A non-linear risk assessment method and apparatus
By generating parameters for the risk assessment model and updating the model parameters using an intelligent optimization algorithm, the problem of insufficient adaptability and accuracy in existing risk assessment methods is solved. This enables nonlinear time-varying risk assessment of business processes, improves the timeliness and accuracy of risk warnings, and enhances business processing efficiency and customer satisfaction.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- ULTRAPOWER SOFTWARE
- Filing Date
- 2023-01-18
- Publication Date
- 2026-06-23
AI Technical Summary
Existing risk assessment methods lack adaptability to the interrelationships and changes between business processes, resulting in low accuracy of risk assessments and an inability to identify and promptly address liability risks.
By generating parameters for the risk assessment model, including the risk normalization coefficient, the associated liability risk coefficient, and the liability risk base, risk assessment is performed based on nonlinear time-varying risk values. The model parameters are then updated using an intelligent optimization algorithm to monitor and quantify the liability risks of each processing step in real time.
It improves the adaptability and accuracy of risk assessment, enabling timely detection of potential business process handling liability risks, achieving timely and accurate risk warnings, and enhancing business processing efficiency and customer service quality.
Smart Images
Figure CN116011820B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of computer technology, and in particular to a nonlinear risk assessment method and apparatus. Background Technology
[0002] Business processing systems commonly used in enterprises can handle multiple business processes. The risk assessment model set up in the business processing system can help enterprises standardize IT processing processes, find risk points in business processes, and thus achieve the goal of effective internal risk control. Among them, the risk point in the business process means that if a certain processing step in the business process is not completed on time, it will lead to the risk that the entire business process cannot be completed on time.
[0003] Existing risk assessment methods mainly involve setting monitoring points between each processing step of a business process. When a monitoring point detects that a corresponding processing step has not been completed in a timely manner, it outputs a pre-set fixed risk value. Then, based on the fixed risk values corresponding to each processing step, the total fixed risk value of each business process in the business processing system is determined. When the total fixed risk value reaches the warning threshold, a risk alarm is issued according to the preset alarm rules.
[0004] However, the risk assessment method based on fixed risk values mentioned above only sets a fixed risk value for each processing step, without considering the correlation between different business processes, or the impact of changes in business processes and time on the priority of business processes and the risk values of each processing step. Therefore, the risk assessment has poor adaptability and low accuracy. Summary of the Invention
[0005] This application provides a nonlinear risk assessment method and apparatus, which can be used to solve the technical problems of poor adaptability and low accuracy of traditional risk assessment methods.
[0006] In a first aspect, embodiments of this application provide a nonlinear risk assessment method, the method comprising:
[0007] Based on each processing flow and processing steps in the business processing system, parameters of the risk assessment model are generated. The parameters of the risk assessment model include the risk normalization coefficient corresponding to each processing flow, the associated liability risk coefficient of any processing step in the processing flow that is not completed and the liability risk base of the incomplete processing step to its own responsible person.
[0008] Within a preset data observation period, the nonlinear time-varying risk value of each processing step in each processing flow of the business processing system is obtained based on the risk assessment model.
[0009] Based on each nonlinear time-varying risk value, the dynamic comprehensive risk value of the business processing system is determined;
[0010] When the dynamic comprehensive risk value of the business processing system exceeds the system risk warning threshold, a system risk warning message is generated.
[0011] In conjunction with the first aspect, in one possible implementation of the first aspect, the method further includes:
[0012] After the data observation period ends, the parameters of the risk assessment model are globally optimized based on a preset intelligent optimization algorithm.
[0013] If an optimal solution exists, the parameters of the risk assessment model are updated to the optimal solution to obtain the updated risk assessment model.
[0014] In the next data observation period, the nonlinear time-varying risk value of each processing step of each processing flow in the business processing system is obtained based on the updated risk assessment model.
[0015] In conjunction with the first aspect, in one possible implementation of the first aspect, the global optimization of the parameters of the risk assessment model based on a preset intelligent optimization algorithm includes:
[0016] The dynamic comprehensive risk value of the business processing system is determined as the fitness function of the genetic algorithm, and the normalization coefficients of each risk and the associated liability risk coefficients are determined as the optimization variables of the genetic mutation population.
