Administrative law enforcement compliance early warning method based on multi-source data fusion

Abstract

The invention relates to an administrative law enforcement compliance early warning method based on multi-source data fusion, and the method comprises the steps: collecting a data set related to administrative law enforcement from a plurality of data sources, obtaining more comprehensive and richer information, and reducing the data one-sidedness or error caused by a single data source; through reverse verification of the data set, risk indexes in the law enforcement process can be extracted, and based on the extracted risk indexes, a risk early warning mechanism can be established, potential law enforcement risks can be found and early warned in time, and decision support is provided for law enforcement. Through comparative analysis of risk indexes by a rule engine, non-compliance parts in law enforcement behaviors can be found in time, and corresponding measures are taken for improvement, so that the compliance of the law enforcement behaviors is improved; through the state circulation management of the compliance evaluation result, the law enforcement behavior can be dynamically tracked and monitored, problems can be timely found and solved, and the law enforcement efficiency is improved.

Description

Technical Field

[0001] This application relates to the technical field of compliance monitoring and analysis, and particularly to an administrative law enforcement compliance warning method based on multi-source data fusion. Background Art

[0002] Existing compliance monitoring and analysis often rely on manual verification of paper files, resulting in a lack of real-time performance. Since relevant data are scattered in different systems, there is a lack of linkage analysis ability. Due to overly rigid rules, it is difficult to flexibly handle various complex situations in actual law enforcement. Due to the lack of a two-way feedback mechanism between law enforcement departments and supervision departments, the efficiency of warning processing is low.

[0003] It should be noted that the above introduction to the technical background is only for the convenience of clearly and completely explaining the technical solutions of this application and facilitating the understanding of those skilled in the art. It cannot be considered that the above technical solutions are well-known to those skilled in the art just because these solutions are described in the background art part of this application. Summary of the Invention

[0004] The purpose of this application is to provide an administrative law enforcement compliance warning method to solve the problems of lack of real-time performance in existing compliance monitoring and analysis, lack of linkage analysis ability, difficulty in flexibly handling complex situations, and low efficiency of warning processing.

[0005] To solve the above problems, an administrative law enforcement compliance warning method involved in this application adopts the following technical solutions: Collect a data set related to administrative law enforcement based on multiple data sources; Conduct reverse verification on the data set to obtain risk indicators of the law enforcement process in the data set; Compare and analyze the risk indicators by a preset rule engine to obtain a compliance evaluation result of the law enforcement behavior, and perform state transition on the compliance evaluation result according to a preset state set.

[0006] To solve the above problems, an administrative law enforcement compliance warning system involved in this application adopts the following technical solutions: A collection module, which is used to collect a data set related to administrative law enforcement based on multiple data sources; A first acquisition module, which is used to conduct reverse verification on the data set to obtain risk indicators of the law enforcement process in the data set; A second acquisition module, which compares and analyzes the risk indicators by a preset rule engine to obtain a compliance evaluation result of the law enforcement behavior, and performs state transition on the compliance evaluation result according to a preset state set.

[0007] To solve the above problems, an electronic device involved in the present application includes: a processor; a memory for storing executable instructions of the processor; wherein, the processor is configured to execute the instructions to implement the law enforcement compliance warning method involved in the present application.

[0008] To solve the above problems, a non-transitory computer-readable storage medium involved in the present application, when the instructions in the storage medium are executed by a processor of an electronic device, enables the electronic device to execute the law enforcement compliance warning method involved in the present application.

[0009] To solve the above problems, a computer program product involved in the present application includes a computer program, and the computer program implements the law enforcement compliance warning method involved in the present application when executed by a processor in a communication device.

[0010] The beneficial effects of the present application are as follows: By collecting data sets related to administrative law enforcement from multiple data sources, more comprehensive and richer information can be obtained, reducing data one-sidedness or errors caused by a single data source; by performing reverse verification on the data sets, risk indicators in the law enforcement process can be extracted, and based on the extracted risk indicators, a risk warning mechanism can be established to timely discover and warn of potential law enforcement risks, providing decision-making support for law enforcement; by comparing and analyzing the risk indicators through a rule engine, non-compliance issues in law enforcement actions can be timely discovered and corresponding measures can be taken for improvement, thereby enhancing the compliance of law enforcement actions; by managing the state transition of the compliance assessment results, dynamic tracking and monitoring of law enforcement actions can be achieved, timely discovering and solving problems, and improving law enforcement efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] To more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings required for the embodiments: Figure 1 A flowchart of a law enforcement compliance warning method provided by an embodiment of the present application; Figure 2 A flowchart of another law enforcement compliance warning method provided by an embodiment of the present application; Figure 3 A structural schematic diagram of a law enforcement compliance warning system provided by an embodiment of the present application; Figure 4 A structural schematic diagram of an electronic device provided according to an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0012] To make the technical objectives, technical solutions, and beneficial effects of this application clearer, the following further elaborates on the technical solutions of this application in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not used to limit this application, that is, the described embodiments are only a part of the embodiments of this application, rather than all of the embodiments. Generally, the components of the embodiments of this application described and illustrated in the accompanying drawings herein can be arranged and designed in various different configurations.

