the matching. In one embodiment, the processing level output module 1150 can be used to perform the operation S250 described above, which will not be repeated here.

[0192] According to an embodiment of the present invention, the parsing rule entry acquisition module 1120 may include a layout analysis unit, a text extraction unit and a parsing entry acquisition unit.

[0193] The layout analysis unit may be used to input the plurality of rule entries into a layout analysis model to obtain a valid layout area. In one embodiment, the layout analysis unit may be used to perform the operation S310 described above, which will not be described in detail herein.

[0194] The text extraction unit may be used to extract texts of multiple rule entries based on the effective layout area using optical character recognition technology. In one embodiment, the text extraction unit may be used to perform the operation S320 described above, which will not be described in detail here.

[0195] The parsed item acquisition unit can be used to perform data cleaning and word segmentation on the text information of the multiple rule items and the multiple system items, and obtain the parsed items using a topic modeling algorithm. In one embodiment, the parsed item acquisition unit can be used to perform the operation S330 described above, which will not be repeated here.

[0196] According to an embodiment of the present invention, the entry tag group acquisition module 1130 may include a relationship extraction unit and a tag assignment unit.

[0197] The relationship extraction unit may be used to automatically identify key entities and relationships in the parsed entries using a relationship extraction algorithm. In one embodiment, the relationship extraction unit may be used to perform the operation S410 described above, which will not be described in detail herein.

[0198] The label assignment unit may be used to assign corresponding labels to each entry based on the key entities and relationships using a pre-built label system. In one embodiment, the label assignment unit may be used to perform the operation S420 described above, which will not be described in detail here.

[0199] According to an embodiment of the present invention, the device 1100 for processing internal enterprise system documents may further include a label system building module.

[0200] According to an embodiment of the present invention, the tag system construction module may include a text analysis unit, a tag classification setting unit, a tag description setting unit and a tag system forming unit.

[0201] The text analysis unit may be used to perform text analysis on the first text information and the second text information. In one embodiment, the text analysis unit may be used to perform the operation S510 described above, which will not be described in detail herein.

[0202] The label classification setting unit can be used to set label classification based on the result of text analysis and the expert experience of compliance management, wherein the label classification includes a label category that clarifies the essence of the content and a label subcategory that clarifies the content link. In one embodiment, the label classification setting unit can be used to perform the operation S520 described above, which will not be repeated here.

[0203] The tag description setting unit may be used to set a tag description for each tag in the tag classification, wherein the tag description includes an association relationship between tags. In one embodiment, the tag description setting unit may be used to perform the operation S530 described above, which will not be described in detail here.

[0204] The label system forming unit may be used to form a label system based on the label classification and label description. In one embodiment, the label system forming unit may be used to perform the operation S540 described above, which will not be described in detail herein.

[0205] According to an embodiment of the present invention, the entry association relationship acquisition module 1140 may include a history entry acquisition unit, a history tag assignment unit, a knowledge extraction unit, a triple representation construction unit and an associated knowledge graph acquisition unit.

[0206] The historical entry acquisition unit can be used to acquire first text information of historical regulatory rules and second text information of historical enterprise internal systems, wherein the first text information and the second text information respectively include multiple historical entries. In one embodiment, the historical entry acquisition unit can be used to perform the operation S610 described above, which is not repeated here.

[0207] The history tag assignment unit may be used to assign corresponding tags to the first text information and the second text information using a pre-built tag system to obtain a history entry tag group. In one embodiment, the history tag assignment unit may be used to perform the operation S620 described above, which will not be described in detail here.

[0208] The knowledge extraction unit may be used to extract knowledge from the history entry tag group to obtain entities and inter-entity relationships in the history entry tag group. In one embodiment, the knowledge extraction unit may be used to perform the operation S630 described above, which will not be described in detail here.

[0209] The triple representation construction unit can be used to construct a triple representation based on the entity and the relationship between entities, wherein the triple representation includes the form of entity, relationship and entity. In one embodiment, the triple representation construction unit can be used to perform the operation S640 described above, which will not be repeated here.

[0210] The associated knowledge graph acquisition unit may be used to construct the associated knowledge graph based on the triple representation. In one embodiment, the associated knowledge graph acquisition unit may be used to perform the operation S650 described above, which will not be described in detail here.

[0211] According to an embodiment of the present invention, the processing level output module 1150 may include a first processing level output module and a second processing level output module.

[0212] According to an embodiment of the present invention, the first processing level output module may include a target information acquisition unit, a cosine similarity calculation unit and an enterprise internal system modification unit.

[0213] The target information acquisition unit may be used to acquire target supervision rule information and corresponding enterprise internal system information based on the item association relationship. In one embodiment, the target information acquisition unit may be used to perform the operation S710 described above, which will not be described in detail here.

