Business handling risk prediction method and device and electronic equipment

Abstract

The invention discloses a business handling risk prediction method and apparatus, and an electronic device. Relates to the field of big data, and comprises the following steps: obtaining business handling information of a target object in a target time period, and processing the business handling information to obtain M pieces of business feature information, M being a positive integer; the M pieces of business feature information are input into a risk prediction model, the risk coefficient of the target object under each risk feature is obtained, and N is a positive integer; and determining whether the target object has a business handling risk according to the N risk coefficients, and determining a risk level of the target object according to the N risk coefficients when the business handling risk exists. Through the method and the device, the problem of relatively low accuracy and efficiency of determining the business handling risk in related technologies is solved.

Description

Technical Field

[0001] This application relates to the field of big data, and more specifically, to a method, apparatus, and electronic device for predicting business processing risks. Background Technology

[0002] When engaging in business cooperation with enterprises, it is usually necessary to determine the enterprise's compliance risks and, based on these risks, as to determine whether the enterprise is capable of conducting business.

[0003] In current corporate legal compliance management systems, risk prediction and abnormal behavior identification primarily rely on manually formulated rules and expert experience. While this traditional approach has some effectiveness in specific scenarios, its limitations are becoming increasingly apparent when facing an increasingly complex and diverse compliance environment. First, the process of manually formulating rules is time-consuming and labor-intensive, and may not reflect emerging business models and operational behaviors in a timely manner, leading to rule lag. Second, expert experience is highly subjective; different experts may assess risks significantly differently in different scenarios, lacking uniformity and objectivity. Finally, traditional methods are difficult to flexibly transfer and adjust across different business scenarios; each scenario requires independent manual analysis and rule setting, which is not only inefficient but may also lead to redundant resource investment.

[0004] There is currently no effective solution to the problem of low accuracy and efficiency in identifying business processing risks in related technologies. Summary of the Invention

[0005] This application provides a method, apparatus, and electronic device for predicting business processing risks, in order to solve the problem of low accuracy and efficiency in determining business processing risks in related technologies.

[0006] According to one aspect of this application, a method for predicting business processing risks is provided. The method includes: acquiring business processing information of a target object within a target time period, processing the business processing information to obtain M business characteristic information, where M is a positive integer; inputting the M business characteristic information into a risk prediction model to obtain a risk coefficient for the target object under each risk characteristic, where N is a positive integer; determining whether the target object has business processing risks based on the N risk coefficients, and if business processing risks exist, determining the risk level of the target object based on the N risk coefficients.

[0007] Optionally, processing the business processing information to obtain M business feature information includes: retrieving M preset business features from the database and determining the calculation method for each preset business feature to obtain M calculation methods; obtaining the calculation parameters required for each calculation method to obtain M sets of calculation parameters; determining whether the business processing information contains the M sets of calculation parameters; if the business processing information does not contain any parameter from any set of calculation parameters, filling the business processing information to obtain updated business processing information; if the business processing information contains the M sets of calculation parameters, calculating the M business feature information using the business processing information and the M calculation methods.

[0008] Optionally, the risk prediction model is trained as follows: obtain historical business processing information of multiple preset objects, and a risk coefficient set for each preset object, wherein each risk coefficient set contains multiple risk coefficients associated with the preset object; take each historical business processing information and the corresponding risk coefficient set as a set of first sample data to obtain multiple sets of first sample data, and use the multiple sets of first sample data to train the isolated forest algorithm model to obtain the risk prediction model.

[0009] Optionally, determining whether a target object has business processing risks based on N risk coefficients includes: determining the importance score of the risk feature to which each risk coefficient belongs, resulting in N importance scores; performing a weighted summation of the N risk coefficients and the N importance scores to obtain a risk score; determining whether the risk score is greater than a first threshold; if the risk score is greater than the first threshold, determining that the target object has business processing risks; if the risk score is less than or equal to the first threshold, determining that the target object does not have business processing risks.

[0010] Optionally, determining the risk level of the target object based on N risk coefficients includes: sequentially obtaining the second threshold of the risk coefficient of each risk feature, and sequentially comparing the risk coefficient of each risk feature with the corresponding second threshold to obtain N comparison results; selecting abnormal results from the N comparison results, and determining the risk level of the risk feature to which the abnormal result belongs as the risk level of the target object, wherein the abnormal result indicates that the risk coefficient is greater than the corresponding second threshold.

[0011] Optionally, after determining the risk level of the target object based on N risk coefficients, the method further includes: determining the risk characteristics to which the abnormal risk coefficients among the N risk coefficients belong, wherein the abnormal risk coefficients are risk coefficients greater than the corresponding second threshold; inputting the risk characteristics and business processing information into a decision tree model to obtain the reasons for the existence of risk characteristics, wherein the decision tree model is trained through multiple sets of second sample data, and each set of second sample data includes historical business processing information and historical risk characteristics of a historical object, as well as the reasons for the existence of historical risk characteristics.

