Big data legal supervision model atomic operation and assembly method

Abstract

The application discloses a big data legal supervision model atomic operation and assembly method, comprising: an atomic operation layer, which is used for realizing basic functions in a legal supervision process; a supervision model layer, which comprises a plurality of legal supervision models for different supervision targets; and an early warning push layer, which is used for sending early warning information. The application defines preconditions and postconditions of atomic operations, uses predicate logic to establish logical relationships between the atomic operations, and further constructs a complete legal supervision model. This method based on predicate logic makes the model have certain interpretability, and dynamic assembly of the atomic operations enhances flexibility and generalization of the model. The application uses formalization description capability of the predicate logic to establish logical correlations between legal thresholds, behaviors of supervision objects and supervision targets, realizes fine analysis and dynamic evaluation of the behaviors of the supervision objects, and displays visualization of an atomic operation path of reachable supervision targets.

Description

Technical Field

[0001] This invention relates to the field of legal supervision model technology, and in particular to the application of big data in legal supervision models. Background Technology

[0002] With the continuous development of information technology, big data technology has been widely applied in various fields. In the field of legal supervision, the application of big data technology has also attracted widespread attention. Compared with traditional manual supervision, big data legal supervision has stronger coverage, real-time performance, and accuracy, and can better discover abnormal behaviors and potential risks hidden in massive amounts of data.

[0003] Existing big data legal supervision models mainly include the following representative schemes:

[0004] 1. A Big Data Legal Supervision Model Based on Machine Learning

[0005] These models typically leverage the vast amounts of historical regulatory data accumulated by regulatory agencies, combining it with techniques such as natural language processing and graph mining to build regulatory early warning models. For example, some models analyze unstructured data such as company announcements and news reports to identify keywords or patterns that may indicate violations and issue warnings for these risks. Other models use structured data such as corporate financial data and transaction records, employing machine learning algorithms such as anomaly detection and clustering to discover potential violations. These machine learning-based models improve the coverage and timeliness of regulation to some extent, but they also face challenges such as a lack of interpretability and susceptibility to data bias.

[0006] 2. Big Data Legal Supervision Model Based on Knowledge Graph

[0007] These models construct knowledge graphs covering multiple domains, including legal norms, regulatory policies, and corporate information, and utilize techniques such as graph reasoning and semantic analysis to discover complex legal risk relationships. Compared to single machine learning models, this knowledge graph-based approach offers stronger interpretability and reasoning capabilities. However, it also suffers from high knowledge acquisition costs and complex graph construction, limiting its practical application.

[0008] 3. Blockchain-based Big Data Legal Supervision Model

[0009] Blockchain technology, with its immutability and traceability, is considered highly suitable for the field of big data-driven legal oversight. Some exploratory studies have attempted to use blockchain to record various data points in the regulatory process, creating immutable regulatory traceability. Simultaneously, blockchain's smart contract functionality can be used to automate the execution of regulatory rules, improving regulatory efficiency. However, the performance bottlenecks and privacy protection issues inherent in blockchain technology currently limit its application in large-scale legal regulatory scenarios.

[0010] In general, existing big data legal supervision models still face some challenges in terms of standardization, data sharing, and model interpretability. For example, different regulatory departments use different regulatory standards and data formats, making it difficult to achieve cross-departmental joint supervision; the ownership of regulatory data also restricts the sharing and utilization of data; and the "black box" nature of machine learning models reduces the interpretability and credibility of regulatory results. Summary of the Invention

[0011] To overcome the shortcomings of existing technologies, the main objective of this invention is to provide an atomic operation and assembly method for big data legal supervision models, addressing several key issues in existing regulatory models. First, they lack modularity and flexibility. Existing regulatory models are typically end-to-end designs, making it difficult to customize and adjust them according to the needs of different regulatory scenarios. This not only limits the applicability of the models in different regulatory fields but also reduces their interpretability. Second, the models lack adaptability. As regulatory practice deepens, regulatory data and needs are constantly changing. However, static models often struggle to absorb new regulatory experiences in a timely manner, resulting in difficulties in continuously improving regulatory effectiveness. Finally, interpretability needs enhancement. Most existing models are based on black-box machine learning algorithms, making it difficult to explain their internal logic. This limits the understanding and trust of regulatory authorities in the model's decisions.

[0012] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0013] A legal supervision system based on an atomic operation big data model includes:

[0014] The atomic operation layer includes multiple independently executable atomic operation units, which are used to implement basic functions in the legal supervision process, including data reading, information extraction, anomaly detection, authorization verification, and error handling.

[0015] The supervision model layer includes several legal supervision models targeting different supervision objectives. These models are constructed using a predicate logic-based atomic operation assembly method, defining preconditions and postconditions for each atomic operation. Based on different legal supervision objectives, the supervision model layer utilizes this predicate logic-based atomic operation assembly method to build legal supervision models for specific scenarios. These models can adapt to different regulatory needs and data characteristics.

[0016] The early warning push layer is used to send early warning information to improve the timeliness and accuracy of legal supervision.