[0214] The cosine similarity calculation unit may be used to calculate the cosine similarity between the target regulatory rule information and the corresponding enterprise internal system information. In one embodiment, the cosine similarity calculation unit may be used to perform the operation S720 described above, which will not be described in detail here.

[0215] The enterprise internal system change unit may be used to output the corresponding enterprise internal system information as a high processing level in response to the cosine similarity being less than a preset threshold, wherein the high processing level indicates that the corresponding enterprise internal system information should be changed based on the target regulatory rule information. In one embodiment, the enterprise internal system change unit may be used to perform the operation S730 described above, which will not be described in detail herein.

[0216] According to an embodiment of the present invention, the second processing level output module may include a weight calculation module, an adjustment sequence acquisition unit and a sequential adjustment unit.

[0217] The weight calculation module may be used to calculate the weight based on the corresponding tag entry of the corresponding enterprise internal system information in response to the number of the corresponding enterprise internal system information being greater than 1. In one embodiment, the weight calculation module may be used to perform the operation S810 described above, which will not be described in detail here.

[0218] The adjustment sequence acquisition unit may be used to sort the corresponding internal enterprise system information in descending order according to the weight to acquire the adjustment sequence. In one embodiment, the adjustment sequence acquisition unit may be used to perform the operation S820 described above, which will not be described in detail here.

[0219] The sequential adjustment may be used to adjust the internal system information of the enterprise according to the adjustment sequence. In one embodiment, the sequential adjustment may be used to perform the operation S830 described above, which will not be described in detail here.

[0220] According to an embodiment of the present invention, the weight calculation module may include a weight acquisition unit and a sorting unit.

[0221] The weight acquisition unit may be used to acquire the tag usage frequency and the tag marking order based on the corresponding tag entry. In one embodiment, the weight acquisition unit may be used to perform the operation S910 described above, which will not be described in detail here.

[0222] The sorting unit may be used to calculate the sum of the tag usage frequency and the tag marking order to obtain the weight. In one embodiment, the sorting unit may be used to perform the operation S920 described above, which will not be described in detail here.

[0223] According to an embodiment of the present invention, the device 1100 for processing internal enterprise system documents may further include an unrelated regulatory rule information processing module.

[0224] According to an embodiment of the present invention, the unrelated regulatory rule information processing module may include an unrelated regulatory rule information acquisition unit and an enterprise internal system reformulation unit.

[0225] The unrelated regulatory rule information acquisition unit can be used to acquire unrelated regulatory rule information based on the item association relationship, and the unrelated regulatory rule information is used to represent regulatory rule information that cannot be associated with the internal system of the enterprise. In one embodiment, the unrelated regulatory rule information acquisition unit can be used to perform the operation S1010 described above, which will not be repeated here.

[0226] The enterprise internal system reformulation unit may be used to output a processing prompt based on the unrelated regulatory rule information to prompt the enterprise to reformulate the enterprise internal system. In one embodiment, the enterprise internal system reformulation unit may be used to perform the operation S1020 described above, which will not be repeated here.

[0227] According to an embodiment of the present invention, any multiple modules among the information acquisition module 1110, the parsed rule entry acquisition module 1120, the item tag group acquisition module 1130, the item association relationship acquisition module 1140 and the processing level output module 1150 can be combined in one module for implementation, or any one of the modules can be split into multiple modules. Alternatively, at least part of the functions of one or more of these modules can be combined with at least part of the functions of other modules and implemented in one module. According to an embodiment of the present disclosure, at least one of the information acquisition module 1110, the parsed rule entry acquisition module 1120, the item tag group acquisition module 1130, the item association relationship acquisition module 1140 and the processing level output module 1150 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation methods of software, hardware and firmware or in an appropriate combination of any of them. Alternatively, at least one of the information acquisition module 1110, the parsing rule item acquisition module 1120, the item tag group acquisition module 1130, the item association relationship acquisition module 1140 and the processing level output module 1150 can be at least partially implemented as a computer program module, which can perform corresponding functions when the computer program module is executed.

[0228] Fig.12 The block diagram schematically shows an electronic device suitable for a method for processing a rule file according to an embodiment of the present invention.

[0229] like Fig.12 As shown, the electronic device 1200 according to an embodiment of the present invention includes a processor 1201, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1202 or a program loaded from a storage part 1208 into a random access memory (RAM) 1203. The processor 1201 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 1201 may also include an onboard memory for caching purposes. The processor 1201 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present invention.