[0013] The terms used in the embodiments of this application are only for the purpose of describing specific embodiments and are not intended to limit the embodiments of this application. The singular forms "a" and "the" used in the embodiments of this application and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0014] It should be understood that although the terms first, second, third, etc. may be used in the embodiments of this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from each other. For example, without departing from the scope of the embodiments of this application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the words "if" and "when" as used herein may be interpreted as "when" or "while" or "in response to determining".

[0015] The following details the embodiments of this application. The examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals represent the same or similar elements throughout. The embodiments described below by referring to the accompanying drawings are exemplary and are intended to explain this application and should not be construed as a limitation to this application.

[0016] It should be noted that the method provided by any one of the embodiments of this application can be executed alone, or can be executed together with the possible implementation methods in other embodiments, or can be executed together with any one of the technical solutions in the related art.

[0017] The current compliance monitoring and analysis overly rely on the manual method of flipping through paper files. This traditional approach is not only inefficient but also difficult to instantly capture compliance defects such as "procedure skipping" and "document omission", thus severely restricting the timeliness and accuracy of supervision. In the current information environment, law enforcement personnel information, case processing flow records, and legal provision databases are stored separately in multiple isolated systems, lacking a comprehensive platform that can dynamically integrate and conduct correlation analysis on these key information, greatly hindering the comprehensiveness and in-depthness of supervision work. Moreover, the existing compliance monitoring and analysis usually adopt a warning mechanism based on fixed thresholds, so it seems powerless when facing complex and changeable law enforcement scenarios such as omission of collective discussions on major cases. It is not only difficult to respond flexibly due to rigid rules but also leads to a high false alarm rate, affecting the effectiveness and credibility of supervision. In the existing compliance monitoring and analysis system, there is a lack of an effective two-way communication and feedback mechanism between the law enforcement department and the supervision department, which directly results in the failure to quickly convert warning information into actual actions, with low processing efficiency and a significant discount on the effectiveness of the warning system, and fails to form a virtuous cycle of closed-loop management.

[0018] The law enforcement compliance warning method of the embodiments of the present application will be described below with reference to the accompanying drawings.

[0019] Figure 1 It is a schematic flowchart of a law enforcement compliance warning method provided by the embodiments of the present application.

[0020] As Figure 1 shown, the law enforcement compliance warning method includes but is not limited to the following steps: S101, collect a data set related to administrative law enforcement based on multiple data sources.

[0021] In a feasible implementation, administrative law enforcement data usually comes from multiple different data sources, such as internal databases of government departments, third-party data providers, public data sources, etc. Appropriate data collection tools and methods can be used, such as ETL tools (Extract, Transform, Load, that is, the three processes of extraction, transformation, and loading), database connection technologies, API interfaces (Application Programming Interface), etc., to collect data from each data source. As an example, ETL tools can be used to connect the collected data to a big data platform. During the data connection process, it is necessary to maintain data consistency, including consistency in data volume, number of fields, field values, etc. (which can be achieved through data verification, etc.). Then, the collected data is cleaned to remove duplicate data, invalid data, error data, etc.; and data in different formats and standards is converted into a unified format and standard for subsequent data analysis and processing; the cleaned and transformed data is aggregated into a data set, where the data set reflects a complete data view related to administrative law enforcement.

[0022] S102, perform reverse verification on the data set to obtain risk indicators of the law enforcement process in the data set.

[0023] In a feasible implementation, reverse verification usually involves multiple steps such as data analysis, process review, and risk identification. First, clarify the purpose and criteria of reverse verification. For example, identify which links are most likely to lead to improper law enforcement or complaints; then trace back from the results (such as complaints, adverse consequences) to each link of the law enforcement process; data analysis tools (such as Excel, SQL, Python, etc.) can be used to analyze the data set to find patterns or trends related to risks; then, based on the results of data analysis, determine the risk indicators in the law enforcement process. These risk indicators may include the frequency, type, duration, personnel involved, etc. of law enforcement actions; then assign a risk value to each risk indicator to quantify its potential impact, and this operation can be carried out through methods such as expert scoring and historical data comparison; then, according to the risk value and occurrence probability, prioritize the risk indicators.