[0017] In one embodiment, the atomic operation unit of the atomic operation layer includes:

[0018] The pre-operation unit, in the preparation phase before the legal supervision process begins, acquires data related to legal events and senses the environment; it reads and senses information from databases including personnel information, monitoring information, legal documents, judicial judgments, case registrations, administrative penalty records, and other data related to legal events; it obtains the necessary data from various sources, including government databases (such as courts, procuratorates, and public security bureaus), public records (such as business registrations and social security information), social media data, news reports, and other information sources related to legal events, including basic case information (such as case information and party information), legal documents, penalty records, historical case data, and other data closely related to legal events.

[0019] After the operation unit is run, in the analysis and feedback phase after the supervision process ends, it performs content comparison, threshold setting and result in-depth analysis, and conducts in-depth analysis of the detected abnormal results, such as statistical analysis and pattern recognition of abnormal transaction amounts, abnormal action trajectories, and frequency of occurrence, to determine whether they reflect potential illegal behavior or risks.

[0020] During the operation phase of the monitoring process, the operating unit performs information extraction. Based on the real-time extracted information, such as names, vehicles, companies, amounts, times, and locations, and according to the abnormal thresholds set by prior knowledge, it performs real-time detection of potential problems and directly pushes some simple judgments to the early warning push layer, such as the use of medical insurance for deceased persons.

[0021] In one embodiment, the operations of the pre-run operation unit include:

[0022] The data read operation connects to different databases based on the selected database type, reads the stored information, and stores it.

[0023] Environment-aware operations acquire current terminal environment information to perform corresponding atomic operations based on different environments, such as adapting to reading from the operating system and database;

[0024] Permission-aware operation: Using static and dynamic permission-aware methods, it is determined whether the person currently performing the operation has the permission to perform this operation. The dynamic permission-aware method is based on the time factor.

[0025] Error-aware operations determine whether an error has occurred during execution, thus affecting the normal operation of the supervised model. When an error is detected, the error location is recorded and saved.

[0026] In one embodiment, the operation of the post-run operation unit includes:

[0027] Content comparison involves comparing information related to business implementation extracted during the operation process, and extracting various types of information including similarities, differences, and frequency.

[0028] Threshold setting: Thresholds are set for the various types of information. When the statistical data, such as information frequency, amount statistics, time interval, data distribution, etc., exceed the set thresholds, they are marked as abnormal data and an early warning is triggered.

[0029] In one embodiment, the legal supervision model is constructed using the following method:

[0030] Define the preconditions and postconditions of atomic operations; the preconditions refer to the conditions that must be met before a specific atomic operation is executed, and these conditions ensure that the execution of the operation is reasonable and effective. Only when these conditions are met can the atomic operation proceed smoothly; the postconditions refer to the conditions that need to be met after a specific atomic operation is executed. They are used to verify whether the execution result of the operation meets expectations and to ensure the correctness and effectiveness of the operation. The specific definitions of the two are mainly based on requirements analysis, condition identification, and logical reasoning.

[0031] Predicate logic is used to establish logical relationships between atomic operations, and reachable atomic operation paths are sought based on different supervision objectives;

[0032] By assembling atomic operation units that meet the preconditions, a complete legal supervision model process is constructed, ensuring the model's correctness and interpretability.

[0033] In one embodiment, the process of establishing logical relationships between atomic operations using predicate logic, seeking reachable atomic operation paths based on different supervision objectives, assembling atomic operation units that satisfy preconditions, and constructing a complete legal supervision model flow is implemented as follows:

[0034] By using predicate logic to establish logical relationships between atomic operations, first, the input, output, and function of each atomic operation are clarified, and the sequential relationship between operations is defined using predicate logic; then, logical reasoning is used to verify whether the logical relationships between each operation are consistent, ensuring that no contradictions or errors occur during execution.

[0035] Based on the specific monitoring objectives, analyze the required final results and intermediate steps, determine the required types of atomic operations, and find ways to achieve the monitoring objectives using both active and passive methods.

[0036] By assembling atomic operation units that satisfy specific data inputs or environmental states, and arranging atomic operation units that conform to logical relationships and preconditions in sequence to form a complete operation sequence, a complete legal supervision model is constructed that includes various operations before, during, and after operation and can achieve the supervision objectives.

[0037] In one embodiment, atomic operations are selected and assembled based on various characteristics, constrained by supervisory business logic, functions, and domain knowledge. First, a comprehensive analysis of the supervisory business logic, functions, and domain knowledge is performed to clarify the required functions and objectives, guiding the selection of atomic operations. Then, based on the business logic and domain knowledge, atomic operations that meet the current supervisory objectives are selected. Finally, the selected atomic operations are assembled according to logical relationships in a bottom-up hierarchical structure, ensuring the coherence of data flow and control flow at each level, resulting in the legal supervision model. The legal supervision model has four levels from bottom to top: information source, preliminary information, intermediate judgment, and result judgment. Information sources include open-source data, government regulatory data, and specific data sources that are not open-source or government regulatory data, potentially including industry internal data, social media data, etc. Preliminary information refers to information obtained through information extraction operations. For a specific supervisory objective, the required basic information is first obtained from the information source, then logical judgments involving content comparison and threshold setting are performed, and finally, it is determined whether each piece of information exceeds the threshold and reaches an abnormal value, thereby achieving the final supervisory objective judgment.