[0012] Optionally, after determining the risk level of the target object based on N risk coefficients, the method further includes: constructing a reinforcement learning model, wherein the risk features are used as state space objects in the model, the adjustment action of the second threshold is used as action space objects in the model, and a reward function is set in the model; iteratively training the reinforcement learning model using multiple sets of third sample data, and adjusting the second threshold of each risk feature through the reinforcement learning model, wherein each set of third sample data includes historical business processing information and historical risk features of a historical object, as well as risk coefficients.

[0013] According to another aspect of this application, a business processing risk prediction device is provided. The device includes: a first acquisition unit, configured to acquire business processing information of a target object within a target time period, and process the business processing information to obtain M business characteristic information, where M is a positive integer; a first input unit, configured to input the M business characteristic information into a risk prediction model to obtain a risk coefficient of the target object under each risk characteristic, where N is a positive integer; and a first determination unit, configured to determine whether the target object has business processing risk based on the N risk coefficients, and, if business processing risk exists, determine the risk level of the target object based on the N risk coefficients.

[0014] According to another aspect of the present invention, a computer program product is also provided, comprising a computer program that, when executed by a processor, implements a business processing risk prediction method provided in the foregoing embodiments of the present application.

[0015] According to another aspect of the present invention, an electronic device is also provided, comprising one or more processors and a memory; the memory stores computer-readable instructions, and the processor is configured to execute the computer-readable instructions, wherein the computer-readable instructions, when executed, perform a business processing risk prediction method provided in the foregoing embodiments.

[0016] This application employs the following steps: obtaining business processing information of a target object within a target time period, processing the business processing information to obtain M business feature information, where M is a positive integer; inputting the M business feature information into a risk prediction model to obtain the risk coefficient of the target object under each risk feature, where N is a positive integer; determining whether the target object has business processing risk based on the N risk coefficients, and if business processing risk exists, determining the risk level of the target object based on the N risk coefficients. This solves the problem of low accuracy and efficiency in determining business processing risk in related technologies. By processing business processing information and using business feature information and a risk prediction model to determine the risk coefficients of the target object, and then determining the risk level through multiple risk coefficients, the technical effect of improving the accuracy of business processing risk identification is achieved. Attached Figure Description

[0017] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0018] Figure 1 This is a flowchart of a business processing risk prediction method provided according to an embodiment of this application;

[0019] Figure 2 This is a schematic diagram of an optional model prediction result provided according to an embodiment of this application;

[0020] Figure 3 This is a schematic diagram of a business processing risk prediction device provided according to an embodiment of this application;

[0021] Figure 4 This is a schematic diagram of an electronic device provided according to an embodiment of this application. Detailed Implementation

[0022] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0023] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present application.

[0024] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of this application described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0025] It should be noted that the methods, devices, and electronic equipment for predicting business processing risks as defined in this disclosure can be used in the field of big data, or in any field other than big data. The application fields of the methods, devices, and electronic equipment for predicting business processing risks as defined in this disclosure are not limited.

[0026] It should be noted that all information, user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, and displayed data) used in this application are information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of related data all comply with the relevant laws, regulations, and standards of the relevant regions, have taken necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entry points for users to choose to authorize or refuse use. If the user chooses to refuse, the process proceeds to the expert decision-making process. For example, this system has interfaces with relevant users or institutions. Before obtaining relevant information, a request to obtain the information needs to be sent to the aforementioned user or institution through the interface, and the relevant information is obtained only after receiving consent from the aforementioned user or institution.

[0027] The embodiments or examples disclosed herein are not exhaustive, but merely illustrative of some embodiments or examples, and are not intended to limit the scope of protection of this disclosure. Unless otherwise specified, each step in a particular embodiment or example can be implemented as an independent embodiment, and the steps can be arbitrarily combined. For example, a solution after removing some steps in a particular embodiment or example can also be implemented as an independent embodiment, and the order of the steps in a particular embodiment or example can be arbitrarily interchanged. Furthermore, optional methods or examples in a particular embodiment or example can be arbitrarily combined; moreover, embodiments or examples can be arbitrarily combined. For example, some or all steps of different embodiments or examples can be arbitrarily combined, and a particular embodiment or example can be arbitrarily combined with optional methods or examples of other embodiments or examples.

[0028] For ease of description, the following explains some of the nouns or terms used in the embodiments of this application:

[0029] Isolation Forest is an unsupervised learning algorithm for anomaly detection. It uses a tree structure to partition the data between randomly selected features and randomly selected feature values, thereby isolating outlier data points.

[0030] Decision tree model: It is a supervised learning algorithm that can divide data into different categories or results through a series of questions, similar to a multi-stage decision-making process with a tree structure.

[0031] According to an embodiment of this application, a method for predicting business processing risks is provided.

[0032] Figure 1 This is a flowchart of a business processing risk prediction method provided according to an embodiment of this application. For example... Figure 1 As shown, the method includes the following steps:

[0033] Step S101: Obtain the business processing information of the target object within the target time period, and process the business processing information to obtain M business feature information, where M is a positive integer.

