Health assessment methods, devices, equipment and media

Abstract

This invention discloses a health assessment method, apparatus, device, and medium, applicable to monitoring intelligent agents. The method includes: acquiring contract data of the assessment object and corresponding business data; performing natural language processing on the contract data to obtain text processing results, and determining a compliance score for the contract data based on the text processing results; performing statistical analysis on the business data to obtain statistical comparative analysis results, and determining anomaly scores for the business data based on the statistical comparative analysis results; and determining the health assessment result of the assessment object based on the compliance score and the anomaly score. This method uses natural language processing on contract data to identify key information, thereby accurately assessing contract compliance. Simultaneously, it performs multi-dimensional feature analysis on business data to identify potential operational anomalies. Combining compliance and anomaly scores, it determines a multi-dimensional health assessment result for the assessment object, improving the accuracy and objectivity of the health assessment results.

Description

Technical Field

[0001] This invention relates to the field of intelligent agent technology, and in particular to a health assessment method, apparatus, device, and medium. Background Technology

[0002] Traditional corporate compliance, legal, and internal audit models have significant limitations. Their focus is primarily on manual review by professionals, and is largely concentrated on in-process approvals (such as contract review) and post-event inspections (such as annual audits). Given the massive and frequent business activities of modern enterprises (e.g., hundreds or thousands of contracts signed daily, and continuously generated business and financial data), this manual model struggles to achieve comprehensive coverage and real-time monitoring.

[0003] Therefore, the identification of corporate risks is often delayed and one-sided, frequently only being discovered after problems have accumulated to a certain extent or caused actual losses. More importantly, the traditional model lacks a unified, quantifiable standard for continuously assessing the health of various business activities, making it difficult for management to form accurate and forward-looking insights into compliance risks and operational quality, and hindering the shift from passive response to proactive prevention. Therefore, there is an urgent need for a technological solution that can comprehensively and in real time assess the health of corporate operations. Summary of the Invention

[0004] This invention provides a health assessment method, apparatus, equipment, and medium to address the problems of lagging, narrow coverage, and lack of quantitative standards in enterprise health assessment in related technologies.

[0005] In a first aspect, the present invention provides a health assessment method, applied to a monitoring intelligent agent, comprising: Obtain the contract data of the assessment object and the corresponding business data; Natural language processing is performed on the contract data to obtain text processing results, and compliance scores are determined based on the text processing results. Perform statistical analysis on business data, obtain statistical comparison analysis results, and determine the anomaly score of business data based on the statistical comparison analysis results; Based on compliance scores and anomaly scores, the health assessment results of the assessed objects are determined.

[0006] In some embodiments, natural language processing is performed on contract data to obtain text processing results, including: Perform character recognition on the contract text data in the contract data to identify text keywords; Semantic analysis is performed on contract approval records in contract data to determine approval semantics. Text keywords and approval semantics are the results of text processing.

[0007] In some embodiments, a compliance score for contract data is determined based on the text processing results, including: The text keywords in the text processing results are matched with risk words in the preset risk word library to determine the risk matching results of the text keywords; The approval semantics of the text processing results are matched with the approval nodes in the preset approval process to determine the node matching results of the approval semantics. Compliance scores are determined based on risk matching results and node matching results.

[0008] In some embodiments, statistical analysis is performed on business data to obtain statistical comparative analysis results, including: Data features are extracted from business data to obtain candidate business feature vectors; Cluster analysis is performed on the candidate business feature vectors to determine the distribution clusters of each candidate business feature vector; Candidate business feature vectors whose outlier parameters in the distribution cluster are less than the preset outlier threshold are identified as the results of statistical comparative analysis.

[0009] In some embodiments, anomaly scores for business data are determined based on statistical comparative analysis results, including: A similarity analysis is performed on the target business feature vector and the preset standard feature vector in the statistical comparative analysis results to determine the feature similarity parameters; Based on preset internal control rules, security verification is performed on the target business feature vector to determine the results of business violations. Anomaly scores are determined based on feature similarity parameters and business violation results.

[0010] In some embodiments, the health assessment result of the assessed object is determined based on compliance score and anomaly score, including: Obtain the contract priority corresponding to the contract data, and the business scenario corresponding to the business data; Based on contract priority and business scenario, determine the first scoring weight for compliance scoring and the second scoring weight for anomaly scoring; Based on the first and second scoring weights, the compliance score and the abnormal score are weighted and fused to obtain the health assessment result of the assessed object.