[0017] With the goal of globally minimizing the fitness function, intelligent optimization iteration is performed.
[0018] In conjunction with the first aspect, in one possible implementation of the first aspect, the risk normalization coefficient corresponding to each processing step is generated in the following way:
[0019] Based on the type of each processing flow, all processing flows in the business processing system are classified.
[0020] The processing flow is quantitatively modeled according to its importance and priority to obtain a processing type risk coefficient vector. The processing type risk coefficient vector includes the risk normalization coefficient corresponding to each type, and the risk normalization coefficient corresponding to each type is determined according to the importance of the type.
[0021] The risk normalization coefficient corresponding to the type to which the processing flow belongs is determined as the risk normalization coefficient corresponding to the processing flow.
[0022] In conjunction with the first aspect, in one possible implementation of the first aspect, the processing type risk coefficient vector is represented by the following formula:
[0023]
[0024] Among them, R C This is the risk coefficient vector for the processing type. Let be the risk normalization coefficient corresponding to the i-th type of processing flow, and The integer i is greater than or equal to 0 and less than or equal to 1, where i is an integer greater than or equal to 1 and less than or equal to N, and N is the total number of processing flow types.
[0025] In conjunction with the first aspect, in one possible implementation of the first aspect, the associated liability risk coefficient for the responsible parties of each processing step in the processing flow due to the failure of any processing step in the processing flow is generated in the following manner:
[0026]
[0027] in, This is the associated liability risk coefficient matrix corresponding to the i-th type of processing flow. The associated liability risk coefficient for the person responsible for the j-th processing step in the i-th processing flow if the j-th processing step is not completed. The risk coefficient representing the associated liability of the person responsible for the (j-1)th processing step in the (j-1)th processing step due to the failure of the j-th processing step in the i-th processing flow. M represents the associated liability risk coefficient for the person responsible for the (j+1)th processing step in the i-th processing flow if the j-th processing step is not completed, and M is the total number of processing steps in the i-th processing flow.
[0028] In conjunction with the first aspect, in one possible implementation of the first aspect, obtaining the nonlinear time-varying risk value of each processing step of each processing flow in the business processing system based on the risk assessment model includes:
[0029] The nonlinear time-varying risk value of each processing step in each processing flow of the business processing system is obtained using the following formula:
[0030]
[0031] Where, r i,n (t) represents the nonlinear time-varying risk value of the nth processing step in the i-th processing flow at time t. This represents the risk normalization coefficient corresponding to the i-th type of processing procedure within the current data observation period. The associated liability risk coefficient is the risk factor for the responsible persons of each processing step in the i-th type of processing flow if the nth processing step is not completed within the current data observation period. The comprehensive risk value generated by the i-th type of processing flow at time t is... Determined by the following formula:
[0032]
[0033] in, The comprehensive risk liability value generated by the i-th type of processing flow at time t is determined based on the associated liability risk coefficient and the liability risk base. The associated liability risk coefficient for the person responsible for the l-th processing step in the i-th processing flow if the j-th processing step is not completed. M represents the base risk of liability for the person responsible for the j-th processing step if the j-th processing step is not completed at time t, and M represents the total number of processing steps in the i-th type of processing flow.
[0034] In conjunction with the first aspect, in one possible implementation of the first aspect, determining the dynamic comprehensive risk value of the business processing system based on various nonlinear time-varying risk values includes:
[0035] The dynamic comprehensive risk value of the business processing system is determined by the following formula:
[0036]
[0037] Among them, S R (t) represents the dynamic comprehensive risk value of the business processing system, R C To handle the type risk coefficient vector, The comprehensive risk liability value generated by the i-th type of processing flow at time t is determined based on the associated liability risk coefficient and the liability risk base, where i is an integer greater than or equal to 1 and less than or equal to N, and N is the total number of processing flow types. The risk normalization coefficient corresponding to the i-th type of processing flow is... The associated liability risk coefficient for the person responsible for the l-th processing step in the i-th processing flow if the j-th processing step is not completed. M represents the base risk of liability for the person responsible for the j-th processing step if the j-th processing step is not completed at time t, and M represents the total number of processing steps in the i-th type of processing flow.