[0024] S103, the preset rule engine conducts comparative analysis on the risk indicators to obtain the compliance evaluation result of the law enforcement action, and performs status transition on the compliance evaluation result according to the preset state set.

[0025] In a feasible implementation, the rule engine defines the evaluation criteria for the compliance of law enforcement actions, which may include legal and regulatory requirements, law enforcement procedure regulations, principles for the protection of citizens' rights and interests, etc.; then it uses the rule engine to conduct a comparative analysis of the collected risk indicators to determine whether the law enforcement actions comply with the preset compliance criteria. Among them, the risk indicators involve the legality of law enforcement entities, the compliance of law enforcement procedures, the protection of citizens' rights and interests, etc.; the rule engine will automatically process the data according to the embedded rules and generate a compliance evaluation result.

[0026] In a feasible implementation, a set of preset status sets can be defined according to the possible results of the compliance evaluation of law enforcement actions. For example, statuses such as "compliant", "pending review", "non-compliant", "rectified" can be defined, and each status has a clear meaning and corresponding processing flow; then according to the compliance evaluation result, determine the current status of the law enforcement action; then according to the preset status transition logic, transfer the law enforcement action from the current status to the next status. For example, if the evaluation result is "non-compliant", it may be transferred to the "pending review" or "rectified" status. Among them, during the status transition process, corresponding actions may need to be triggered, such as sending notifications, updating the database, calling APIs, etc.

[0027] In a feasible implementation, a monitoring mechanism can also be established to track the status transition of law enforcement actions in real time. If abnormal status transitions or compliance issues are found, the alarm mechanism will be triggered in a timely manner, and corresponding corrective measures will be taken. Then, regularly evaluate and optimize the status transition logic and the rule engine to ensure the accuracy and efficiency of the system.

[0028] In summary, for the method for early warning of administrative law enforcement compliance based on multi-source data fusion provided by the embodiments of the present application, by collecting data sets related to administrative law enforcement from multiple data sources, more comprehensive and richer information can be obtained, reducing data one-sidedness or errors caused by a single data source; through reverse verification of the data set, risk indicators in the law enforcement process can be extracted, and based on the extracted risk indicators, a risk early warning mechanism can be established to timely discover and warn of potential law enforcement risks, providing decision-making support for law enforcement; through the comparative analysis of risk indicators by the rule engine, non-compliances in law enforcement actions can be discovered in a timely manner, and corresponding measures can be taken for improvement, thereby enhancing the compliance of law enforcement actions; through the status transition management of the compliance evaluation result, dynamic tracking and monitoring of law enforcement actions can be realized, problems can be discovered and solved in a timely manner, and law enforcement efficiency can be improved.

[0029] Figure 2 It is a schematic flowchart of another method for early warning of law enforcement compliance provided by the embodiments of the present application.

[0030] As Figure 2As shown, the law enforcement compliance warning method includes but is not limited to the following steps: S201. Collect initial data related to administrative law enforcement from multiple autonomous data sources.

[0031] In a feasible implementation manner, it is first necessary to clarify the data collection objectives, that is, which initial data related to administrative law enforcement need to be obtained. This may include the identity information of law enforcement objects, violation records, records of law enforcement processes, feedback on law enforcement results, etc. These data sources include but are not limited to: government open data, third-party data platforms, internal data of law enforcement agencies, as well as social media and Internet data, where: Government open data is data related to administrative law enforcement publicly released by government departments, such as enterprise credit information, administrative penalty records, etc. These data have high authority and accuracy and are important sources for initial data collection; Third-party data platforms collect and organize data related to administrative law enforcement and provide them for public query. The data sources of these data platforms are extensive, but there may be certain errors and inaccuracies, which need to be verified when used; Internal data of law enforcement agencies usually includes a large amount of data related to administrative law enforcement, such as law enforcement records, case files, etc. These data have high confidentiality, but can be collected through internal channels on the premise of meeting relevant laws and regulations; There may also be data related to administrative law enforcement in social media and Internet data, such as public feedback on law enforcement actions, hot spots of public opinion attention, etc. Although these data have certain reference value, attention needs to be paid to the authenticity and reliability of the data.