[0038] In one embodiment, the early warning push layer utilizes visualization technology to provide a visual display and reporting function for the regulatory situation.

[0039] This invention also provides a method for atomic operation and assembly of a big data legal supervision model, used to form the aforementioned legal supervision system based on the atomic operation big data model, and further used to realize legal supervision, including the following steps:

[0040] First, based on the legal supervision model, dynamic permission awareness is used to determine whether permission is granted;

[0041] Then, various information extraction operations are performed from the information source to extract basic information;

[0042] Subsequently, some information is compared and processed to generate secondary information such as frequency.

[0043] Finally, based on the information obtained, it is determined whether any data exceeds the set threshold and becomes abnormal data. When an anomaly is found, the path of the above atomic operation is recorded to form a model that meets the supervision objectives. By recording the path of atomic operation, the system can flexibly adjust and optimize the legal supervision model used in subsequent operations to adapt to specific regulatory needs.

[0044] In one embodiment, the present invention can find a path to achieve the supervision objective through one of the following methods:

[0045] Method 1, Proactive

[0046] Throughout the entire process from data reading to early warning push, a legal supervision model involving the supervision target is autonomously constructed by combining all the atomic operations involved.

[0047] Method 2, Passive

[0048] Redundant analysis based on predicate logic is used to obtain the various atomic operation paths required to reach the abnormal data.

[0049] Compared to existing technologies, this invention subdivides the entire regulatory process into a series of atomic operation units, using predicate logic to describe the preconditions and postconditions between these atomic operations, thereby achieving refined modeling and dynamic assembly of the regulatory process. This greatly enhances the modularity and flexibility of the model, and based on interpretable atomic operation units and predicate logic descriptions, significantly improves the interpretability of the entire regulatory model, which is beneficial for regulatory authorities to understand and trust the model's decision-making process. Attached Figure Description

[0050] Figure 1 It is a diagram of atomic operations, segmentation, and three-dimensional structure.

[0051] Figure 2 This is a flowchart of the data reading operation.

[0052] Figure 3 This is a flowchart of the environmental perception operation.

[0053] Figure 4 This is a flowchart of the information extraction operation.

[0054] Figure 5 This is a flowchart of the operations after the program runs.

[0055] Figure 6 This is a flowchart of SMS alert push notifications.

[0056] Figure 7This is a diagram of the atomic operation input / output correlation analysis model.

[0057] Figure 8 This is a flowchart for constructing atomic operation paths. Detailed Implementation

[0058] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings and examples. It should be noted that the embodiments described herein are only for explaining the present invention and are not intended to limit the present invention. Furthermore, the technical features involved in the embodiments of the present invention can be combined with each other unless otherwise specified.

[0059] To address the aforementioned problems of this invention, a more flexible, interpretable, and adaptable big data legal supervision model needs to be designed. On the one hand, this model should possess good modularity and scalability, capable of adapting to the business needs and data characteristics of different regulatory departments. On the other hand, it should also have strong interpretability, accurately identifying potential violations and clearly explaining the basis for identification, thereby enhancing the credibility of regulatory results. Furthermore, the model should also possess a certain degree of self-adaptability, continuously optimizing and updating itself as regulatory practice evolves, thereby continuously improving the timeliness and accuracy of supervision.

[0060] To address this, this invention proposes a legal supervision system and method based on an atomic operation big data model. Its core idea is to subdivide the entire regulatory process into a series of atomic operation units, and use predicate logic to describe the preconditions and postconditions between these atomic operation units, thereby achieving refined modeling and dynamic assembly of the regulatory process. This not only improves the modularity and flexibility of the model but also lays the foundation for enhancing its interpretability. Furthermore, an incremental learning mechanism can be incorporated, enabling the model to continuously absorb new regulatory practice experience, improving the timeliness and accuracy of regulatory effectiveness.

[0061] In this invention, atomic operations refer to the basic units that constitute the regulatory process, such as data reading, information extraction, anomaly detection, authorization verification, and error handling. These atomic operation units have a certain degree of independence and can be flexibly assembled according to the needs of different regulatory scenarios. This invention divides the regulatory process into three layers; that is, the legal supervision system based on the atomic operation big data model mainly includes these three layers: the atomic operation layer, the supervision model layer, and the early warning push layer. Wherein:

[0062] The atomic operation layer can be further refined into multiple independently executable atomic operation units to implement basic functions in the legal supervision process, including data reading, information extraction, anomaly detection, authorization verification, and error handling.