[0038] In conjunction with the first aspect, in one possible implementation of the first aspect, the base amount of liability risk arising to the responsible party from the incomplete processing steps is generated in the following manner:
[0039]
[0040] in, Let this be the base amount of liability risk for the person responsible for the j-th processing step if the j-th processing step is not completed at time t. Let k be the processing time limit for the j-th processing step, and k0, k1, a, and b are adjustable parameters.
[0041] Secondly, embodiments of this application provide a nonlinear risk assessment device, the device comprising:
[0042] The parameter generation unit is used to generate parameters for the risk assessment model based on each processing flow and processing steps in the business processing system. The parameters of the risk assessment model include the risk normalization coefficient corresponding to each processing flow, the associated liability risk coefficient of any processing step in the processing flow that is not completed, and the liability risk base of the incomplete processing step to its own responsible person.
[0043] The risk monitoring unit is used to obtain the nonlinear time-varying risk value of each processing step of each processing flow in the business processing system based on the risk assessment model within a preset data observation period.
[0044] The risk assessment unit is used to determine the dynamic comprehensive risk value of the business processing system based on various nonlinear time-varying risk values.
[0045] The risk warning unit is used to generate system risk warning information when the dynamic comprehensive risk value of the business processing system exceeds the system risk warning threshold.
[0046] This application provides a nonlinear risk assessment method and apparatus. In the risk assessment method, parameters of a pre-built risk assessment model are generated based on each processing flow and processing step within a business processing system. Then, within a preset data observation period, the nonlinear time-varying risk values of each processing step in each processing flow are monitored based on the risk assessment model. The dynamic comprehensive risk value of the business processing system is determined in real time based on these nonlinear time-varying risk values, and an alarm is issued when the dynamic comprehensive risk value exceeds a threshold. The entire method performs real-time nonlinear quantitative modeling of the responsibility risk of each processing step in the processing flow, considering the correlation of responsibility between different processing steps, and the impact of changes in business processes and completion times on the risk values of each processing step. This allows the risk assessment model to perform risk assessment based on nonlinear time-varying risk values, exhibiting strong adaptability and providing timely and accurate risk assessment and early warning. Attached Figure Description
[0047] Figure 1 This is a schematic diagram of risk monitoring in a traditional risk assessment model;
[0048] Figure 2 This is a schematic diagram of the workflow of a nonlinear risk assessment method provided in an embodiment of this application;
[0049] Figure 3 This is a schematic diagram of risk monitoring using the risk assessment model provided in the embodiments of this application;
[0050] Figure 4 This is a schematic diagram of the structure of a nonlinear risk assessment device provided in an embodiment of this application. Detailed Implementation
[0051] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.
[0052] The following section will first introduce the existing risk assessment methods.
[0053] Existing risk assessment methods primarily involve setting monitoring points between each processing step in a business process. When a monitoring point detects that a corresponding processing step has not been completed in a timely manner, a pre-set fixed risk value is output. Then, based on the fixed risk values corresponding to each processing step, the total fixed risk value for all business processes within the business processing system is determined. When the total fixed risk value reaches a warning threshold, a risk alarm is issued according to preset alarm rules. Risk monitoring for each processing step can be implemented using a risk assessment model.
[0054] Figure 1 This is a schematic diagram of risk monitoring in a traditional risk assessment model. (Refer to...) Figure 1 As shown, traditional risk assessment models set up risk detection points and monitoring units between each processing step in a business process. When any processing step n is not processed in a timely manner, a pre-set fixed risk value rn is generated. The risk detection point monitors the processing information flow of the corresponding processing step, while the monitoring unit, based on the risk detection point's monitoring of the processing information flow, adds monitoring logic, such as monitoring whether the maximum allowable processing time has been exceeded and outputting the corresponding risk value. The fixed risk value corresponding to each processing step can be preset based on human experience; for example, r1 = 0.13, r2 = 0.21, ..., r(n-1) = 0.27, rn = 0.33.