[0034] It should be noted that the target object refers to the enterprise or entity that needs to be predicted for legal compliance risks, the target period refers to a specific time window used to collect and analyze data, business processing information, that is, detailed information on all business activities carried out by the target object within the target period, including but not limited to procurement bidding, contract signing, financial transactions, etc., and business characteristic information, that is, the characteristics extracted through the preprocessing and analysis of business processing information, which may include, for example, procurement amount, bidding method, etc.

[0035] Specifically, the execution entity in this embodiment can be a risk prediction system. When performing risk prediction, the system first needs to collect all business activity records of the target object within a specified time period. Information can be retrieved from multiple data sources, such as internal databases, data released by external regulatory agencies, and compliance events mentioned in news reports. Then, the system preprocesses this data, including data cleaning (removing erroneous or irrelevant data), data integration (merging datasets from different sources), and data normalization (ensuring consistent data formats). Finally, M business feature information is extracted from the preprocessed business transaction information and used as input for the subsequent risk prediction model.

[0036] For example, Organization A conducted multiple procurement activities within a certain period. The system will collect all data related to these procurements, including but not limited to the qualifications of bidders, the detailed content of the tender notices, the date and reason for issuing the winning bid notices, the final terms of the contracts, and the finance department's audit records for these transactions. The system will then extract features from this information, such as amount, date, and tendering method, which are relevant to risk prediction.

[0037] Step S102: Input M business feature information into the risk prediction model to obtain the risk coefficient of the target object under each risk feature, where N is a positive integer.

[0038] It should be noted that the risk prediction model refers to a model built using the Isolation Forest algorithm, which can analyze input business characteristic information and identify potential abnormal patterns or risks. The risk coefficient is a quantitative indicator output by the model, representing the degree of abnormality of a specific risk characteristic; a high risk coefficient means that there may be a significant risk.

[0039] Specifically, after obtaining M business characteristic information, these M characteristics can be input into the risk prediction model. The risk prediction model will analyze each characteristic, using the Isolation Forest algorithm to calculate its path length within the Isolation Forest, thereby assessing its degree of anomaly. In this way, the model will generate N risk coefficients, corresponding to N different risk characteristics, and then determine the business processing risk based on these risk coefficients.

[0040] For example, when the system inputs multiple feature information into the model, the Isolation Forest algorithm constructs multiple isolated trees based on historical compliance data and current feature values, and evaluates the degree of anomaly of this feature through the structure of the trees. If the model finds that the average path length of the "bidding method violation risk feature coefficient" is much smaller than the average path length of normal samples, it will be assigned a high risk coefficient, indicating that there is a business processing risk under this risk feature of the target object.

[0041] Step S103: Determine whether the target object has business processing risks based on N risk coefficients, and if business processing risks exist, determine the risk level of the target object based on N risk coefficients.

[0042] It should be noted that business processing risk refers to the risks that may arise when performing business operations. The risk level is a quantitative representation of the degree of business processing risk, which is usually divided into low, medium, high or other more detailed levels, and is used to guide subsequent compliance actions or resource allocation.

[0043] Specifically, given N risk coefficients, the system comprehensively assesses whether the target entity's business faces risk based on these N risk coefficients derived from the model. At this point, it distinguishes between normal business and risky business using preset thresholds. If any risk coefficient exceeds the preset threshold, the system determines that the target entity faces business processing risk. Furthermore, the system determines the risk level based on the overall situation of all risk coefficients, for example, through weighted averages or other statistical methods, thereby ensuring the accuracy of the risk level determination.

[0044] For example, after performing risk prediction on the target entity, the results show that the risk coefficient for "Contract Signing Violation Risk Characteristic Coefficient" is 0.9 (out of 1), indicating a high potential risk in the contract signing process. The risk coefficient for "Financial Record Consistency Risk Characteristic Coefficient" is 0.2, indicating a low compliance risk in financial records. The system will comprehensively consider all risk coefficients. If the overall risk exceeds a pre-set threshold (e.g., 0.5), it will determine that there is a high risk of contract signing violation, requiring further review and measures to ensure the security of business cooperation with the target entity.

[0045] The business processing risk prediction method provided in this application involves acquiring business processing information of a target object within a target time period, processing the information to obtain M business feature information, where M is a positive integer; inputting the M business feature information into a risk prediction model to obtain the risk coefficient of the target object under each risk feature, where N is a positive integer; determining whether the target object has business processing risk based on the N risk coefficients, and if business processing risk exists, determining the risk level of the target object based on the N risk coefficients. This solves the problem of low accuracy and efficiency in determining business processing risk in related technologies. By processing business processing information and using business feature information and a risk prediction model to determine the risk coefficients of the target object, and then determining the risk level through multiple risk coefficients, the technical effect of improving the accuracy of business processing risk identification is achieved.

[0046] To ensure the accuracy of business feature information, optionally, in the business processing risk prediction method provided in this application embodiment, processing the business processing information to obtain M business feature information includes: obtaining M preset business features from the database and determining the calculation method for each preset business feature to obtain M calculation methods; obtaining the calculation parameters required for each calculation method to obtain M sets of calculation parameters; determining whether the business processing information contains the M sets of calculation parameters; if the business processing information does not contain any parameter from any set of calculation parameters, filling the business processing information to obtain updated business processing information; if the business processing information contains the M sets of calculation parameters, calculating the M business feature information using the business processing information and the M calculation methods.