[0063] The supervision model layer includes several legal supervision models targeting different supervision objectives. Based on these objectives, this invention utilizes a predicate logic-based atomic operation assembly method. By defining the preconditions and postconditions of atomic operations, and establishing the logical relationships between them using predicate logic, complete legal supervision models can be constructed for specific scenarios, i.e., different supervision objectives. These models can adapt to different regulatory needs and data characteristics. This predicate logic-based method not only ensures the effectiveness and interpretability of the models but also achieves flexibility and adaptability through the dynamic assembly of atomic operations.

[0064] The early warning push layer is mainly used to send early warning information to improve the timeliness and accuracy of legal supervision.

[0065] The specific implementation of this invention includes constructing atomic operation segmentation, designing a three-dimensional structure of the supervised model, implementing atomic operations, and assembling atomic operations for predicate logic. Figure 1 This is a block diagram of the big data legal supervision model based on atomic operations of this invention, for reference. Figure 1 It mainly includes the following contents:

[0066] 1. Construct a three-dimensional structure for the atomic operation partitioning and design supervision model.

[0067] Based on an examination of existing supervision models, the main problem facing the field of big data legal supervision is the difficulty in sharing and integrating data among multiple administrative departments, while the rational division of atomic operations also faces challenges. Therefore, the goal of this invention's atomic operation segmentation modeling design is to achieve fine-grained decomposition of the legal supervision model, decomposing it into combinations of atomic functions based on supervisory functions. Different atomic functions correspond to one atomic operation. Conversely, by combining different atomic operations, a legal supervision model with specific functions can be formed according to requirements, aiming to achieve fine-grained decomposition and flexible combination of the supervision model. Based on the previous requirements analysis, the atomic operation model of the legal supervision model has the following functions:

[0068] 1) It can improve the interpretability of the legal supervision model to some extent. Since the legal supervision model is constructed from a combination of atomic operation models, and the functions of atomic operations are clearly defined, the combination of atomic operations can realize the flow characteristics of data within the legal supervision model, thereby improving its interpretability. 2) It can improve cross-departmental collaboration to some extent. Due to the clarity and interpretability of atomic operations, it can encourage data contributions from different departments, thus improving inter-departmental collaboration. Based on these characteristics, atomic operation modeling in the legal supervision model needs to consider both the rationality of atomic operation segmentation and the analyzability and predictability of atomic operations.

[0069] The hierarchical structure composed of atomic operations predefined in this invention is as follows: Figure 1 As shown, the structure is roughly divided into three layers. The bottom layer consists of various atomic operations, such as data reading, information extraction, anomaly detection, permission verification, and error handling. The atomic operations have a single, clear function, which helps improve the interpretability of the supervision model. These atomic operations can also connect to data sources providing information, allowing for customized information extraction and analysis for different regulatory issues. Different information can be extracted to solve different specific problems, and different indicators can be weighted and judged under different evaluation standards. For example, in medical insurance events, information such as people, time, events, locations, purchase frequency, and purchase amount can be extracted, and abnormal information can be filtered out through content comparison and threshold judgment. The middle layer is the supervision model layer for specific regulatory scenarios, corresponding to the specific reasons for early warning cases, such as various illegal and irregular issues. The atomic operations at the bottom layer can be flexibly combined according to the requirements of different cases to form different supervision models for specific regulatory scenarios, such as Legal Supervision Model 1, Legal Supervision Model 2, and Legal Supervision Model 3 in the figure. The top layer is the event early warning layer, which performs the final abnormal information aggregation and alarm push operations, such as SMS alarm push operations for abnormal insurance fraud, and email push operations. This layer can aggregate the abnormal information extracted from previous model runs, package it into early warning information, and push it to relevant personnel. In the system of this invention, an atomic operation is only responsible for completing a specific function, which helps to improve the interpretability of the operation. Through this layered design, using combinations of different atomic operations to build a supervision model can improve the overall flexibility and interpretability of the supervision model, and also achieve refined control over different regulatory areas. The big data legal supervision model design method based on atomic operations is expected to break through many bottlenecks existing in current regulatory models.

[0070] 2. Specific implementation of atomic operations

[0071] Based on the division of atomic operations in terms of operation time, the atomic operation units of the atomic operation layer of the present invention include: pre-run operation units, post-run operation units, and in-run operation units.

[0072] Before the legal supervision process begins, the pre-operation unit acquires relevant data and assesses the environment for legal events during the preparation phase. The main function of data acquisition is to connect to different databases based on the selected database type, read the various information stored in the databases, and store this data for subsequent targeted extraction and further processing. The databases used include personnel information, monitoring information, legal documents, judicial judgments, case registrations, administrative penalty records, and other data related to legal events. The unit acquires necessary data from various sources, including government databases (such as courts, procuratorates, and public security bureaus), public records (such as business registrations and social security information), social media data, and news reports. This data includes basic case information (such as case details and party information), legal documents, penalty records, historical case data, and other data closely related to legal events.