[0032] Furthermore, according to the characteristics and collection objectives of the above initial data, select appropriate collection methods. For example, government open data can be downloaded through official websites or obtained through API interfaces; third-party data platforms can be obtained through query interfaces or crawler technologies; internal data of law enforcement agencies need to be applied for and extracted according to internal procedures; social media and Internet data can be collected through channels such as search engines and social media platforms.

[0033] Next, design or select appropriate collection tools according to the collection methods. For example, when using web crawler technology, a crawler program needs to be designed to capture web page data; when using an API interface, code needs to be written to call the interface and parse the returned data.

[0034] S202. Preprocess the initial data to obtain a normalized data set.

[0035] In a feasible implementation, heterogeneous data cleaning can be performed on the initial data. As an example, KNN (K-nearest neighbor) outlier detection can be used to identify and process outliers or noisy data in the initial data, which may be caused by equipment failures, data transmission errors, or human factors. Through KNN outlier detection, the distance between each data point and other data points can be calculated, and outliers can be identified based on the distance. These outliers can be removed or specially processed in subsequent data processing to improve the quality and accuracy of the data.

[0036] In a feasible implementation, TF-IDF (Term Frequency-Inverse Document Frequency) weighting can be used to extract and process key information in the initial data for heterogeneous data cleaning. For example, in the initial data related to administrative law enforcement, there may be a large amount of text descriptions or records. Through TF-IDF weighting, it can be identified which words or phrases appear frequently in the document and play an important role in distinguishing different documents. This information can be used for subsequent tasks such as text classification, abstract generation, or information extraction. At the same time, TF-IDF weighting can also help remove high-frequency words or noise words that are not important for document classification or analysis, thereby improving the efficiency and accuracy of data processing.

[0037] After completing the heterogeneous data cleaning of the initial data, according to business requirements and data characteristics, a metadata model is defined, which includes data attributes, data types, data lengths, data formats, etc.; then the cleaned data is mapped into the metadata model to generate structured metadata; the generated structured metadata can also be verified to ensure the accuracy and integrity of the data.

[0038] In a feasible implementation, the structured metadata can be stored in an appropriate storage medium, such as a relational database, a NoSQL database, etc.; a data management mechanism can also be established, including data backup, data recovery, data permission management, etc., to ensure the security and availability of the structured metadata.

[0039] In a feasible implementation, after obtaining the structured metadata, perform sharding processing and integrity verification on each piece of structured metadata. Sharding processing is the process of splitting large data into multiple smaller data segments (or shards), which helps with data transmission, storage, and parallel processing. Further, determine the size of each shard according to the characteristics of the data (such as size, type, importance, etc.) and the requirements of transmission and storage. Among them, the shard size should be appropriate, considering both transmission efficiency and avoiding increased processing complexity caused by excessive sharding. Then assign a unique identifier (such as a shard number, sequence number, etc.) to each shard so that the complete data can be reassembled at the receiving end. Then transmit the shards to the target location in sequence or in parallel and store them in the corresponding storage medium.

[0040] After completing the sharding processing, perform an integrity verification operation. CRC check (Cyclic Redundancy Check) can be used. Among them, append a CRC code after each shard of data as the verification information for that shard. After the receiving end receives the shard data, use the same CRC algorithm to verify the shard data. If the calculated CRC code is the same as the received CRC code, it is considered that the shard data is correct. Otherwise, it is considered that an error occurred during data transmission and error handling (such as retransmission) is required. For incorrect shard data, processing methods such as retransmission, discarding, or marking as incorrect can be adopted. When retransmitting, the sending end can be requested to resend the incorrect shard data until the correct data is received.

[0041] After sharding processing and integrity verification operations, obtain the verification data corresponding to the structured metadata.

[0042] In a feasible implementation, according to a preset mapping mode model, define the weight assignment strategy for each data source or data item. Weight assignment can be based on multiple factors, such as the reliability of the data source, the historical accuracy of the data, the freshness of the data (timestamp), and the diversity of the data sources. Then calculate the weight of each data source according to the defined strategy. For example, if a certain data source has a high historical accuracy, a higher weight is given. Multiply each data item by its corresponding data source weight. Calculate the sum of all weighted data items. Divide each weighted data item by the weighted sum to obtain a normalized data set.

[0043] In a feasible implementation, a Schema Mapping model can be used to align the verification data schemas of different data sources to ensure the correct correspondence between data items; then, according to the weight assignment strategy, weights are evaluated for the verification data of each data source based on the defined factors, which may involve using statistical methods, machine learning models, or expert systems to dynamically calculate weights and dynamically adjust the weights according to real-time feedback or new data source information; the calculated weights are applied to the corresponding data items to obtain weighted data; the weighted data from multiple data sources are merged into a unified data set; then the sum of the weighted data set is calculated; each weighted data item is divided by the sum to obtain a normalized data set.