[0055] However, existing risk assessment methods lack dynamic optimization modeling for the risk value of each processing step, and lack real-time risk value correlation and early warning methods for the responsible persons of each processing step. This results in many potential risks related to processing responsibilities not being identified and warned of in a timely manner. Furthermore, existing risk assessment methods also lack methods for comprehensive, intelligent, and quantitative optimization modeling of the overall risk of all business processes in the business processing system, and real-time early warning methods for the comprehensive risk value of the person in charge of the overall business processing. This leads to the inability to accurately assess, warn, and handle overall comprehensive risk in real time. When a new processing process is added, the correlation information of the corresponding process cannot be updated to the fixed risk assessment model in a timely and accurate manner. For example, a mobile network operator's newly built primary-level mobile hall backend system experienced a failure in the timely processing of a service upgrade, resulting in complaints related to account management, payment / recharge services. The failure remained unresolved for 20 minutes, leading to 32 new complaints. As another example, a newly built IoT business system of an operator experienced an error in an instance of the aggregated IoT-WEB module. Due to the failure to handle the error within one hour, complaints related to government and enterprise services and IoT services were generated, resulting in 52 new complaints. The two cases above illustrate that existing fixed risk assessment methods cannot automatically adapt to changes in associated process risks and overall business processing risks during new business processes, and their shortcomings are very obvious.
[0056] Based on the above introduction, in order to address the technical problems of poor adaptability and low accuracy of existing risk assessment methods, this application discloses a nonlinear risk assessment method through the following embodiments. The nonlinear risk assessment method provided in this application utilizes a risk assessment model to monitor risks based on the nonlinear time-varying risk values of each processing step, thereby taking into account the correlation between various business processes. Furthermore, after each data observation period, an optimization algorithm can be used to globally optimize and update the parameters of the risk assessment model, thereby significantly improving the accuracy of risk assessment. See also... Figure 2 The flowchart shown illustrates a nonlinear risk assessment method provided in this application, which specifically includes the following steps:
[0057] 101: Generate parameters for the risk assessment model based on each processing flow and processing steps in the business processing system.
[0058] The parameters of the risk assessment model include the risk normalization coefficient corresponding to each processing step, the associated liability risk coefficient of any processing step in the processing step not being completed, and the liability risk base of the incomplete processing step to its own responsible person.
[0059] Figure 3 This is a schematic diagram of risk monitoring for the risk assessment model provided in this application embodiment. (Refer to...) Figure 3As shown, the risk assessment model provided in this application embodiment can be used to obtain the nonlinear time-varying risk value of each processing step in each processing flow of the business processing system.
[0060] In some embodiments, the risk normalization coefficient corresponding to each processing step can be generated through the following steps:
[0061] The first step is to classify all processing processes in the business processing system based on the type of each processing process.
[0062] For example, the complaint handling process in the business processing system can be divided into N types according to the type of complaint, such as family business complaints, account management and payment complaints, billing and payment complaints, electronic channel complaints, and personal business complaints.
[0063] The second step is to quantitatively model various processing procedures according to their importance and priority, and obtain the risk coefficient vector of each processing type.
[0064] The processing type risk coefficient vector includes the risk normalization coefficient corresponding to each type, and the risk normalization coefficient corresponding to each type is determined by assigning values according to the importance of the type.
[0065] Specifically, the treatment type risk coefficient vector can be represented by the following formula (1):
[0066]
[0067] In formula (1), R C To handle the type risk coefficient vector, Let be the risk normalization coefficient corresponding to the i-th type of processing flow, and Greater than or equal to 0 and less than or equal to 1, i is an integer greater than or equal to 1 and less than or equal to N, and N is the total number of processing flow types. In formula (1), They are arranged in order of importance and priority. The specific value can be set and adjusted by those skilled in the art according to actual needs. For example, It can be assigned the value 0.12. It can be assigned the value 0.27. It can be assigned the value 0.31. It can be assigned a value of 0.43.
[0068] The third step is to determine the risk normalization coefficient corresponding to the type of processing procedure as the risk normalization coefficient corresponding to the processing procedure.
[0069] This application embodiment also comprehensively and quantitatively models the processing responsibility risks and related (supervisory) responsibility risks of various processing procedures. Processing responsibility risk refers to the responsibility risk arising from the dereliction of duty by the person responsible for a processing step in failing to perform the procedure actions in a timely manner. Related responsibility risk refers to the (supervisory) related responsibility risk that a subsequent party bears when a procedure action is not completed in a timely manner and the subsequent party fails to fulfill its supervisory responsibility. For example, in the i-th type of complaint processing procedure, there are M processing steps, which may involve complaint intelligent classification and review processing responsibility risks, complaint automatic order creation and review processing responsibility risks, automatic order dispatch and review processing responsibility risks, complaint processing responsibility risks, processing result review processing responsibility risks, and report review processing responsibility risks. Furthermore, there may also be related responsibility risks in complaint intelligent classification and review, complaint automatic order creation and review, automatic order dispatch and review, complaint processing, processing result review, and report review.