[0011] In some embodiments, after determining the health assessment result of the assessed object based on compliance score and anomaly score, the method further includes: When the health assessment result is less than the preset health threshold, abnormal factors and abnormal early warning information are identified based on contract data and business data. Based on the abnormal factors, identify the business roles responsible for handling abnormal early warning information; Anomaly warning information is sent to the corresponding job-specific intelligent agent to trigger the job-specific intelligent agent to process the anomaly warning information.

[0012] Secondly, this disclosure provides a health assessment device, comprising: The data acquisition module is used to acquire the contract data of the evaluation object and the corresponding business data. The semantic processing module is used to perform natural language processing on contract data, obtain text processing results, and determine the compliance score of the contract data based on the text processing results. The data statistical analysis module is used to perform statistical analysis on business data, obtain statistical comparison analysis results, and determine the anomaly score of business data based on the statistical comparison analysis results. The health assessment module is used to determine the health assessment result of the assessed object based on compliance score and abnormality score.

[0013] Thirdly, the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-mentioned health assessment method.

[0014] Fourthly, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described health assessment method.

[0015] The aforementioned health assessment methods, devices, equipment, and media can be applied to monitoring intelligent agents, including: acquiring contract data of the assessment object and corresponding business data; performing natural language processing on the contract data to obtain text processing results, and determining the compliance score of the contract data based on the text processing results; performing statistical analysis on the business data to obtain statistical comparative analysis results, and determining the anomaly score of the business data based on the statistical comparative analysis results; and determining the health assessment result of the assessment object based on the compliance score and the anomaly score. This method uses natural language processing on contract data to identify its key information, thereby accurately assessing contract compliance. Simultaneously, it performs multi-dimensional feature analysis on business data to identify potential operational anomalies. Combining compliance and anomaly scores, it determines the multi-dimensional health assessment result of the assessment object, improving the accuracy and objectivity of the health assessment results. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart of a health assessment method according to an embodiment of the present invention; Figure 2 This is a flowchart of a health assessment method according to an embodiment of the present invention; Figure 3 This is a flowchart of a health assessment method according to an embodiment of the present invention; Figure 4 This is a flowchart of a health assessment method according to an embodiment of the present invention; Figure 5 This is a flowchart of a health assessment method according to an embodiment of the present invention; Figure 6 This is a flowchart of a health assessment method according to an embodiment of the present invention; Figure 7 This is a schematic diagram of a health assessment device according to an embodiment of the present invention; Figure 8 This is a schematic block diagram of a computer device according to an embodiment of the present invention. Detailed Implementation

[0018] As an example, such as Figure 1 As shown, a health assessment method is provided that can be applied to monitoring intelligent agents, including the following steps: S101, Obtain the contract data of the evaluation object and the corresponding business data; S102, Perform natural language processing on the contract data to obtain text processing results, and determine the compliance score of the contract data based on the text processing results; S103, Perform statistical analysis on business data, obtain statistical comparison analysis results, and determine the anomaly score of business data based on the statistical comparison analysis results; S104. Based on compliance scores and anomaly scores, determine the health assessment results of the assessed object.

[0019] As an example, in step S101, the evaluation object may be a project, supplier or customer, etc. Its contract data may include the contract content, contract approval records, contract signatory information and performance terms, etc. Business data may include business content, transaction flow and operational indicators and other multi-dimensional data. This disclosure does not limit the specific content included above.

[0020] In one embodiment, the monitoring agent can be an artificial intelligence (AI) device with data monitoring and analysis decision-making capabilities, capable of autonomously performing data collection, processing, and evaluation tasks. This disclosure does not limit the specific implementation form of the monitoring agent, which can be deployed on the cloud or local server and connect with the enterprise's internal database to obtain contract data and business data.

[0021] As an example, in step S102, the monitoring agent can perform natural language processing on the contract data, extract relevant information from the contract text data (such as key clauses, risk statements, qualifications of the signing entity, etc.), obtain the text processing result, and analyze the text processing result to determine whether there is any risk, thereby determining the compliance score of the contract data.

[0022] Natural language processing can include word recognition, syntactic analysis, semantic understanding and other processing methods, which are not limited in this disclosure. When analyzing the text processing results, compliance judgment can be made based on preset legal and regulatory databases, industry standards and internal corporate rules, so as to determine the compliance score, but it is not limited to this.