[0044] In some embodiments, correlation modeling can be performed on verification data including law enforcement officers, cases, and legal provisions. For example, an association between a law enforcement officer and a case is established, which represents the relationship of the law enforcement officer's participation in the case (such as investigation, trial, etc. operations; record the specific time of participation); an association between a case and a legal provision is established, which represents the legal provision cited by the case (reflecting the degree of fit between the provision and the case content, and this degree of fit can be obtained by analyzing the potential relevance and importance between nodes to obtain the aggregated weight of node features, and this aggregated weight is used to predict the matching degree between the law enforcement officer and the case). Then, according to the results of the correlation modeling, a normalized data set is obtained.

[0045] S203, perform reverse verification on the data set to obtain the risk indicators of the law enforcement process in the data set.

[0046] In a feasible implementation, the scope of the law enforcement process that needs to be verified can be identified from the data set; then, according to the actual steps of the law enforcement process, taking the closing link of the scope of the law enforcement process as the starting point, each node in the scope of the law enforcement process is traversed in reverse to obtain the time stamp, document integrity detection result, and legal provision citation result related to the node; then, graph attention network modeling is performed on the time stamp, document integrity detection result, and legal provision citation result to obtain the risk indicators of the law enforcement process in the data set.

[0047] In some embodiments, the timestamp, the document integrity detection result, and the legal clause citation result can be integrated into graph-structured data (the graph-structured data includes different types of nodes and edges; the nodes include law enforcement records, timestamps, document integrity detection results, and legal clause citation results; the edge types include the association between law enforcement records and timestamps, the association between law enforcement records and document integrity detection results, and the association between law enforcement records and legal clause citation results). The graph-structured data is modeled using a graph attention network to obtain a first ratio of nodes with discontinuous timestamps to the total number of nodes, a second ratio of nodes with tampered document content to the total number of nodes, and a third ratio of nodes with semantic similarity lower than a preset similarity threshold to the total number of nodes. Further, the integrity of the relevant documents of each node is detected to obtain the current hash value of the document; the historical hash value of the relevant document is compared with the current hash value, and the nodes with mismatched hash values are counted; the nodes with mismatched hash values are used as the nodes with tampered document content to obtain the second ratio of the nodes with tampered document content to the total number of nodes. Then, the first ratio, the second ratio, and the third ratio are weighted and averaged to obtain the risk indicator of the law enforcement process in the data set.

[0048] It should be noted that reverse verification is a process verification based on the case type. For ordinary cases (classified according to the nature of the case, the degree of social harm, and the punishment intensity), starting from the current stage of the ordinary case, the previous stage is checked in reverse (for example, if the current stage is punishment execution, the integrity of "filing → investigation → decision" needs to be verified), and it is checked whether the documents in each stage exist and the timestamps are continuous; if any link is missing, a "procedure jump" warning is triggered; if it is compliant, the next stage is allowed to proceed. For major cases, the legal review opinions and collective discussion records are verified; it is detected whether the "facts, reasons, basis, and right of defense" fields are included in the notice; if the verification fails, a "major procedural violation" is triggered, and then the case is warned and frozen; if the inspection is successful, the next stage is allowed to proceed.

[0049] It should be added that in some embodiments, if the warning state is triggered, the case will be pushed to the law enforcement department, and the law enforcement department will submit a feedback report, which will be reviewed by the department at the next higher level of the law enforcement department. Among them, the department at the next higher level can use the BERT model to analyze the text similarity between the feedback report and the preset evaluation rules. If the text similarity ≥ the threshold (as an example, the threshold can be 0.7), the feedback report will be adopted, the warning will be closed, and the law enforcement department will continue to handle the case; if the text similarity < the threshold, it is considered that the reason for the feedback is insufficient, the warning will be triggered again, and the law enforcement department will be notified for secondary processing.

[0050] S204. Compare and analyze the risk indicators by a preset rule engine to obtain the compliance evaluation result of the law enforcement behavior, and perform state transition on the compliance evaluation result according to the preset state set.

[0051] In a feasible implementation manner, business rules can be determined according to the condition-action structure. For example, if "the law enforcement exceeds the authority and is not approved in a single time, it is determined as illegal law enforcement". Then match the historical administrative law enforcement data with the business rules to obtain a violation case library, and use the violation case library as the rule engine.