[0073] Figure 2 This is a flowchart of the data reading operation. In this operation, different problems and different companies may choose different databases as information sources, so the first step is to determine the type of database to connect to. Currently, seven databases can be connected: Kingbase, PostgreSQL, MySQL, GaussDB, Oracle, SQLite, and MSSQL. A redundant approach is used in the implementation. Corresponding connection functions are written for each database type. Users need to enter the required database information as prompted; for example, for MySQL, the hostname, port number, username, password, and database name are required. After a successful connection, the information in the database will be temporarily stored in a Python cursor for further extraction and processing.

[0074] The pre-run operation also includes a series of pre-run perception operations. The main function of the environment perception operation is to obtain various environmental information of the current terminal, so as to facilitate the execution of different operations in different environments, making the operation more adaptable and able to run well in different environments. Figure 3Shows the flowchart of the environmental perception operation. Usually as the first step after model execution, this operation obtains the environmental information of the current device and software to ensure the effective operation of subsequent programs. In this operation, by calling the platform library in Python, the current environment can be sensed, and basic environmental information such as the name of the operating system, the version of the operating system, the version number of the current Python interpreter, the name of the machine's hardware, the name of the processor, and the network name of the current system can be extracted. This is convenient for using redundant methods later to perform corresponding operations according to different environments, thereby improving the adaptability of the operations and ensuring that the system can operate well in various environments and enhancing robustness. At the same time, this information will also be stored in a txt file for subsequent query and verification.

[0075] The main function of the permission perception operation is to judge whether the person performing the current operation has the permission to perform this operation. This permission perception operation not only uses static permission perception methods such as account passwords to judge personnel permissions, but also introduces dynamic permission judgments, that is, time factors. For the same user (a set of username and password), they may also have different permissions at different times, that is, permissions may be updated dynamically over time. Introducing a time variable in the permission perception operation can more effectively judge the user's permissions at the current moment. Finally, the main function of the error perception operation is to judge whether an error has occurred due to human operation errors or other reasons during the user's execution process, resulting in the failure of the supervision model to operate normally. If the above situation occurs, the model will save and record the location where the error occurred for subsequent operators to analyze and modify.

[0076] During the operation stage of the running operation unit during the supervision process, according to the supervision cases and supervision objectives, various types of information closely related to specific operations are extracted from various information sources (i.e., different types of databases) in the previous step. For different business combinations, the supervision model may need to extract different types of information. To cope with the diversity of information types in different operations, these operations are further divided into multiple small operations. For example, if the target case is a medical insurance case, information such as insured persons, medicine purchasers, pharmacy names, purchase amounts, and types of purchased medicines in the database needs to be extracted. Since there are a large number of information types for different operations, but the processes of these operations are basically the same. For example, person name information extraction, company information extraction, hospital information extraction, amount information extraction, etc. The main function of these atomic operations is to read data from the database and extract information specifically. Depending on the extraction, this information may be directly used for early warning of abnormal situations, such as using the ghost medical insurance of deceased persons in medical insurance issues, or only storing preliminary information for subsequent further comparison or judgment, such as exceeding the amount limit or multiple card swipes among family members.

[0077] Figure 4 This is a flowchart of the information extraction operation. In this operation, firstly, based on the case and the monitoring objective, it's necessary to determine what kind of information extraction is required; information irrelevant to the case is not extracted. Then, based on the data saved in the database from the data reading operation, the corresponding information is extracted. In some information extractions, such as monetary information extraction, sorting the extracted information in descending order has several advantages. First, for target information like monetary amounts, descending order makes it easier to identify anomalies and facilitates threshold judgment. Second, when the extraction target is a group of information containing many details, sorting them by time makes the results more readable and facilitates subsequent operations. Third, when obtaining information such as personnel frequency, sorting names in descending order at this step groups information about the same person together, facilitating subsequent content comparison operations to calculate frequency information. After the above process is completed, the information extraction operation stores the extracted primary information in a txt file for subsequent content comparison operations to obtain secondary information such as similarities, differences, and frequency information.

[0078] The post-operation unit's analysis and feedback phase after the monitoring process concludes is broadly divided into two parts: content comparison and threshold setting. Content comparison involves comparing the content of various business-related information extracted during operation, extracting secondary information including similarities, differences, and frequency, and storing this secondary information for further evaluation to confirm the presence of any anomalies. Threshold setting is the crucial operation for determining whether information is abnormal. Its main function is to set thresholds for various types of information; when certain data exceeds the set thresholds after statistical analysis, it is classified as abnormal data and an alert is issued.

[0079] For certain information, such as extracted monetary amounts, content comparison may not be necessary; a threshold setting can be used to determine if it exceeds a certain threshold. However, for other information, such as medical insurance cases, if someone uses their medical insurance for their entire family, it may result in a large number of medical insurance reimbursement records in a short period. To issue an alert, content comparison is needed to analyze the frequency of the individual's medical insurance reimbursements and compare it with the set threshold to determine if it is abnormal data. Threshold setting involves setting thresholds for various types of information. When the statistically analyzed data, such as information frequency, monetary statistics, time intervals, and data distribution, exceeds the set threshold, it is marked as abnormal data and an alert is triggered.