[0023] As an example, in step S103, the monitoring agent can perform statistical analysis on the business data to determine the changing trend and distribution characteristics of the business data, thereby obtaining the statistical comparison analysis results. Then, based on the statistical comparison analysis results, the degree of deviation between the business data and normal data is determined, and the abnormal score of the business data is determined.

[0024] The data statistical analysis may include descriptive statistics, trend analysis, distribution tests, and other processing methods, which are not limited in this disclosure. When conducting anomaly analysis on business data, the business data may be compared with historical business benchmarks and industry reference standards to quantify the degree of deviation of the business data, thereby determining the anomaly score, but is not limited to this.

[0025] As an example, in step S104, the monitoring agent dynamically weights and fuses the compliance score and the anomaly score to obtain the health assessment result. When fusing the compliance score and the anomaly score, the weighting coefficient can be dynamically adjusted according to the type of the assessment object, the risk preference of the business scenario, etc., to achieve scenario-based adaptation of the health assessment result. This disclosure does not limit the weight adjustment strategy.

[0026] In summary, this disclosure proposes a health assessment method applicable to monitoring agents, comprising: acquiring contract data of the assessment object and corresponding business data; performing natural language processing on the contract data to obtain text processing results, and determining a compliance score for the contract data based on the text processing results; performing statistical analysis on the business data to obtain statistical comparative analysis results, and determining anomaly scores for the business data based on the statistical comparative analysis results; and determining the health assessment result of the assessment object based on the compliance score and the anomaly score. This method accurately assesses contract compliance by using natural language processing on contract data to identify key information, while simultaneously performing multi-dimensional feature analysis on business data to identify potential operational anomalies. By combining compliance and anomaly scores, a multi-dimensional health assessment result for the assessment object is determined, improving the accuracy and objectivity of the health assessment results.

[0027] As an example, such as Figure 2 As shown, step S102, which involves performing natural language processing on the contract data to obtain the text processing result, includes: S201, Perform character recognition on the contract text data in the contract data to determine text keywords; S202, Semantic analysis is performed on the contract approval records in the contract data to determine the approval semantics. Text keywords and approval semantics are the results of text processing.

[0028] As an example, in step S201, optical character recognition (OCR) technology can be used to recognize characters in the contract text data, extract key information from the contract text, and thus determine text keywords. The keywords can be pre-defined related terms or words that appear in the contract text with a frequency exceeding a preset frequency threshold. It should be understood that the keyword extraction strategy can be dynamically adjusted in contracts of different types and for different customers to adapt to the compliance requirements and business characteristics of different industries. For example, in financial contracts, the focus is on clauses such as interest rates and liability for breach of contract, while in supply chain contracts, the focus is on elements such as delivery cycle and payment method.

[0029] For example, OCR technology is used to identify key fields in the contract text, such as Party A, Party B, amount, and performance period, and these key fields are used as text keywords.

[0030] As an example, in step S202, the contract approval record can be semantically parsed using natural language processing technology to identify key nodes, approval opinions, and approval time limits in the approval process, thereby determining the approval semantics. In other words, the approval semantics indicate the execution status of each stage in the contract approval process.

[0031] The approval semantics can include information such as whether the approval has been approved and the completeness of the approval chain, in order to assess the compliance level of the contract in the internal management process.

[0032] For example, natural language processing technology is used to parse the approval opinion text in the approval record, identify key decision information such as "agree", "reject", and "supplementary materials required", as well as process attributes such as "approver's name" and "approval time", and use them as the output of semantic parsing.

[0033] As an example, such as Figure 3 As shown, step S102, which is to determine the compliance score of the contract data based on the text processing results, includes: S301, Match the text keywords of the text processing result with the risk words in the preset risk word library to determine the risk matching result of the text keywords; S302, match the approval semantics of the text processing result with the approval nodes in the preset approval process, and determine the node matching result of the approval semantics; S303 determines the compliance score based on risk matching results and node matching results.

[0034] As an example, in step S301, the monitoring agent can compare the text keywords with the risk words in the preset risk word library, determine whether there are words in the text keywords that are the same as or similar to the risk words, and mark the matching words as risk items to form a risk matching result.