[0047] It should be noted that the preset business features can be features closely related to business compliance, that is, feature names included in the M business feature information. The business feature information includes feature name and feature value. For example, the preset business features can be: number of violations, violation frequency, etc. When obtaining business feature information, it is first necessary to determine the preset business features required for risk prediction and determine the calculation method for each preset business feature. Through the above calculation method, each preset business feature can be calculated so that the model can perform risk prediction operation based on the value of the preset business feature. The calculation method can be a formula.

[0048] After determining the calculation method for each preset business feature, it is necessary to determine the parameters required for each calculation method, obtain M sets of calculation parameters, and obtain each calculation parameter from the business processing information of the target object, such as the number of violation cases, the number of bidding, etc. Then, the preset business features can be calculated according to each set of calculation parameters to obtain the feature data of each preset business feature, thereby combining the preset business features and feature data into business feature information.

[0049] It should be noted that when the calculation parameters for a certain calculation method cannot be obtained from the business processing information, the missing calculation parameters can be filled in using a preset data filling method to ensure data integrity. The filling method can be based on historical averages, predicted values, or other reasonable estimates, thereby ensuring the accurate calculation of business feature information.

[0050] This embodiment determines whether the parameters in the business processing information are complete by establishing calculation formulas for each business characteristic information. Then, it calculates each business characteristic information according to the calculation parameters and formulas, thereby accurately and completely extracting business characteristic information related to legal compliance risk prediction from the messy business processing information, providing a data foundation for the calculation of risk coefficients and the assessment of business risks in subsequent steps.

[0051] To ensure the accuracy of the prediction results output by the risk prediction model, optionally, in the business processing risk prediction method provided in this application embodiment, the risk prediction model is trained in the following manner: obtaining historical business processing information of multiple preset objects, and a risk coefficient set for each preset object, wherein each risk coefficient set contains multiple risk coefficients associated with the preset object; taking each historical business processing information and the corresponding risk coefficient set as a set of first sample data, obtaining multiple sets of first sample data, and using the multiple sets of first sample data to train the isolated forest algorithm model to obtain the risk prediction model.

[0052] Specifically, before using a risk prediction model, it needs to be trained. First, historical business transaction information of the target entities needs to be collected and prepared. These target entities can be different departments or employees within the company, or external entities that transact with the company. Historical business transaction information includes records of past business activities, such as transaction details, compliance records, and audit reports. This data is used to train the model to identify normal and abnormal behavioral patterns.

[0053] Furthermore, it is necessary to obtain a risk coefficient set for each preset object. This risk coefficient set can be based on historical data and expert evaluation, quantifying the degree of abnormality of each preset object under different risk characteristics. Each set can contain multiple risk coefficients, which reflect the historical performance of the preset object on specific risk characteristics, such as "risk coefficient for violations in bidding methods" and "risk coefficient for violations in contract signing."

[0054] Furthermore, the collected historical information can be bundled with risk coefficients to form a training dataset. Each combination of information and coefficients represents the business status and corresponding risk level of a specific object at a specific point in time, providing material for the Isolation Forest algorithm to learn and identify abnormal patterns.

[0055] Figure 2 This is a schematic diagram of an optional model prediction result provided according to an embodiment of this application, such as... Figure 2 As shown, the circled point represents an abnormal risk point, while the other points are controllable risk points, thus allowing the risk coefficient to be determined based on the prediction results.

[0056] By integrating multiple sets of sample data, the model learns to distinguish between normal and abnormal data. The Isolation Forest algorithm identifies outliers in the data—samples with high anomaly scores—by constructing multiple decision trees. During training, the model learns which combinations of business characteristics and risk coefficients represent high risk, enabling it to identify potential anomalies based on similar data features in future predictions. The trained model can predict the risk level based on new input business processing information and risk characteristics, outputting a risk score that reflects the risk level of the target object under the current business processing situation, thus outputting the risk coefficient for each risk feature.

[0057] This embodiment trains a risk prediction model, effectively utilizing historical business data and risk coefficients to improve the accuracy and reliability of predicting corporate legal compliance risks. During model training, the Isolation Forest algorithm identifies which feature combinations are most likely to lead to abnormal risks, enabling the model to quickly and accurately assess risk levels when faced with new data.

[0058] To accurately determine whether business processing risks exist, optionally, in the business processing risk prediction method provided in this application embodiment, determining whether a target object has business processing risks based on N risk coefficients includes: determining the importance score of the risk feature to which each risk coefficient belongs, obtaining N importance scores; performing a weighted summation of the N risk coefficients and the N importance scores to obtain a risk score; determining whether the risk score is greater than a first threshold; if the risk score is greater than the first threshold, determining that the target object has business processing risks; if the risk score is less than or equal to the first threshold, determining that the target object does not have business processing risks.