[0080] Figure 5This is a flowchart illustrating the entire process of content comparison and threshold setting after execution. In this operation, the information about the supervised cases extracted in the previous information extraction step is first input. Then, based on actual requirements, secondary information such as similarities, differences, and frequency is extracted. This can be achieved using the jieba and collections libraries in Python. Next, the acquired frequency and other information are sorted in descending order. Similar to the first step of descending order in the information extraction operation, sorting this information in descending order makes it easier to identify anomalies and facilitates threshold judgment. Then, based on the actual situation of supervision, thresholds are set for each piece of information, and the processed information is compared with the set thresholds. If abnormal data exceeding the thresholds is found, this data is recorded and summarized in the alert push.

[0081] Furthermore, in-depth analysis can be introduced, specifically, to conduct in-depth analysis of detected abnormal results, such as statistical analysis and pattern recognition of abnormal transaction amounts, abnormal movement trajectories, and frequency of occurrence, in order to determine whether they reflect potential illegal activities or risks.

[0082] The top-level early warning push layer of the monitoring model, serving as the result layer, primarily functions to extract and summarize abnormal information after the model's execution, then package it and push the early warning information to relevant personnel via SMS, email, or other communication methods. Furthermore, it can utilize visualization technology to provide visual displays and reporting functions for the regulatory situation.

[0083] Figure 6 This is a flowchart for SMS alert push notifications. The alert push operation primarily involves connecting to an external communication platform to send SMS messages to the target mobile phone. First, it needs to read and record the abnormal alert information output from the previous step. Then, it needs to obtain the communication platform's SMS signature, request address, and request parameters, and incorporate these into the message sending function. Finally, this function is called, accepting a username and password. An access control check determines if the user has permission to send messages. If permission is granted, the SMS sending operation is executed, and the sent parameters and SMS information are returned after successful transmission.

[0084] 3. Assembly of atomic operations based on predicate logic

[0085] Based on information initially extracted from the database, this invention studies how to achieve legal supervision goals through the combination of atomic operations, constructing atomic operation paths for a legal supervision model—a global analysis of legal supervision. This invention, based on a predicate logic-based atomic operation input / output correlation analysis method, proposes an atomic operation path that starts from basic initial conditions such as names and combines a series of atomic operations to reach the supervision target. The core idea of ​​this principle is to utilize the relationship between atomic operation inputs and outputs; that is, subsequent atomic operations occur on the premise of previous atomic operations, such as content comparison operations. This, in turn, increases the categories of input information, including frequency information, similarity information, and threshold information. Furthermore, this invention proposes an atomic operation input / output correlation analysis model. By utilizing the output information of the previous atomic operation and the input information required for the subsequent atomic operation, it discovers atomic operation paths in the legal supervision model that can achieve the supervision goal.

[0086] To construct a legal supervision model, this invention requires defining the preconditions and postconditions of atomic operations. Then, predicate logic is used to establish the logical relationships between atomic operations, clarifying the inputs, outputs, and functions of each atomic operation. Predicate logic is used to define the pre- and post-operational relationships between operations, and logical reasoning is used to verify the consistency of these relationships, ensuring that no contradictions or errors occur during execution. This allows for the search of achievable atomic operation paths based on different supervision objectives. Specifically, based on the specific supervision objective, the required final result and intermediate steps are analyzed to determine the necessary atomic operation types and find ways to achieve the supervision objective. Finally, atomic operation units that satisfy specific data inputs or environmental states are assembled and arranged sequentially according to the logical relationships and preconditions to form a complete operation sequence. This constructs a complete legal supervision model that includes various operations before, during, and after execution, capable of achieving the supervision objective, ensuring the model's correctness and interpretability.

[0087] Predicate logic is an extension and development of propositional logic. It is built upon the further decomposition of simple propositions, breaking them down into two parts: entities and predicates. This addresses some simple yet common reasoning processes that propositional logic cannot handle. An entity refers to an objective entity that exists independently of subjective perception; it may be a concrete object or an abstract concept. A predicate is typically used to describe the properties of an object or the relationships between objects. The legal supervision model designed in this invention is constrained by supervisory business logic, functions, and domain knowledge. It selects atomic operations and assembles them by combining their various characteristics. First, a comprehensive analysis of the supervisory business logic, functions, and domain knowledge is conducted to clarify the required functions and objectives, guiding the selection of atomic operations. Then, based on the business logic and domain knowledge, atomic operations that meet the current supervisory objectives are selected. Finally, following a bottom-up hierarchical structure, the selected atomic operations are assembled according to logical relationships, ensuring the coherence of data flow and control flow at each level, resulting in the legal supervision model that achieves the goal of solving various supervisory objectives.

[0088] The legal supervision model of this invention has four levels from bottom to top: information source, preliminary information, intermediate judgment, and result judgment. The information source includes open source data, government regulatory data, and specific data sources that are not open source data or government regulatory data, which may include industry internal data, social media data, etc. The preliminary information refers to information such as names, times, and locations obtained by information extraction operations. For a specific supervision target, the basic information required is first obtained from the information source, then logical judgments of content comparison and threshold setting are performed, and finally it is judged whether each piece of information exceeds the threshold and reaches an abnormal value, thereby achieving the final judgment of the supervision target.