[0230] In RAM 1203, various programs and data required for the operation of electronic device 1200 are stored. Processor 1201, ROM 1202 and RAM 1203 are connected to each other via bus 1204. Processor 1201 performs various operations of the method flow according to the embodiment of the present invention by executing the program in ROM 1202 and / or RAM 1203. It should be noted that the program can also be stored in one or more memories other than ROM 1202 and RAM 1203. Processor 1201 can also perform various operations of the method flow according to the embodiment of the present invention by executing the program stored in the one or more memories.

[0231] According to an embodiment of the present invention, the electronic device 1200 may further include an input / output (I / O) interface 1205, which is also connected to the bus 1204. The electronic device 1200 may further include one or more of the following components connected to the I / O interface 1205: an input portion 1206 including a keyboard, a mouse, etc.; an output portion 1207 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 1208 including a hard disk, etc.; and a communication portion 1209 including a network interface card such as a LAN card, a modem, etc. The communication portion 1209 performs communication processing via a network such as the Internet. A drive 1210 is also connected to the I / O interface 1205 as needed. A removable medium 1211, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 1210 as needed, so that a computer program read therefrom is installed into the storage portion 1208 as needed.

[0232] The present invention also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiment; or may exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present invention is implemented.

[0233] According to an embodiment of the present invention, the computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, an apparatus or a device. For example, according to an embodiment of the present invention, the computer-readable storage medium may include the ROM 1202 and / or RAM 1203 described above and / or one or more memories other than ROM 1202 and RAM 1203.

[0234] The embodiment of the present invention also includes a computer program product, which includes a computer program, and the computer program contains program code for executing the method shown in the flowchart. When the computer program product is run in a computer system, the program code is used to enable the computer system to implement the method provided by the embodiment of the present invention.

[0235] The computer program executes the above functions defined in the system / device of the embodiment of the present invention when it is executed by the processor 1201. According to the embodiment of the present invention, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0236] In one embodiment, the computer program may rely on tangible storage media such as optical storage devices, magnetic storage devices, etc. In another embodiment, the computer program may also be transmitted and distributed in the form of signals on a network medium, and downloaded and installed through the communication part 1209, and / or installed from the removable medium 1211. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0237] In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 1209, and / or installed from the removable medium 1211. When the computer program is executed by the processor 1201, the above functions defined in the system of the embodiment of the present invention are performed. According to the embodiment of the present invention, the system, device, means, module, unit, etc. described above can be implemented by a computer program module.

[0238] According to an embodiment of the present invention, the program code for executing the computer program provided by the embodiment of the present invention can be written in any combination of one or more programming languages, specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, such as Java, C++, python, "C" language or similar programming languages. The program code can be executed completely on the user computing device, partially on the user device, partially on the remote computing device, or completely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (e.g., using an Internet service provider to connect through the Internet).

[0239] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0240] The embodiments of the present invention are described above. However, these embodiments are only for the purpose of illustration, and are not intended to limit the scope of the present invention. Although each embodiment is described above, it does not mean that the measures in each embodiment cannot be used in combination advantageously. Without departing from the scope of the present invention, those skilled in the art may make various substitutions and modifications, which should all fall within the scope of the present invention.

Claims

1. A method for processing rule and system documents, characterized in that: The method comprises: Obtaining regulatory rule information and enterprise internal system information, wherein the regulatory rule information and enterprise internal system information include a plurality of rule items and a plurality of system items respectively; Performing data preprocessing and text analysis on the multiple rule entries and the multiple system entries to obtain multiple parsed entries; Using a pre-built label system to assign a corresponding label to each resolved item, to obtain a plurality of item label groups, wherein each of the item label groups includes a resolved item and a label corresponding to the resolved item; Using a pre-built association knowledge graph to traverse the item tag group to obtain an item association relationship, wherein the association knowledge graph includes a plurality of entities and a plurality of edges, each of the entities represents the item tag group, each of the edges represents an association relationship between the item tag groups, and the item association relationship is used to indicate an association relationship between the rule item and the system item; and The matching between the rule items and the system items is calculated by using the item association relationship, and the processing level of the internal system information of the enterprise is output based on the matching.

2. The method according to claim 1, characterized in that Pre-build the associated knowledge graph, including: Acquire first text information of historical regulatory rules and second text information of historical enterprise internal systems, wherein the first text information and the second text information respectively include a plurality of historical entries; Assigning corresponding tags to the first text information and the second text information using a pre-built tag system to obtain a history entry tag group; Performing knowledge extraction on the historical item tag group to obtain entities and relationships between entities in the historical item tag group; Based on the entities and the relationships between the entities, constructing a triple representation, wherein the triple representation includes the form of the entity, the relationship, and the entity; and Based on the triple representation, the associated knowledge graph is constructed.