[0059] Specifically, when determining whether there are business processing risks, since there may be multiple risk coefficients and different risk coefficients correspond to different levels of importance of risk features, it is first necessary to determine the importance score of each risk feature, multiply the risk coefficient of each risk feature by the corresponding importance score, and then add the product results to obtain the risk score. This ensures that risk features considered more important have a greater weight in the final score, thereby more accurately reflecting the overall risk level of the target object.

[0060] Furthermore, after obtaining the risk score, the risk score can be compared with the first threshold. If the risk score is greater than the first threshold, it is determined that the target object has business processing risks. At this time, it indicates that there are at least one or more high-weight risk features that are abnormal, that is, the target object has substantial risks in business processing, and measures need to be taken to manage risks or adjust business strategies.

[0061] When the risk score is lower than or equal to the first threshold, it indicates that all abnormalities in risk characteristics are within a controllable range and no high-risk warning has been triggered. At this time, the business activities of the target entity can be considered safe and stable, and no immediate risk management intervention is required.

[0062] For example, suppose there are six risk characteristics, and the importance score for each risk characteristic is as follows:

[0063] The risk coefficients for the following six risk characteristics are as follows: Bidding method violation risk coefficient: 0.25; Collusive bidding violation risk coefficient: 0.15; Winning bidder determination violation risk coefficient: 0.10; Winning bid result violation risk coefficient: 0.20; Contract signing violation risk coefficient: 0.15; Procurement process violation risk coefficient: 0.15. The risk coefficients for these six characteristics are: 0.12; 0.17; 0.08; 0.19; 0.07; 0.13. The first threshold is set to 0.7.

[0064] The weighted summation calculation process is as follows: 0.12×0.25+0.17×0.15+0.08×0.10+0.19×0.20+0.07×0.15+0.13×0.15=0.1075.

[0065] Since the calculated risk score is 0.1075, which is lower than the first threshold of 0.7, it can be determined that there is currently no business processing risk for the target object.

[0066] This embodiment determines whether a target object has business processing risks by calculating a risk score through weighted summation. It effectively integrates information from multiple risk characteristics and ensures that the impact of key risk characteristics is not ignored by setting importance scores, thereby improving the accuracy and efficiency of risk assessment.

[0067] To accurately determine the risk level, optionally, in the business processing risk prediction method provided in this application embodiment, determining the risk level of the target object based on N risk coefficients includes: sequentially obtaining a second threshold for the risk coefficient of each risk feature, and sequentially comparing the risk coefficient of each risk feature with the corresponding second threshold to obtain N comparison results; selecting abnormal results from the N comparison results, and determining the risk level of the risk feature to which the abnormal result belongs as the risk level of the target object, wherein the abnormal result indicates that the risk coefficient is greater than the corresponding second threshold.

[0068] Specifically, the second threshold refers to a pre-set limit value for each risk characteristic; exceeding this value indicates a high risk for that characteristic. Once the risk to the target object is determined, the cause of the risk also needs to be identified. Since different risk causes have different impacts, leading to different risk levels, it is necessary to further compare the risk coefficient of each risk characteristic with the corresponding second threshold, obtaining N comparison results. This allows for the determination of the target object's risk status for each specific risk characteristic, i.e., whether it exceeds the normal range. Risk characteristics exceeding the normal range are then identified as abnormal risk characteristics, thereby determining the target object's risk level.

[0069] After obtaining N comparison results, abnormal results can be identified, that is, comparison results with risk coefficients greater than the second threshold. The risk level of each abnormal result is determined as the risk level of the target object, thereby ensuring the accurate determination of the risk level.

[0070] For example, suppose the following risk characteristics and corresponding second thresholds exist: the second threshold for the risk coefficient of irregular bidding method is 0.7; the second threshold for the risk coefficient of irregular contract signing is 0.6; and the second threshold for the risk coefficient of irregular bidding result is 0.8. The target object has the following risk coefficients: the risk coefficient for irregular bidding method is 0.9; the risk coefficient for irregular contract signing is 0.3; and the risk coefficient for irregular bidding result is 0.4.

[0071] Comparing the risk coefficients above with the second threshold: the risk coefficient for violations in the bidding method (0.9) is greater than the second threshold (0.7), indicating an abnormal result. The risk coefficients for violations in contract signing (0.3) and the risk coefficient for violations in the winning bid (0.4) do not exceed the second threshold, therefore they are not considered abnormal results. The only abnormal result is the "risk coefficient for violations in the bidding method." Since the risk level of the "risk coefficient for violations in the bidding method" is the highest risk level, i.e., "high risk level," the risk level of the target object is high because the risk of violations in the bidding method is relatively high.

[0072] This embodiment distinguishes between normal and abnormal risks by setting a second threshold, and determines the overall risk level of the target object based on the risk level of abnormal risk characteristics, effectively improving the pertinence and accuracy of risk prediction.