[0089] In implementation, the model first uses a dynamic permission-aware operator to determine whether the user has the necessary permissions. Then, it performs various information extraction operations from the information source to extract basic information. Subsequently, it performs content comparison and other processing on some information to generate secondary information such as frequency. Finally, based on this information, it determines whether any data exceeds a set threshold and is considered anomalous. Once an anomaly is detected, the path of the aforementioned atomic operations is recorded, forming a model that meets the supervision objectives.

[0090] To find an atomic operation input / output correlation analysis method that conforms to predicate logic, this invention also proposes an atomic operation input / output correlation analysis model, consisting of four modules: open source data processing, government regulatory data processing, early warning information generation and visualization, such as... Figure 7 As shown.

[0091] The four modules in the diagram are relatively independent yet interconnected. In the open-source data processing and government regulatory data processing sections, a redundant approach is used to design data reading functions capable of accessing multiple databases, ensuring initial data acquisition. The temporary categorization is primarily due to differences in the pre-run perception phase, particularly regarding permissions. In the open-source data section, the main perception operation is determining the security and reliability of the source data; permission perception may be less critical. However, in the government regulatory data section, due to potential confidentiality, permission perception operations may be more complex to prevent data leakage. In the legal supervision model, the atomic operation layer operations used in these two modules mainly consist of pre-run operations and in-run information extraction operations, ultimately providing the extracted information for further comparison or judgment by the supervision model. The early warning module primarily performs post-run operations and early warning push operations. Depending on the supervision model, the input supervision target, various thresholds, and operator permissions enable the filtering and push of abnormal information. The final visualization module primarily displays the current supervision model, recording all atomic operations from data reading to alert push, and sequentially forming the model's operation path. The legal supervision path derived from the current model can be directly used the next time the same supervision target is encountered, significantly improving the efficiency of building the supervision model. The differences in atomic operation paths represent the differences between supervision models, not only being key to distinguishing different supervision models but also greatly enhancing the practicality and flexibility of the supervision model.

[0092] Simultaneously, different atomic operations can be combined to form new atomic operation paths, thus constructing different legal supervision models. This process can be divided into two methods: proactive construction and reactive construction. Proactive construction refers to the operator combining all atomic operations performed throughout the entire process, from reading data before execution to sending early warnings after execution, to autonomously construct a legal supervision model related to the supervision target. The general construction process is as follows: Figure 8 As shown. This approach is generally applicable in two situations. One situation is when the operator can roughly determine which parts of the data might contain anomalies, typically indicating that the operator has already used a similar supervised model to process datasets similar to the current supervised target. The other situation is when preliminary screening of anomaly information has already been completed using passive modeling, and now it's necessary to focus on certain information. For example, if a specific individual is identified in the anomaly report, a supervised model related to that individual can be built to investigate whether there are more specific anomalies or their correlation with other anomalies.

[0093] When faced with complex data where the location of anomalies is unclear, a passive approach can be used to build a supervised model. By performing redundant analysis based on predicate logic, the various atomic operation paths required to reach the data where anomalies occur can be obtained.

[0094] This invention defines a quintuple (Im, Bj, Ml, Tk, Pn) in a passive supervised model f, representing the input and output of a class of supervised models. Here, Im represents the input of various types of information, such as information read from a database. Bj represents the current state of the supervised model, such as information source, preliminary information, intermediate judgment, or result judgment. Ml, Tk, and Pn are time-related information, which are parameters in the dynamic permission-aware service. Because various permissions may change over time, it is necessary to introduce time-related information to determine the current permission.

[0095] Furthermore, there is an atomic operation G: Gi = H(Ix). Gi represents various atomic operations, H represents the atomic operation function, and Ix, etc., represent various input variables, such as information.

[0096] Finally, in the predicate logic association process utilizing atomic operation input / output, the input information source D, time-series signals M, T, P, account password IP, and various thresholds, along with the legal supervision model set C derived from this method, represent the atomic operation path from the lowest level to achieving the current supervision goal. The atomic operations required for each step are represented by the set Bj, the input for each step is the set Im, and the rest are time signals. Each derived legal supervision path can be directly used when encountering the same supervision goal again, or alternative paths can be sought.

[0097] In summary, the atomic operation and assembly method for big data legal supervision models presented in this invention not only effectively solves some key problems existing in current models, but also significantly enhances model effectiveness, interpretability, and adaptability, bringing a new breakthrough to the application of big data in the field of legal supervision. This method is not only theoretically innovative but also has broad prospects for practical application.