[0070] In some embodiments, the associated liability risk coefficient for the responsible party of each processing step in the processing flow due to the failure to complete any processing step in the processing flow can be generated in the following manner:
[0071]
[0072] In formula (2), This is the associated liability risk coefficient matrix corresponding to the i-th type of processing flow. The associated liability risk coefficient for the person responsible for the j-th processing step in the i-th processing flow if the j-th processing step is not completed. The risk coefficient representing the associated liability of the person responsible for the (j-1)th processing step in the (j-1)th processing step due to the failure of the j-th processing step in the i-th processing flow. M represents the associated liability risk coefficient for the person responsible for the (j+1)th processing step in the i-th processing flow if the j-th processing step is not completed, and M is the total number of processing steps in the i-th processing flow.
[0073] Suppose that after a complaint of type i occurs, M processing steps are required. The processing time vector for these M processing steps can be represented by the following formula (3):
[0074]
[0075] In formula (3), To process time-limited vectors, This refers to the processing time limit for the j-th processing step, meaning that the j-th processing step out of these M processing steps needs to be completed within a specified timeframe. Complete before the deadline.
[0076] Then, when the j-th processing step is completed at time t, the base amount of liability risk for the person responsible for the j-th processing step can be determined by the following formula (4):
[0077]
[0078] In formula (4), Let this be the base amount of liability risk for the person responsible for the j-th processing step if the j-th processing step is not completed at time t. Let k be the processing time limit for the j-th processing step, and k0, k1, a, and b are adjustable parameters.<k0,k1,a,b> This is called the risk base parameter quadruple, and the parameters of the risk base parameter quadruple can be flexibly adjusted according to the actual situation. When And if the j-th processing step is not completed, Reaching the maximum liability risk threshold.
[0079] when Furthermore, if the j-th processing step is not completed, the liability risk value for the person responsible for the j-th processing step can be determined by the following formula (5):
[0080]
[0081] In formula (5), The liability risk value arising from the failure to complete the j-th processing step for the person responsible for the j-th processing step. The associated liability risk coefficient for the person responsible for the j-th processing step in the i-th processing flow if the j-th processing step is not completed. This represents the base amount of liability risk incurred by the person responsible for the j-th processing step if the j-th processing step is not completed at time t.
[0082] Accordingly, if the j-th processing step is not completed at time t, the liability risk value for the person responsible for the (j-1)-th processing step can be determined by the following formula (6):
[0083]
[0084] In formula (6), The liability risk value arising from the failure to complete the j-th processing step for the person responsible for the (j-1)-th processing step. The risk coefficient representing the associated liability of the person responsible for the (j-1)th processing step in the (j-1)th processing step due to the failure of the j-th processing step in the i-th processing flow. This represents the base amount of liability risk incurred by the person responsible for the j-th processing step if the j-th processing step is not completed at time t.
[0085] Accordingly, if the j-th processing step is not completed at time t, the comprehensive liability risk value for the responsible parties of all M processing steps can be determined by the following formula (7):
[0086]
[0087] In formula (7), The total liability risk value arising from the failure to complete the j-th processing step for all M processing steps. The associated liability risk coefficient for the person responsible for the l-th processing step in the i-th processing flow if the j-th processing step is not completed. This represents the base amount of liability risk incurred by the person responsible for the j-th processing step if the j-th processing step is not completed at time t.
[0088] When at time t, and Once the j-th processing step is completed, that is, once the j-th processing step is completed ahead of schedule, the sum of all risks and related risks arising from the j-th processing step will be zero.
[0089] Accordingly, the comprehensive risk liability value generated by the i-th type of processing procedure at time t can be determined by the following formula (8):
[0090]
[0091] In formula (8), Let be the comprehensive risk liability value generated by the i-th type of processing flow at time t. The associated liability risk coefficient for the person responsible for the l-th processing step in the i-th processing flow if the j-th processing step is not completed. M represents the base risk of liability for the person responsible for the j-th processing step if the j-th processing step is not completed at time t, and M represents the total number of processing steps.