[0073] Optionally, in the business processing risk prediction method provided in this application embodiment, after determining the risk level of the target object based on N risk coefficients, the method further includes: determining the risk characteristic to which the abnormal risk coefficient belongs among the N risk coefficients, wherein the abnormal risk coefficient is a risk coefficient greater than the corresponding second threshold; inputting the risk characteristic and business processing information into a decision tree model to obtain the reason for the existence of risk characteristics, wherein the decision tree model is trained through multiple sets of second sample data, and each set of second sample data includes historical business processing information and historical risk characteristics of a historical object, as well as the reason for the existence of historical risk characteristics.

[0074] Specifically, to determine the reasons for the risks posed by the target object, firstly, risk coefficients exceeding a second threshold need to be identified, and their corresponding risk characteristics are defined as anomalous risk characteristics. These identified anomalous risk characteristics, along with relevant business processing information, are then fed into a decision tree model. The decision tree model is trained on historical data, learning which historical business processing information is related to anomalies in the risk characteristics. Therefore, when the current risk characteristics and business processing information of the target object are input, the model can predict the specific reasons for these anomalous risk characteristics, and thus provide the target object with the reasons for the anomalies. This allows the target object to take targeted measures to ensure that its business processing can proceed normally.

[0075] For example, after comparing the risk coefficient with the second threshold, the "risk characteristic coefficient of bidding method violation" and the "risk characteristic coefficient of winning bid result violation" were determined as abnormal risk coefficients. Then, the "risk characteristic of bidding method violation" and the "risk characteristic of winning bid result violation" along with relevant business processing information, such as specific bidding documents, winning bid notices, and contract texts, were input into the decision tree model. Through analysis, the decision tree model may identify that the abnormality is caused by the failure of the bidding announcement to be fully open and transparent, or by the excessively frequent revision of the contract terms after the winning bid, which leads to increased doubts and uncertainties about the winning bid result.

[0076] This embodiment uses a decision tree model to conduct an in-depth analysis of the causes of abnormal risk coefficients, determine the abnormal causes of abnormal risk characteristics, and then assist the target object in making adjustments based on the abnormal causes so that it can carry out business transactions normally.

[0077] Optionally, in the business processing risk prediction method provided in this application embodiment, after determining the risk level of the target object based on N risk coefficients, the method further includes: constructing a reinforcement learning model, wherein the risk features are used as state space objects in the model, the adjustment action of the second threshold is used as action space objects in the model, and a reward function is set in the model; the reinforcement learning model is iteratively trained using multiple sets of third sample data, and the second threshold of each risk feature is adjusted through the reinforcement learning model, wherein each set of third sample data includes historical business processing information and historical risk features of a historical object, as well as risk coefficients.

[0078] It should be noted that reinforcement learning is a machine learning technique that enables an agent (or model) to learn how to take actions to maximize some cumulative reward through interaction with the environment.

[0079] Specifically, the goal of the reinforcement learning model is to optimize the accuracy of risk detection by adjusting the second threshold. In the model, the "state space object" refers to the current state of each risk feature, represented by the risk coefficient; the "action space object" refers to the actions that adjust the second threshold; and the "reward function" is designed to evaluate the impact of actions (i.e., the adjustment of the second threshold) on risk detection performance, such as increasing the detection rate or reducing the false positive rate.

[0080] When building a reinforcement learning model, it's necessary to set up a reward function, a state space object, and an action space object. The reward function is crucial in guiding the model's learning; it defines the positive or negative feedback the model receives after taking a specific action. For example, if adjusting the second threshold reduces the false positive rate, the model receives a positive reward; conversely, it receives negative feedback. In this way, the model gradually learns when to adjust the second threshold to achieve optimal risk detection performance.

[0081] Furthermore, after constructing the reinforcement learning model, multiple sets of third-sample data are needed to iteratively train the model. These third-sample data include historical business processing information, historical risk characteristics, and risk coefficients. This allows the model to understand the complex relationship between risk characteristics and the second threshold in historical data, and how adjusting the second threshold can optimize overall risk detection performance. During training, the model will try different second threshold adjustment strategies to observe and learn which adjustments improve risk detection performance. Over time, the model will automatically optimize the second threshold based on feedback from the reward function, achieving a more accurate state and thus improving the accuracy and rationality of risk feature identification.

[0082] This embodiment uses a reinforcement learning model to dynamically adjust the second threshold, automatically optimizing the risk detection threshold based on changes in historical data and actual business scenarios, which greatly improves the accuracy and adaptability of risk assessment.

[0083] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0084] This application also provides a business processing risk prediction device. It should be noted that this business processing risk prediction device can be used to execute the business processing risk prediction method provided in this application. The following describes the business processing risk prediction device provided in this application.

[0085] Figure 3 This is a schematic diagram of a business processing risk prediction device provided according to an embodiment of this application. For example... Figure 3 As shown, the device includes: a first acquisition unit 31, a first input unit 32, and a first determination unit 33.

[0086] The first acquisition unit 31 is used to acquire the business processing information of the target object within the target time period, and process the business processing information to obtain M business feature information, where M is a positive integer.

[0087] The first input unit 32 is used to input M business feature information into the risk prediction model to obtain the risk coefficient of the target object under each risk feature, where N is a positive integer.