Claims

1. A legal supervision system based on an atomic operation big data model, characterized in that, include: The atomic operation layer includes multiple independently executable atomic operation units, which are used to implement basic functions in the legal supervision process, including data reading, information extraction, anomaly detection, authorization verification, and error handling. The supervision model layer includes several legal supervision models for different supervision objectives. These legal supervision models are constructed based on the atomic operation assembly method of predicate logic by defining the preconditions and postconditions of atomic operations. The early warning push layer is used to send early warning information; The atomic operation units of the atomic operation layer include: The pre-operation unit acquires relevant data on legal events and senses the environment during the preparation phase before the legal supervision process begins. After the operation unit is run, in the analysis and feedback phase after the supervision process ends, it performs content comparison, threshold setting, and in-depth analysis of results. It conducts in-depth analysis of the detected abnormal results to determine whether they reflect potential illegal behavior or risks. During the operation phase of the monitoring process, the operating unit performs information extraction. Based on the real-time extracted information and the abnormal threshold set according to prior knowledge, it performs real-time detection of possible problems and pushes some simple judgments directly to the early warning push layer. The operations of the pre-run operation unit include: The data read operation connects to different databases based on the selected database type, reads the stored information, and stores it. Environment-aware operation: acquires information about the current terminal environment and performs corresponding atomic operations based on different environments; Permission-aware operation: Using static and dynamic permission-aware methods, it is determined whether the person currently performing the operation has the permission to perform this operation. The dynamic permission-aware method is based on the time factor. Error-aware operations determine whether an error has occurred during execution, thus affecting the normal operation of the supervised model. When an error is detected, the error location is recorded and saved. The operations of the operation unit after operation include: Content comparison involves comparing information related to business implementation extracted during the operation process, and extracting various types of information including similarities, differences, and frequency. Threshold setting: Set thresholds for the various types of information. When the statistically analyzed data exceeds the set thresholds, it is marked as abnormal data and an early warning is triggered. The legal supervision model is constructed using the following method: Define the preconditions and postconditions of atomic operations; the preconditions refer to the conditions that must be met before a specific atomic operation is executed, and the postconditions refer to the conditions that must be met after a specific atomic operation is executed. Predicate logic is used to establish logical relationships between atomic operations, and reachable atomic operation paths are sought based on different supervision objectives; By assembling atomic operation units that meet the preconditions, a complete legal supervision model process is constructed to ensure the correctness and interpretability of the model. The process of establishing logical relationships between atomic operations using predicate logic, seeking reachable atomic operation paths based on different supervision objectives, assembling atomic operation units that satisfy preconditions, and constructing a complete legal supervision model flow is implemented as follows: By using predicate logic to establish logical relationships between atomic operations, first, the input, output, and function of each atomic operation are clarified, and the sequential relationship between operations is defined using predicate logic; then, logical reasoning is used to verify whether the logical relationships between each operation are consistent, ensuring that no contradictions or errors occur during execution. Based on the specific monitoring objectives, analyze the required final results and intermediate steps, determine the required types of atomic operations, and find ways to achieve the monitoring objectives using both active and passive methods. By assembling atomic operation units that satisfy specific data inputs or environmental states, and arranging atomic operation units that conform to logical relationships and preconditions in sequence to form a complete operation sequence, a complete legal supervision model is constructed that includes various operations before, during, and after operation and can achieve the supervision objectives. Constrained by supervisory business logic, functions, and domain knowledge, atomic operations are selected and assembled by combining their various characteristics. First, a comprehensive analysis of the supervisory business logic, functions, and domain knowledge is conducted to clarify the required functions and objectives, guiding the selection of atomic operations. Then, based on the business logic and domain knowledge, atomic operations that meet the current supervisory objectives are selected. Finally, following a bottom-up hierarchical structure, the selected atomic operations are assembled according to logical relationships, ensuring the coherence of data flow and control flow at each level, resulting in the legal supervision model. The legal supervision model has four levels from bottom to top: information source, preliminary information, intermediate judgment, and result judgment. Information sources include open-source data, government regulatory data, and specific data sources that are not open-source or government regulatory data. Preliminary information refers to information obtained through information extraction operations. For a specific supervisory objective, the required basic information is first obtained from the information source, then logical judgments involving content comparison and threshold setting are performed, and finally, it is determined whether each piece of information exceeds the threshold and reaches an anomaly value, thereby achieving the final supervisory objective judgment. Find a path to achieve the supervision objective using one of the following methods: Method 1, Proactive Throughout the entire process from data reading to early warning push, a legal supervision model involving the supervision target is autonomously constructed by combining all the atomic operations involved. Method 2, Passive Redundant analysis based on predicate logic is used to obtain the various atomic operation paths required to reach the abnormal data.

2. The legal supervision system based on the atomic operation big data model according to claim 1, characterized in that, The early warning push layer utilizes visualization technology to provide a visual display and reporting function for the regulatory situation.

3. A method for atomic operation and assembly of a legal supervision model based on an atomic operation big data model in a legal supervision system as described in claim 1, characterized in that, Includes the following steps: First, determine whether you have the necessary permissions through dynamic permission awareness; Then, various information extraction operations are performed from the information source to extract basic information; Subsequently, some information is compared and processed to generate secondary information; Finally, based on the information obtained, it is determined whether any data exceeds the set threshold and becomes abnormal data. When an anomaly is found, the path of the atomic operation is recorded to form a model that meets the supervision objective.

Images