[0035] Optionally, the risk matching results may also include the specific type of the matched words, the semantic similarity with the risk words, the risk level of the corresponding risk words, and the contextual position information of the matched item in the contract, so as to provide fine-grained support for subsequent risk assessment.

[0036] As an example, in step S302, the monitoring agent can align key nodes in the approval semantics with standardized nodes in the preset approval process to identify risk situations such as unauthorized approval or missing nodes, thereby determining the node matching result.

[0037] For example, natural language processing models can be used to identify approval nodes in the approval semantics, such as the submitter, approver, and approval time, and automatically compare them with the company's preset standard approval process to determine whether there are missing nodes or disordered order.

[0038] Optionally, the approval semantics can also include the approval opinions of each approval node. The monitoring agent can perform semantic analysis on each approval opinion, analyze the semantic sentiment and key expressions in the approval opinion, and match the analysis results with the preset risk semantics to identify whether there are abnormal semantics such as ambiguous expressions or illegal commitments, such as conditional expressions such as "suspend execution" or "supplementary materials required", and then generate structured data (i.e. the above node matching results) for subsequent compliance scoring calculation.

[0039] As an example, in step S303, the monitoring agent, based on the risk matching results and node matching results, and in conjunction with preset scoring rules, quantitatively assesses the compliance of the contract data and generates a compliance score. The scoring rules can be weighted according to factors such as the number, type, and risk level of risk items in the risk matching results, and the degree of node deviation in the node matching results, ensuring that the score accurately reflects the compliance level of the contract in both text content and approval process.

[0040] In one embodiment, such as Figure 4 As shown, step S103 involves performing statistical analysis on the business data to obtain statistical comparison analysis results, including: S401, extract features from business data to obtain candidate business feature vectors; S402, perform cluster analysis on the candidate business feature vectors to determine the distribution clusters of each candidate business feature vector; S403, the candidate business feature vectors in the distribution cluster whose outlier parameters are less than the preset outlier threshold are determined as the statistical comparison analysis results.

[0041] As an example, in step S401, the monitoring agent can use a deep learning model to extract candidate business feature vectors that can characterize the business behavior of an enterprise based on multi-dimensional information such as transaction amount, transaction frequency, and business content in the business data.

[0042] The candidate operational characteristic vector may include, but is not limited to, the temporal characteristics of transaction patterns, the topological characteristics of related party networks, and the distribution characteristics of capital flows, in order to comprehensively characterize the operational behavior of the evaluation object.

[0043] As an example, in step S402, the monitoring agent can perform cluster analysis (e.g., K-means or DBSCAN clustering algorithm) on the extracted candidate business feature vectors to determine the distribution clusters of each candidate business feature vector, so as to group and classify different types of candidate business feature vectors, identify the set of candidate business feature vectors with the same or similar data types, and provide a data foundation for subsequent anomaly detection.

[0044] As an example, in step S403, the monitoring agent calculates the outlier parameter of each candidate business feature vector based on the clustering results. The outlier parameter is used to measure the distance or density difference between each candidate business feature vector and the center of its distribution cluster.

[0045] When the outlier parameter is less than the preset outlier threshold, it indicates that the candidate business feature vector deviates little from its distribution cluster, meaning it belongs to normal data and can be retained as a target business feature vector for subsequent anomaly scoring calculations. When the outlier parameter is greater than or equal to the preset outlier threshold, it indicates that the candidate business feature vector deviates significantly and is identified as an anomaly data point, which will be removed to avoid interfering with subsequent analysis. This screening mechanism ensures that the target business feature vectors entering the anomaly scoring calculation stage truly reflect the normal business data of the evaluation object, thereby improving the accuracy and reliability of anomaly scoring.

[0046] In one embodiment, such as Figure 5 As shown, step S103, which is to determine the anomaly score of business data based on the statistical comparison analysis results, includes: S501, Perform similarity analysis on the target business feature vector and the preset standard feature vector in the statistical comparison analysis results, and determine the feature similarity parameters; S502, based on preset internal control rules, performs security verification on the target business feature vector to determine the business violation results; S503 determines the anomaly score based on feature similarity parameters and business violation results.