3. The method according to claim 2, characterized in that Pre-built labeling system, including: Performing text analysis on the first text information and the second text information; Based on the results of text analysis and using the expert experience of compliance management, label classification is set, wherein the label classification includes a major label category that clarifies the essence of the content and a minor label category that clarifies the content links; Setting a tag description for each tag in the tag classification, wherein the tag description includes an association relationship between tags; and A label system is formed based on the label classification and label description.

4. The method according to any one of claims 1 to 3, characterized in that: The associated knowledge graph is stored using the Neo4j graph database.

5. The method according to claim 1, characterized in that The calculating the matching between the rule item and the system item by using the item association relationship, and outputting the processing level of the enterprise internal system information based on the matching specifically includes: Based on the association relationship between the items, obtain the target regulatory rule information and the corresponding enterprise internal system information; Calculating the cosine similarity between the target regulatory rule information and the corresponding enterprise internal system information; and In response to the cosine similarity being less than a preset threshold, the corresponding enterprise internal system information is output as a high processing level, wherein the high processing level indicates that the corresponding enterprise internal system information should be changed based on the target regulatory rule information.

6. The method according to claim 5, characterized in that The method further comprises: In response to the number of the corresponding internal enterprise system information being greater than 1, calculating a weight based on a corresponding tag entry of the corresponding internal enterprise system information; Sort the corresponding internal enterprise system information in descending order according to the weight to obtain an adjustment order; and The internal system information of the enterprise is adjusted and processed according to the adjustment sequence.

7. The method according to claim 6, characterized in that The weight calculation based on the corresponding tag entry of the corresponding enterprise internal system information specifically includes: Based on the corresponding tag entry, obtaining the tag usage frequency and the tag marking order; and The weight is obtained by calculating the sum of the tag usage frequency and the tag marking order.

8. The method according to any one of claims 5 to 7, characterized in that: After traversing the item tag group using the pre-built associated knowledge graph to obtain the item association relationship, the method further includes: Based on the item association relationship, obtaining unrelated regulatory rule information, wherein the unrelated regulatory rule information is used to represent regulatory rule information that cannot be associated with the internal system of the enterprise; and Based on the unrelated regulatory rule information, processing prompts are output to prompt the enterprise to reformulate its internal system.

9. The method according to any one of claims 1 to 3, 5 to 7, characterized in that: The regulatory rule information includes image information, and the enterprise internal system information is text information; The performing data preprocessing and text analysis on the multiple rule items and the multiple system items to obtain multiple parsed items specifically includes: Inputting the plurality of rule entries into a layout analysis model to obtain a valid layout area; Based on the effective layout area, extracting text information of multiple rule entries using optical character recognition technology; and Data cleaning and word segmentation are performed on the text information of the multiple rule entries and the multiple system entries, and the parsed entries are obtained using a topic modeling algorithm.

10. The method according to claim 9, characterized in that The method of using the pre-built label system to assign a corresponding label to each parsed item to obtain a plurality of item label groups specifically includes: Automatically identifying key entities and relationships in the parsed entries using a relationship extraction algorithm; as well as Based on the key entities and relationships, a pre-built tag system is used to assign corresponding tags to each entry.

11. The method according to claim 10, characterized in that The method of using the pre-built label system to assign a corresponding label to each entry specifically includes: Based on the pre-built label system, an automatic labeling model is used to assign corresponding labels to each entry, wherein the automatic labeling model includes a BERT feature representation layer for encoding each character, an IDCNN-BiLSTM feature learning layer for deep feature extraction, and a CRF knowledge reasoning layer for prediction.

12. A device for processing internal system documents of an enterprise, characterized in that: The device comprises: An information acquisition module is used to: acquire regulatory rule information and enterprise internal system information, wherein the regulatory rule information and enterprise internal system information include a plurality of rule items and a plurality of system items respectively; A parsed rule entry acquisition module is used to: perform data preprocessing and text analysis on the multiple rule entries and the multiple system entries to obtain multiple parsed entries; An item label group acquisition module is used to: assign a corresponding label to each parsed item using a pre-built label system to obtain a plurality of item label groups, wherein each item label group includes a parsed item and a label corresponding to the parsed item; An item association relationship acquisition module is used to: traverse the item tag group using a pre-built association knowledge graph to obtain item association relationships, wherein the association knowledge graph includes multiple entities and multiple edges, each of the entities represents the item tag group, each of the edges represents the association relationship between the item tag groups, and the item association relationship is used to indicate the association relationship between the rule item and the system item; and The processing level output module is used to calculate the matching between the rule items and the system items by using the item association relationship, and output the processing level of the internal system information of the enterprise based on the matching.

13. An electronic device comprising: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors are enabled to execute the method according to any one of claims 1 to 11.

14. A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, causes the processor to execute the method according to any one of claims 1 to 11.

Images