[0052] In a feasible implementation manner, legal provisions related to administrative law enforcement can be converted into a Rete rule network (a Rete rule network is a network composed of nodes and edges connecting these nodes. Each node represents a condition or action in the rule, and the edges represent the logical relationships between the conditions). Then, use a fuzzy membership function to quantify the fuzzy expressions in the Rete rule network into a compliance degree interval. As an example, map fuzzy concepts such as "compliance", "partial compliance", "non-compliance", etc. to specific numerical intervals. For example, when the value of a certain indicator falls within [90, 100], it is considered fully compliant, and the membership degree is 1; when the value of a certain indicator falls within [70, 90], it is considered partially compliant, and the membership degree varies between 0 and 1, and the specific value is determined according to the position of the indicator value in this interval; when the value of a certain indicator is below 70, it is considered non-compliant, and the membership degree is 0. In the Rete rule network, map the condition part of each rule to the corresponding fuzzy membership function. This may require converting the legal provisions into a format that can match the membership function. For example, if the rule condition is based on a certain numerical indicator, the numerical value can be directly used as the input of the membership function. If the condition is based on the description of legal provisions (such as "low risk", "medium risk", etc.), these legal provision descriptions need to be first converted into numerical values or levels, and then mapped to the membership function. Then, perform cluster analysis on the historical administrative law enforcement data by the compliance degree interval to obtain a violation case library, and use the violation case library as the rule engine.

[0053] Since the historical administrative law enforcement data includes information such as violation types, violation times, violation locations, and violation degrees, use statistical methods to mine the laws and trends behind the historical violation data. For example, the high occurrence frequency of violation types, the distribution characteristics of violation times, the hot spots of violation locations, etc. can be analyzed; according to the results of the statistical analysis, dynamically adjust the warning thresholds related to the law enforcement process. For example, if the high occurrence frequency of a certain violation type increases, the warning threshold of this type can be appropriately reduced to detect and intervene in potential violations earlier; when a loophole or problem is found in a certain link, the warning threshold can be adjusted accordingly to enhance the monitoring and warning capabilities of this link, thereby improving the accuracy of reverse verification.

[0054] Furthermore, historical administrative law enforcement data can be cleaned to remove missing values and outliers to ensure data integrity and accuracy; appropriate clustering algorithms such as K-means, hierarchical clustering, etc. can be selected to perform clustering analysis on historical administrative law enforcement data by compliance degree intervals to obtain a violation case library; then key violation features can be extracted from the violation case library as the antecedents of the rules; according to the handling results of the cases and regulatory requirements, the consequents of the rules, i.e., corresponding law enforcement actions or penalty measures, can be determined; the extracted rules can be converted into a format recognizable by a computer, such as conditional statements in a programming language or the dedicated syntax of a rule engine; then based on the dedicated syntax, a suitable rule engine framework can be selected to build a rule engine.

[0055] Next, the risk indicator data is input into the rule engine obtained in the above embodiment, and the rule engine performs reasoning and judgment based on the input data and the defined rules to obtain the compliance evaluation result of the law enforcement action.

[0056] In a feasible embodiment, entity relationship extraction can also be performed on historical administrative law enforcement data by manual review opinions to obtain the relationships between entities; the relationships are fed back into the rule engine to iteratively optimize the rule engine.

[0057] In a feasible embodiment, according to the finite state machine model, a state set of compliance evaluation operations and an event set that triggers state transitions can be determined. Among them, the state set includes not reminded, reminded, pending review, adopted, and not adopted; the event set includes violation detection, manual review, feedback submission, and threshold update; then the corresponding relationship between the elements of the state set and the event set can be determined. As an example, a Petri net can be used to determine the corresponding relationship between the elements of the state set and the event set. Since a Petri net consists of places, transitions, and flow relations, in the Petri net of compliance evaluation: places can correspond to the states of the finite state machine model; transitions can correspond to the events of the finite state machine model; the flow relation defines the connection between places and transitions, indicating the state transition. Then, based on the corresponding relationship, state transitions are performed on the compliance evaluation result.

[0058] In some embodiments, when a violation detection event occurs, if the current state is "not reminded", it transitions to the "reminded" state; when a manual review event occurs, if the current state is "reminded", it transitions to the "pending review" state; when a feedback submission event occurs, if the current state is "pending review", if the review is passed, it transitions to the "adopted" state; if the review fails, it transitions to the "not adopted" state; when a threshold update event occurs, the violation detection event, the manual review event, and the feedback submission event are re-reviewed.