[0047] As an example, in step S501, the feature similarity parameter between the target business feature vector and the preset standard feature vector can be calculated by using cosine similarity or Euclidean distance algorithm to quantify the degree of deviation between the business data and the preset security standard. When the feature similarity parameter is lower than or equal to the preset threshold, it indicates that there is a significant difference between the business data and the standard pattern, which may pose a potential risk. When the feature similarity parameter is higher than the preset threshold, it indicates that the business data is highly consistent with the standard pattern and there is no significant risk.

[0048] As an example, in step S502, the monitoring agent performs multi-dimensional security verification on the target business feature vector based on the enterprise's internal control rule base, identifies whether there are any operational behaviors or data anomalies that violate compliance requirements, and outputs the business violation results.

[0049] For example, if the internal control rule includes "contract approval requires the signatures of both legal and financial departments", then if any approval node record is missing in the target business feature vector, it is judged as a violation, and a corresponding violation mark is generated as part of the business violation result.

[0050] As an example, in step S503, the system performs a weighted calculation of the feature similarity parameter and the business violation result to comprehensively evaluate the degree of abnormality of the business data, thereby determining the abnormality score.

[0051] In the weighted calculation, the weight of the feature similarity parameter can be dynamically adjusted according to the risk sensitivity of the business scenario. Higher weights are given to high-risk scenarios to strengthen the deviation warning capability. The weight of the business violation result can be set in stages according to the severity of the violation. Higher weights are given to major violations to highlight the impact of compliance risks.

[0052] For example, based on preset weighting coefficients, the deviation score corresponding to the feature similarity parameter and the violation score corresponding to the business violation result are weighted and summed to obtain the final anomaly score. The deviation score can be mapped based on the difference between the feature similarity parameter and the preset threshold, and the violation score is obtained by looking up a table based on the violation type and severity level, but it is not limited to this, and this disclosure does not limit it.

[0053] In one embodiment, such as Figure 6 As shown, step S104, which is to determine the health assessment result of the assessment object based on the compliance score and the anomaly score, includes: S601, obtain the contract priority corresponding to the contract data, and the business scenario corresponding to the business data; S602, based on contract priority and business scenario, determines the first scoring weight for compliance scoring and the second scoring weight for anomaly scoring; S603, based on the first and second scoring weights, performs a weighted fusion of compliance scores and abnormal scores to obtain the health assessment result of the assessed object.

[0054] As an example, in step S601, the contract priority is determined by classifying the contract amount, the qualifications of the partner, and the scope of business impact. For example, contracts can be divided into three priorities: high, medium, and low. Contracts with high value or involving important partners are marked as high priority. Business scenarios are classified according to the department (such as procurement, sales, and R&D), process type, and the frequency of historical risk occurrence.

[0055] As an example, in step S602, differentiated first and second score weights can be configured for compliance scores and anomaly scores according to different contract priorities and business scenarios. For example, high-priority contracts are given higher weights for compliance scores in procurement scenarios to strengthen process compliance requirements; while in high-risk business scenarios such as R&D data management, the weight of anomaly scores is increased to more sensitively capture potential risks and ensure the accuracy of risk identification and scenario adaptability.

[0056] As an example, in step S603, the compliance score and the abnormal score can be weighted and fused according to the first scoring weight and the second scoring weight, and the calculation result can be determined as the health assessment result of the assessment object to reflect the overall health level of the enterprise's operation.

[0057] Optionally, the health assessment results may also include specific risk warnings and improvement suggestions. For example, for anomalies found in the assessment, the system automatically generates structured risk warnings, including risk type, related terms, scope of impact, and suggested corrective measures, and pushes them to the relevant responsible parties.

[0058] In one embodiment, after step S104, i.e., determining the health assessment result of the assessment object based on the compliance score and the anomaly score, the following steps may also be included: S701, When the health assessment result is less than the preset health threshold, determine the abnormal factors and abnormal early warning information based on contract data and business data; S702, Based on abnormal factors, determine the business positions that handle abnormal early warning information; S703 sends the abnormal warning information to the job-specific intelligent agent corresponding to the business position, so as to trigger the job-specific intelligent agent to process the abnormal warning information.

[0059] As an example, in step S701, when the health assessment result is lower than the preset health threshold, the monitoring agent can trace and identify abnormal factors that cause the health assessment result to decline based on contract data and business data, such as missing contract approval process, key terms deviating from compliance requirements, or abnormal transaction patterns in business data, and generate abnormal warning information including risk type, occurrence stage, related responsible party, and severity.