[0092] In this embodiment of the application, the comprehensive risk value generated by the i-th type of processing flow at time t can be determined by the following formula (9):
[0093]
[0094] In formula (9), Let be the comprehensive risk value generated by the i-th type of processing flow at time t. The comprehensive risk liability value generated by the i-th type of processing flow at time t is determined based on the associated liability risk coefficient and the liability risk base. The associated liability risk coefficient for the person responsible for the l-th processing step in the i-th processing flow if the j-th processing step is not completed. M represents the base risk of liability for the person responsible for the j-th processing step if the j-th processing step is not completed at time t, and M represents the total number of processing steps in the i-th type of processing flow.
[0095] In addition, after executing step 101 and before executing step 102, the method provided in this application embodiment may further include: setting initial values for the parameters of the risk assessment model. Specifically, the initial values can be determined based on experience or actual conditions, and this application embodiment does not impose specific limitations on this.
[0096] 102: Within a preset data observation period, obtain the nonlinear time-varying risk value of each processing step in each processing flow of the business processing system based on the risk assessment model.
[0097] Specifically, the risk assessment model can obtain the nonlinear time-varying risk value of each processing step in each processing flow of the business processing system based on the following formula (10):
[0098]
[0099] In formula (10), r i,n (t) represents the nonlinear time-varying risk value of the nth processing step in the i-th processing flow at time t. This represents the risk normalization coefficient corresponding to the i-th type of processing procedure within the current data observation period. This represents the associated liability risk coefficient for the responsible parties in each processing step of the i-th type of processing flow if the nth processing step is not completed within the current data observation period. Let be the comprehensive risk value generated by the i-th type of processing flow at time t.
[0100] In some embodiments, when the comprehensive risk liability value generated by the i-th type of processing flow at time t When the risk exceeds the first risk warning threshold, a first risk warning message can be generated. For example, a risk warning can be sent to the relevant responsible person via a Super SIM card message. At the same time, a processing interface can be provided to the relevant responsible person through a smartphone terminal APP or Super SIM card H5 mini-program to perform timely complaint handling actions.
[0101] 103: Based on each nonlinear time-varying risk value, determine the dynamic comprehensive risk value of the business processing system.
[0102] Specifically, the dynamic comprehensive risk value of the business processing system can be determined using the following formula (11):
[0103]
[0104] In formula (11), S R(t) represents the dynamic comprehensive risk value of the business processing system, R C To handle the type risk coefficient vector, The comprehensive risk liability value generated by the i-th type of processing flow at time t is determined based on the associated liability risk coefficient and the liability risk base, where i is an integer greater than or equal to 1 and less than or equal to N, and N is the total number of processing flow types. The risk normalization coefficient corresponding to the i-th type of processing flow is... The associated liability risk coefficient for the person responsible for the l-th processing step in the i-th processing flow if the j-th processing step is not completed. M represents the base risk of liability for the person responsible for the j-th processing step if the j-th processing step is not completed at time t, and M represents the total number of processing steps in the i-th type of processing flow.
[0105] 104: When the dynamic comprehensive risk value of the business processing system exceeds the system risk warning threshold, generate system risk warning information.
[0106] For example, risk warnings can be sent to relevant management personnel through Super SIM card guaranteed delivery messages. At the same time, relevant management personnel can also be provided with a processing interface through smartphone terminal APP or Super SIM card H5 mini program to perform timely complaint handling actions.
[0107] Furthermore, the method provided in the embodiments of this application may also include:
[0108] 105: After the data observation period ends, the parameters of the risk assessment model are globally optimized based on the preset intelligent optimization algorithm.
[0109] In some embodiments, the dynamic comprehensive risk value of the business processing system can be determined as the fitness function F of the genetic algorithm, and the normalized coefficients of each risk and the associated liability risk coefficients can be determined as the optimization variables of the genetic mutation population. That is, the coefficients in the processing type risk coefficient vector and the associated liability risk coefficient matrix are used as the optimization variables of the genetic mutation population. The intelligent optimization iteration is performed with the goal of globally minimizing the fitness function F. Specifically, it can be expressed by the following formula (12):
[0110]
[0111] In formula (12), S R (t) represents the dynamic comprehensive risk value of the business processing system, R C To handle the type risk coefficient vector, Let T be the associated liability risk coefficient matrix corresponding to the i-th type of processing flow, and T be the data observation period.