[0059] In some embodiments, the "not reminded" status shown in the compliance assessment result is used as the initial status; if a violation detection event occurs, the initial status of the compliance assessment result is transferred to the "reminded" status; if a manual review event occurs, the "reminded" status of the compliance assessment result is transferred to the "pending review" status; if the review is passed, the "pending review" status of the compliance assessment result is transferred to the "adopted" status; if the review fails, the "pending review" status of the compliance assessment result is transferred to the "not adopted" status, where the "adopted" status and the "not adopted" status are used as the end statuses; if the continuous event of the "pending review" status of the compliance assessment result exceeds the preset event threshold and is not transferred to the "adopted" status, the compliance assessment operation is transferred to the superior department for handling, and the current compliance assessment process is frozen.

[0060] In summary, for an administrative law enforcement compliance warning method based on multi-source data fusion provided by the embodiments of the present application, by collecting data sets related to administrative law enforcement from multiple data sources, more comprehensive and richer information can be obtained, reducing data one-sidedness or errors caused by a single data source; by performing reverse verification on the data sets, risk indicators in the law enforcement process can be extracted, and based on the extracted risk indicators, a risk warning mechanism can be established to timely discover and warn potential law enforcement risks, providing decision-making support for law enforcement; by comparing and analyzing the risk indicators through a rule engine, non-compliance in law enforcement behaviors can be timely discovered, and corresponding measures can be taken for improvement, thereby improving the compliance of law enforcement behaviors; by managing the status transfer of the compliance assessment results, dynamic tracking and monitoring of law enforcement behaviors can be achieved, timely discovering and solving problems, and improving law enforcement efficiency.

[0061] Figure 3 It is a schematic structural diagram of an enforcement compliance warning system provided by the embodiments of the present application.

[0062] As Figure 3 shown, the enforcement compliance warning system 300 includes: A collection module 301, which is used to collect data sets related to administrative law enforcement based on multiple data sources; A first acquisition module 302, which is used to perform reverse verification on the data set to obtain risk indicators of the law enforcement process in the data set; A second acquisition module 303, which performs comparative analysis on the risk indicators by a preset rule engine to obtain the compliance assessment result of the law enforcement behavior, and performs status transfer on the compliance assessment result according to a preset status set.

[0063] Figure 4 It is a schematic structural diagram of an electronic device provided according to the embodiments of the present application.Figure 4 The electronic device shown is merely an example and should not impose any limitations on the functions and usage scope of the embodiments of this application.

[0064] As Figure 4 shown, the electronic device 400 includes a processor 401, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 402 or the program loaded from the memory 406 into the random access memory (RAM) 403. In the RAM 403, various programs and data required for the operation of the electronic device 400 are also stored. The processor 401, the ROM 402, and the RAM 403 are connected to each other via a bus 404. The input / output (I / O) interface 405 is also connected to the bus 404.

[0065] The following components are connected to the I / O interface 405: a memory 406 including a hard disk, etc.; and a communication part 407 including a network interface card such as a LAN (Local Area Network) card, a modem, etc., and the communication part 407 performs communication processing via a network such as the Internet; a drive 408 is also connected to the I / O interface 405 as required.

[0066] Specifically, according to the embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments of this application include a computer program carried on a computer-readable medium, and the computer program includes program codes for executing the methods shown in the flowcharts. In such an embodiment, the computer program can be downloaded and installed from the network via the communication part 407. When the computer program is executed by the processor 401, the above functions defined in the methods of this application are executed.

[0067] In an exemplary embodiment, a storage medium including instructions is also provided, such as a memory including instructions, and the above instructions can be executed by the processor 401 of the electronic device 400 to complete the above methods. Optionally, the storage medium can be a non-transitory computer-readable storage medium. For example, the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device, etc.

[0068] In this application, a computer-readable storage medium can be any tangible medium that contains or stores a program, which can be used by or in conjunction with an instruction execution system, apparatus, or device. And in this application, a computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on a computer-readable medium can be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination of the above.

[0069] Finally, it should be noted that the above embodiments are only used to illustrate rather than limit the technical solutions of this application. Any equivalent replacement and modification or partial replacement that do not deviate from the spirit and scope of this application should be covered within the scope of the claims of this application.

Claims

1. A method for early warning of administrative law enforcement compliance based on multi-source data fusion, characterized in that, Including: Collecting data sets related to administrative law enforcement based on multiple data sources; Performing reverse verification on the data sets to obtain risk indicators of the law enforcement process in the data sets; Comparatively analyzing the risk indicators by a preset rule engine to obtain a compliance evaluation result of the law enforcement behavior, and performing state transition on the compliance evaluation result according to a preset state set.