[0060] As an example, in step S702, the monitoring agent can intelligently match the abnormal factors in the abnormal warning information with the business positions according to the preset job responsibility mapping rules, and determine the business position responsible for handling the abnormality. For example, when the abnormal factor involves an amount exceeding the limit, the monitoring agent can determine that the relevant approval position in the finance department is responsible for handling it.

[0061] As an example, in step S703, the monitoring agent can push abnormal warning information to the corresponding job agent through the enterprise's internal collaboration platform, triggering it to automatically start the abnormal handling process, including sending reminders to the responsible person, generating a processing task work order, retrieving relevant business vouchers for auxiliary analysis, and tracking the processing progress until the abnormality is resolved.

[0062] Optionally, after receiving an anomaly warning, the job-specific intelligent agent can automatically decompose the task according to the preset handling strategy template, assign it to the specific person in charge, and set the processing time limit; at the same time, the intelligent agent will continuously monitor the processing feedback data and update it in real time.

[0063] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0064] In one embodiment, a health assessment device is provided, which corresponds one-to-one with the health assessment methods described in the above embodiments. For example... Figure 7 As shown, the health assessment device includes a data acquisition module 701, a semantic processing module 702, a data statistical analysis module 703, and a health assessment module 704. Detailed descriptions of each functional module are as follows: The data acquisition module 701 is used to acquire the contract data of the evaluation object and the business data corresponding to the contract data. The semantic processing module 702 is used to perform natural language processing on contract data, obtain text processing results, and determine the compliance score of the contract data based on the text processing results. The data statistical analysis module 703 is used to perform data statistical analysis on business data, obtain statistical comparison analysis results, and determine the anomaly score of business data based on the statistical comparison analysis results. Health assessment module 704 is used to determine the health assessment result of the assessed object based on compliance score and abnormality score.

[0065] In one embodiment, the semantic processing module 702 is further configured to perform character recognition on the contract text data in the contract data to determine text keywords; Semantic analysis is performed on contract approval records in contract data to determine approval semantics. Text keywords and approval semantics are the results of text processing.

[0066] In one embodiment, the semantic processing module 702 is further configured to match the text keywords of the text processing result with risk words in a preset risk word library to determine the risk matching result of the text keywords; The approval semantics of the text processing results are matched with the approval nodes in the preset approval process to determine the node matching results of the approval semantics. Compliance scores are determined based on risk matching results and node matching results.

[0067] In one embodiment, the data statistical analysis module 703 is further configured to extract features from business data to obtain candidate business feature vectors; Cluster analysis is performed on the candidate business feature vectors to determine the distribution clusters of each candidate business feature vector; Candidate business feature vectors whose outlier parameters in the distribution cluster are less than the preset outlier threshold are identified as the results of statistical comparative analysis.

[0068] In one embodiment, the data statistical analysis module 703 is further configured to perform similarity analysis on the target business feature vector and the preset standard feature vector in the statistical comparison analysis results, and determine the feature similarity parameter; Based on preset internal control rules, security verification is performed on the target business feature vector to determine the results of business violations. Anomaly scores are determined based on feature similarity parameters and business violation results.

[0069] In one embodiment, the health assessment module 704 is further configured to obtain the contract priority corresponding to the contract data and the business scenario corresponding to the business data. Based on contract priority and business scenario, determine the first scoring weight for compliance scoring and the second scoring weight for anomaly scoring; Based on the first and second scoring weights, the compliance score and the abnormal score are weighted and fused to obtain the health assessment result of the assessed object.

[0070] In one embodiment, the health assessment module 704 is further configured to, when the health assessment result is less than a preset health threshold, determine abnormal factors and abnormal warning information based on contract data and business data; Based on the abnormal factors, identify the business roles responsible for handling abnormal early warning information; Anomaly warning information is sent to the corresponding job-specific intelligent agent to trigger the job-specific intelligent agent to process the anomaly warning information.

[0071] This invention provides a health assessment device, comprising: a data acquisition module for acquiring contract data of the assessment object and corresponding business data; a semantic processing module for performing natural language processing on the contract data to obtain text processing results, and determining a compliance score for the contract data based on the text processing results; a data statistical analysis module for performing statistical analysis on the business data to obtain statistical comparison analysis results, and determining anomaly scores for the business data based on the statistical comparison analysis results; and a health assessment module for determining the health assessment result of the assessment object based on the compliance score and the anomaly score. This device accurately assesses contract compliance by performing natural language processing on contract data to identify key information, while simultaneously performing multi-dimensional feature analysis on business data to identify potential operational anomalies. By combining compliance and anomaly scores, it determines a multi-dimensional health assessment result for the assessment object, improving the accuracy and objectivity of the health assessment results.