[0112] 106: Check if an optimal solution exists. If an optimal solution exists, proceed to step 107; otherwise, return to step 102 and enter the next data observation cycle. Based on the risk assessment model, obtain the nonlinear time-varying risk value of each processing step in each processing flow of the business processing system.
[0113] 107: Update the parameters of the risk assessment model to the optimal solution to obtain the updated risk assessment model.
[0114] Specifically, once the global minimum value of F is found, the coefficient values of the treatment type risk coefficient vector and the associated liability risk coefficient matrix in the genetic variation population at that time are extracted, and the parameters of the risk assessment model are updated. In this way, based on historical business processing big data and combined with artificial intelligence parameter global optimization technology, the parameters of the nonlinear risk assessment model are periodically optimized and adjusted, thereby making risk assessment and early warning more timely and accurate.
[0115] 108: In the next data observation period, obtain the nonlinear time-varying risk value of each processing step in each processing flow of the business processing system based on the updated risk assessment model. Then return to step 103.
[0116] Thus, the risk assessment method provided in this application, with its innovative risk coefficient modeling technique in a nonlinear risk assessment model, is highly compatible with artificial intelligence optimization algorithms. Compared to existing risk assessment models with fixed risk values, it can provide more timely and accurate early warnings of liability risks, minimizing overall system risk. Furthermore, the responsibility internal control quantitative modeling methods, such as processing type risk coefficient vectors, associated responsibility risk matrices, and time-varying nonlinear responsibility risk bases, can accurately and in real-time identify potential business process handling responsibility risks and possess convenient localized parameter tuning capabilities. Moreover, the method provided in this application innovatively combines intelligent automation technology for risk responsibility models with business processing and risk responsibility internal control technology. This better assists personnel at each stage of the process in automatically executing corresponding tasks at each control point of responsibility. The efficient and organic integration with process and safety responsibility management can significantly improve business processing efficiency, shorten the average processing time of business processes, and enhance business processing satisfaction, thereby achieving a comprehensive improvement in work efficiency and customer service quality. Therefore, the risk assessment method provided in this application has significant technical advantages.
[0117] The following are embodiments of the apparatus described in this application, which can be used to execute the embodiments of the method described in this application. For details not disclosed in the apparatus embodiments of this application, please refer to the embodiments of the method described in this application.
[0118] Figure 4 This is a schematic diagram of the structure of a nonlinear risk assessment device provided in an embodiment of this application. Figure 4As shown, the device has the function of implementing the aforementioned nonlinear risk assessment method. This function can be implemented in hardware or by hardware executing corresponding software. The device may include: a parameter generation unit 401, a risk monitoring unit 402, a risk assessment unit 403, and a risk warning unit 404. Wherein:
[0119] The parameter generation unit 401 is used to generate parameters for the risk assessment model based on each processing flow and processing steps in the business processing system. The parameters of the risk assessment model include the risk normalization coefficient corresponding to each processing flow, the associated liability risk coefficient of any processing step in the processing flow that is not completed, and the liability risk base of the incomplete processing step to its own responsible person.
[0120] Risk monitoring unit 402 is used to obtain the nonlinear time-varying risk value of each processing step of each processing process in the business processing system based on the risk assessment model within a preset data observation period.
[0121] Risk assessment unit 403 is used to determine the dynamic comprehensive risk value of the business processing system based on each nonlinear time-varying risk value.
[0122] Risk warning unit 404 is used to generate system risk warning information when the dynamic comprehensive risk value of the business processing system exceeds the system risk warning threshold.
[0123] In one possible implementation, the device further includes:
[0124] The parameter update unit is used to perform global optimization of the parameters of the risk assessment model based on a preset intelligent optimization algorithm after the data observation period ends, and to update the parameters of the risk assessment model to the optimal solution if an optimal solution exists, thereby obtaining the updated risk assessment model.
[0125] The risk monitoring unit 402 is also used to obtain the nonlinear time-varying risk value of each processing step of each processing process in the business processing system based on the updated risk assessment model in the next data observation period.