[0088] The first determining unit 33 is used to determine whether the target object has business processing risk based on N risk coefficients, and if business processing risk exists, to determine the risk level of the target object based on N risk coefficients.

[0089] The business processing risk prediction device provided in this application embodiment acquires business processing information of a target object within a target time period through a first acquisition unit 31, and processes the business processing information to obtain M business feature information, where M is a positive integer; a first input unit 32 inputs the M business feature information into a risk prediction model to obtain the risk coefficient of the target object under each risk feature, where N is a positive integer; a first determination unit 33 determines whether the target object has business processing risk based on the N risk coefficients, and if business processing risk exists, determines the risk level of the target object based on the N risk coefficients. This solves the problem of low accuracy and efficiency in determining business processing risk in related technologies. By processing business processing information and using business feature information and a risk prediction model to determine the risk coefficients of the target object, and then determining the risk level through multiple risk coefficients, the device achieves the technical effect of improving the accuracy of business processing risk identification.

[0090] Optionally, in the business processing risk prediction device provided in this application embodiment, the first acquisition unit 31 includes: a first determining module, used to acquire M preset business features from the database and determine the calculation method for each preset business feature to obtain M calculation methods; an acquisition module, used to acquire the calculation parameters required for each calculation method to obtain M sets of calculation parameters; a first judging module, used to judge whether there are M sets of calculation parameters in the business processing information; a filling module, used to fill the business processing information to obtain updated business processing information when no parameter in any set of calculation parameters exists in the business processing information; and a first calculation module, used to calculate M business feature information through the business processing information and the M calculation methods when there are M sets of calculation parameters in the business processing information.

[0091] Optionally, in the business processing risk prediction device provided in this application embodiment, the risk prediction model is trained in the following manner: a second acquisition unit is used to acquire historical business processing information of multiple preset objects and a risk coefficient set of each preset object, wherein each risk coefficient set contains multiple risk coefficients associated with the preset object; a first training unit is used to take each historical business processing information and the corresponding risk coefficient set as a set of first sample data, obtain multiple sets of first sample data, and use the multiple sets of first sample data to train the isolated forest algorithm model to obtain the risk prediction model.

[0092] Optionally, in the business processing risk prediction device provided in this application embodiment, the first determining unit 33 includes: a second determining module, used to determine the importance score of the risk feature to which each risk coefficient belongs, to obtain N importance scores; a second calculating module, used to perform a weighted summation of the N risk coefficients and the N importance scores to obtain a risk score; a second judging module, used to judge whether the risk score is greater than a first threshold; a third determining module, used to determine that the target object has business processing risk if the risk score is greater than the first threshold; and a fourth determining module, used to determine that the target object does not have business processing risk if the risk score is less than or equal to the first threshold.

[0093] Optionally, in the business processing risk prediction device provided in this application embodiment, the first determining unit 33 includes: a comparison module, used to sequentially obtain the second threshold of the risk coefficient of each risk feature, and sequentially compare the risk coefficient of each risk feature with the corresponding second threshold to obtain N comparison results; and a fifth determining module, used to select abnormal results from the N comparison results, and determine the risk level of the risk feature to which the abnormal result belongs as the risk level of the target object, wherein the abnormal result indicates that the risk coefficient is greater than the corresponding second threshold.

[0094] Optionally, in the business processing risk prediction device provided in this application embodiment, after determining the risk level of the target object based on N risk coefficients, the device further includes: a second determining unit, used to determine the risk characteristic to which the abnormal risk coefficient belongs among the N risk coefficients, wherein the abnormal risk coefficient is a risk coefficient greater than the corresponding second threshold; and a second input unit, used to input the risk characteristic and business processing information into a decision tree model to obtain the reason for the existence of risk characteristics, wherein the decision tree model is trained through multiple sets of second sample data, and each set of second sample data includes historical business processing information and historical risk characteristics of a historical object, as well as the reason for the existence of historical risk characteristics.

[0095] Optionally, in the business processing risk prediction device provided in this application embodiment, after determining the risk level of the target object based on N risk coefficients, the device further includes: a construction unit for constructing a reinforcement learning model, wherein risk features are used as state space objects in the model, the adjustment action of the second threshold is used as action space objects in the model, and a reward function is set in the model; and a second training unit for iteratively training the reinforcement learning model using multiple sets of third sample data, and adjusting the second threshold of each risk feature through the reinforcement learning model, wherein each set of third sample data includes historical business processing information and historical risk features of a historical object, as well as risk coefficients.

[0096] The aforementioned business processing risk prediction device includes a processor and a memory. The first acquisition unit 31, the first input unit 32, the first determination unit 33, etc., are all stored in the memory as program units. The processor executes the aforementioned program units stored in the memory to realize the corresponding functions.

[0097] The processor contains a kernel, which retrieves the corresponding program units from memory. One or more kernels can be configured; adjusting kernel parameters addresses the low accuracy and efficiency issues in determining business processing risks in related technologies.

[0098] The memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0099] This invention provides a computer-readable storage medium storing a program thereon, which, when executed by a processor, implements the method for predicting business processing risks.