2. The method according to claim 1, wherein The collecting data sets related to administrative law enforcement based on multiple data sources includes: Collecting initial data related to administrative law enforcement from multiple autonomous data sources; Performing heterogeneous data cleaning on each of the initial data to obtain structured metadata; Performing sharding processing and integrity verification on each of the structured metadata to obtain verification data; Performing dynamic weight assignment on the verification data from different data sources by a preset mapping mode model to obtain a normalized data set.

3. The method according to claim 1, wherein The performing reverse verification on the data sets to obtain risk indicators of the law enforcement process in the data sets includes: Identifying the scope of the law enforcement process to be verified from the data sets; Taking the closing link of the scope of the law enforcement process as the starting point according to the actual steps of the law enforcement process, and traversing each node of the scope of the law enforcement process in reverse to obtain the timestamp, document integrity detection result, and legal clause citation result related to the node; Performing graph attention network modeling on the timestamp, the document integrity detection result, and the legal clause citation result to obtain risk indicators of the law enforcement process in the data sets.

4. The method according to claim 3, wherein The performing graph attention network modeling on the timestamp, the document integrity detection result, and the legal clause citation result to obtain risk indicators of the law enforcement process in the data sets includes: Integrating the timestamp, the document integrity detection result, and the legal clause citation result into graph structure data; Using a graph attention network to model the graph structure data to obtain a first ratio of nodes with discontinuous timestamps to the total number of nodes, a second ratio of nodes with tampered document content to the total number of nodes, and a third ratio of nodes with semantic similarity lower than a preset similarity threshold to the total number of nodes; Performing weighted average on the first ratio, the second ratio, and the third ratio to obtain risk indicators of the law enforcement process in the data sets.

5. The method according to claim 3, characterized in that The obtaining the second ratio of nodes with tampered document content to the total number of nodes includes: Performing integrity detection on the relevant documents of each node to obtain the current hash value of the document; Comparing the historical hash value of the relevant document with the current hash value, and counting the nodes with mismatched hash values; Taking the nodes with mismatched hash values as the nodes with tampered document content to obtain the second ratio of nodes with tampered document content to the total number of nodes.

6. The method according to claim 1, wherein The comparatively analyzing the risk indicators by a preset rule engine to obtain a compliance evaluation result of the law enforcement behavior includes: Determining business rules according to the condition-action structure; Matching historical administrative law enforcement data and business rules to obtain a violation case library, and using the violation case library as the rule engine; Input the risk indicators into the rule engine to obtain the compliance evaluation result of the law enforcement behavior.

7. The method according to claim 6, wherein It further includes: Extract the entity relationships from the historical administrative law enforcement data based on the manual review opinions to obtain the relationships between entities; Feed back the relationships to the rule engine to iteratively optimize the rule engine.

8. The method according to any one of claims 1-7, characterized in that, The state transition of the compliance evaluation result according to the preset state set includes: Determine the state set of the compliance evaluation operation and the event set that triggers the state transition according to the finite state machine model, where the state set includes not reminded, reminded, pending review, adopted, and not adopted; the event set includes violation detection, manual review, feedback submission, and threshold update; Determine the corresponding relationship between the elements of the state set and the event set; Perform state transition on the compliance evaluation result based on the corresponding relationship.

9. The method according to claim 8, characterized in that The determination of the corresponding relationship between the elements of the state set and the event set includes: When the violation detection event occurs, if the current state is "not reminded", then transition to the "reminded" state; When the manual review event occurs, if the current state is "reminded", then transition to the "pending review" state; When the feedback submission event occurs, if the current state is "pending review", if the review is passed, then transition to the "adopted" state; if the review is not passed, then transition to the "not adopted" state; When the threshold update event occurs, re-review the violation detection event, the manual review event, and the feedback submission event.

10. The method according to claim 8, wherein The performance of state transition on the compliance evaluation result based on the corresponding relationship includes: Take the "not reminded" state displayed in the compliance evaluation result as the initial state; If the violation detection event occurs, transition the initial state of the compliance evaluation result to the "reminded" state; If the manual review event occurs, transition the "reminded" state of the compliance evaluation result to the "pending review" state; If the review is passed, transition the "pending review" state of the compliance evaluation result to the "adopted" state; if the review is not passed, transition the "pending review" state of the compliance evaluation result to the "not adopted" state, where the "adopted" state and the "not adopted" state are used as the end states; If the continuous event of the "pending review" state of the compliance evaluation result exceeds the preset event threshold and does not transition to the "adopted" state, transfer the compliance evaluation operation to the superior department for handling and freeze the current compliance evaluation process.

Images