[0072] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 8 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database is used for data employed in the health assessment method. The network interface is used for communication with external terminals via a network connection. When executed by the processor, the computer program implements a health assessment method.

[0073] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements a health assessment method.

[0074] In one embodiment, a computer-readable storage medium is provided that stores a computer program, which, when executed by a processor, implements a health assessment method.

[0075] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), IAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0076] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0077] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A health assessment method, characterized in that, Applications to monitoring intelligent agents include: Obtain the contract data of the evaluation object and the corresponding business data; Natural language processing is performed on the contract data to obtain text processing results, and a compliance score for the contract data is determined based on the text processing results. Perform statistical analysis on the business data to obtain statistical comparison analysis results, and determine the anomaly score of the business data based on the statistical comparison analysis results; Based on the compliance score and the anomaly score, the health assessment result of the assessed object is determined.

2. The method according to claim 1, characterized in that, The natural language processing of the contract data to obtain the text processing result includes: Character recognition is performed on the contract text data in the contract data to determine text keywords; Semantic parsing is performed on the contract approval records in the contract data to determine the approval semantics. The text keywords and the approval semantics belong to the text processing result.

3. The method according to claim 1, characterized in that, The process of determining the compliance score of the contract data based on the text processing results includes: The text keywords of the text processing result are matched with risk words in a preset risk word library to determine the risk matching result of the text keywords; The approval semantics of the text processing result are matched with the approval nodes in the preset approval process to determine the node matching result of the approval semantics; The compliance score is determined based on the risk matching results and the node matching results.

4. The method according to claim 1, characterized in that, The statistical analysis of the business data to obtain statistical comparison analysis results includes: Feature extraction is performed on the business data to obtain candidate business feature vectors; Cluster analysis is performed on the candidate business feature vectors to determine the distribution clusters of each candidate business feature vector; Candidate business feature vectors in the distribution cluster whose outlier parameters are less than a preset outlier threshold are identified as the statistical comparison analysis results.

5. The method according to claim 1, characterized in that, The process of determining the anomaly score of the business data based on the statistical comparison analysis results includes: A similarity analysis is performed on the target business feature vector and the preset standard feature vector in the statistical comparison analysis results to determine the feature similarity parameters; Based on preset internal control rules, the target business feature vector is subjected to security verification to determine the business violation result; The anomaly score is determined based on the feature similarity parameter and the business violation result.

6. The method according to claim 1, characterized in that, The process of determining the health assessment result of the assessment object based on the compliance score and the anomaly score includes: Obtain the contract priority corresponding to the contract data, and the business scenario corresponding to the business data; Based on the contract priority and the business scenario, a first scoring weight for the compliance score and a second scoring weight for the anomaly score are determined. Based on the first scoring weight and the second scoring weight, the compliance score and the abnormal score are weighted and fused to obtain the health assessment result of the assessment object.

7. The method according to claim 1, characterized in that, After determining the health assessment result of the assessed object based on the compliance score and the anomaly score, the method further includes: When the health assessment result is less than the preset health threshold, abnormal factors and abnormal early warning information are determined based on the contract data and the business data. Based on the aforementioned abnormal factors, the business roles responsible for handling abnormal early warning information are determined. The abnormal warning information is sent to the job-specific intelligent agent corresponding to the job position, so as to trigger the job-specific intelligent agent to process the abnormal warning information.

8. A health assessment device, characterized in that, include: The data acquisition module is used to acquire the contract data of the evaluation object and the business data corresponding to the contract data; The semantic processing module is used to perform natural language processing on the contract data to obtain text processing results, and to determine the compliance score of the contract data based on the text processing results. The data statistical analysis module is used to perform data statistical analysis on the business data, obtain statistical comparison analysis results, and determine the anomaly score of the business data based on the statistical comparison analysis results. The health assessment module is used to determine the health assessment result of the assessment object based on the compliance score and the anomaly score.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the health assessment method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the health assessment method as described in any one of claims 1 to 7.

Images