[0100] This invention provides a processor for running a program, wherein the program executes the method for predicting business processing risks during runtime.

[0101] Figure 4 This is a schematic diagram of an electronic device provided according to an embodiment of this application, such as... Figure 4 As shown, this embodiment of the invention provides an electronic device 40, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps of the aforementioned method for predicting business processing risks. The device in this document can be a server, PC, PAD, mobile phone, etc.

[0102] This application also provides a computer program product that, when executed on a data processing device, is adapted to perform the steps of initializing the prediction method for the aforementioned business processing risks.

[0103] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0104] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0105] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0106] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0107] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0108] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, like read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0109] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0110] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0111] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for predicting business processing risks, characterized in that, include: Obtain the business processing information of the target object within the target time period, and process the business processing information to obtain M business feature information, where M is a positive integer; The M business characteristic information is input into the risk prediction model to obtain the risk coefficient of the target object under each risk characteristic, where N is a positive integer; The risk level of the target object is determined based on N risk coefficients to determine whether there is a business processing risk.

2. The method according to claim 1, characterized in that, The business processing information is processed to obtain M business feature information, including: M preset business features are obtained from the database, and the calculation method for each preset business feature is determined to obtain M calculation methods; Obtain the calculation parameters required for each calculation method, resulting in M ​​sets of calculation parameters; Determine whether the M sets of calculation parameters exist in the business processing information; If any parameter from any set of calculation parameters is missing in the business processing information, the business processing information is filled in to obtain updated business processing information; If the M sets of calculation parameters exist in the business processing information, the M business feature information are calculated using the business processing information and the M calculation methods.

3. The method according to claim 1, characterized in that, The risk prediction model is trained in the following manner: Obtain historical business processing information for multiple preset objects, as well as a risk coefficient set for each preset object, wherein each risk coefficient set contains multiple risk coefficients associated with the preset object; Each historical business transaction information and its corresponding risk coefficient set are used as a set of first sample data to obtain multiple sets of first sample data. The isolated forest algorithm model is then trained using the multiple sets of first sample data to obtain the risk prediction model.

4. The method according to claim 1, characterized in that, Determining whether the target object faces business processing risks based on N risk coefficients includes: Determine the importance score of the risk characteristic to which each risk coefficient belongs, resulting in N importance scores; The risk score is obtained by weighting and summing the N risk coefficients and the N importance scores. Determine whether the risk score is greater than a first threshold; If the risk score is greater than the first threshold, it is determined that the target object has the business processing risk. If the risk score is less than or equal to the first threshold, it is determined that the target object does not have the business processing risk.

5. The method according to claim 1, characterized in that, Determining the risk level of the target object based on the N risk coefficients includes: The second threshold of the risk coefficient for each risk feature is obtained sequentially, and the risk coefficient of each risk feature is compared with the corresponding second threshold in turn to obtain N comparison results; An abnormal result is selected from the N comparison results, and the risk level of the risk feature to which the abnormal result belongs is determined as the risk level of the target object, wherein the abnormal result indicates that the risk coefficient is greater than the corresponding second threshold.

6. The method according to claim 1, characterized in that, After determining the risk level of the target object based on the N risk coefficients, the method further includes: Determine the risk characteristic to which the abnormal risk coefficient belongs among the N risk coefficients, wherein the abnormal risk coefficient is a risk coefficient that is greater than the corresponding second threshold; The risk characteristics and the business processing information are input into the decision tree model to obtain the reasons why the risk characteristics exist. The decision tree model is trained by multiple sets of second sample data. Each set of second sample data includes the historical business processing information and historical risk characteristics of a historical object, as well as the reasons why the historical risk characteristics exist.

7. The method according to claim 5, characterized in that, After determining the risk level of the target object based on the N risk coefficients, the method further includes: Construct a reinforcement learning model, wherein risk features are used as state space objects in the model, the adjustment action of the second threshold is used as action space objects in the model, and a reward function is set in the model; The reinforcement learning model is iteratively trained using multiple sets of third sample data, and the second threshold of each risk feature is adjusted using the reinforcement learning model. Each set of third sample data includes historical business processing information and historical risk features of a historical object, as well as a risk coefficient.

8. A device for predicting business processing risks, characterized in that, include: The first acquisition unit is used to acquire the business processing information of the target object within the target time period, and process the business processing information to obtain M business feature information, where M is a positive integer; The first input unit is used to input the M business feature information into the risk prediction model to obtain the risk coefficient of the target object under each risk feature, where N is a positive integer; The first determining unit is used to determine whether the target object has business processing risk based on N risk coefficients, and if the business processing risk exists, to determine the risk level of the target object based on the N risk coefficients.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored executable program, wherein, when the executable program is executed, it controls the device on which the computer-readable storage medium is located to perform the business processing risk prediction method according to any one of claims 1 to 7.

10. An electronic device, characterized in that, include: Memory, which stores executable programs; A processor for running the program, wherein the program executes the business processing risk prediction method according to any one of claims 1 to 7 when it runs.

Images