[0126] In one possible implementation, the parameter update unit is specifically used for:
[0127] The dynamic comprehensive risk value of the business processing system is determined as the fitness function of the genetic algorithm, and the normalization coefficients of each risk and the associated liability risk coefficients are determined as the optimization variables of the genetic mutation population.
[0128] With the goal of globally minimizing the fitness function, intelligent optimization iteration is performed.
[0129] In one feasible approach, the risk normalization coefficient for each processing step is generated as follows:
[0130] Based on the type of each processing flow, all processing flows in the business processing system are classified.
[0131] The processing procedures are quantitatively modeled according to their importance and priority, resulting in a processing type risk coefficient vector. This vector includes a risk normalization coefficient for each type, which is determined by assigning values based on the type's importance.
[0132] The risk normalization coefficient corresponding to the type of processing procedure is determined as the risk normalization coefficient corresponding to the processing procedure.
[0133] In one feasible approach, the processing type risk coefficient vector is represented by the following formula:
[0134]
[0135] Among them, R C To handle the type risk coefficient vector, Let be the risk normalization coefficient corresponding to the i-th type of processing flow, and The integer i is greater than or equal to 0 and less than or equal to 1, where i is an integer greater than or equal to 1 and less than or equal to N, and N is the total number of processing flow types.
[0136] In one feasible approach, the associated liability risk coefficient for the responsible parties at each step of the process flow due to the incompleteness of any processing step is generated in the following manner:
[0137]
[0138] in, This is the associated liability risk coefficient matrix corresponding to the i-th type of processing flow. The associated liability risk coefficient for the person responsible for the j-th processing step in the i-th processing flow if the j-th processing step is not completed. The risk coefficient representing the associated liability of the person responsible for the (j-1)th processing step in the (j-1)th processing step due to the failure of the j-th processing step in the i-th processing flow. M represents the associated liability risk coefficient for the person responsible for the (j+1)th processing step in the i-th processing flow if the j-th processing step is not completed, and M is the total number of processing steps in the i-th processing flow.
[0139] In one possible implementation, the risk monitoring unit 402 is specifically used for:
[0140] The nonlinear time-varying risk value of each processing step in each processing flow of the business processing system is obtained using the following formula:
[0141]
[0142] Where, r i,n (t) represents the nonlinear time-varying risk value of the nth processing step in the i-th processing flow at time t. This represents the risk normalization coefficient corresponding to the i-th type of processing procedure within the current data observation period. This represents the associated liability risk coefficient for the responsible parties in each processing step of the i-th type of processing flow if the nth processing step is not completed within the current data observation period. Let be the comprehensive risk value generated by the i-th type of processing flow at time t. Determined by the following formula:
[0143]
[0144] in, The comprehensive risk liability value generated by the i-th type of processing flow at time t is determined based on the associated liability risk coefficient and the liability risk base. The associated liability risk coefficient for the person responsible for the l-th processing step in the i-th processing flow if the j-th processing step is not completed. M represents the base risk of liability for the person responsible for the j-th processing step if the j-th processing step is not completed at time t, and M represents the total number of processing steps in the i-th type of processing flow.
[0145] In one possible implementation, the risk assessment unit 403 is specifically used for:
[0146] The dynamic comprehensive risk value of the business processing system is determined by the following formula:
[0147]
[0148] Among them, S R (t) represents the dynamic comprehensive risk value of the business processing system, R C To handle the type risk coefficient vector, The comprehensive risk liability value generated by the i-th type of processing flow at time t is determined based on the associated liability risk coefficient and the liability risk base, where i is an integer greater than or equal to 1 and less than or equal to N, and N is the total number of processing flow types. The risk normalization coefficient corresponding to the i-th type of processing flow is... The associated liability risk coefficient for the person responsible for the first processing step in the i-th type of processing flow if the j-th processing step is not completed. M represents the base risk of liability for the person responsible for the j-th processing step if the j-th processing step is not completed at time t, and M represents the total number of processing steps in the i-th type of processing flow.
[0149] In one feasible approach, the base amount of liability risk arising from incomplete processing steps to the responsible party is generated in the following manner:
[0150]
[0151] in, Let this be the base amount of liability risk for the person responsible for the j-th processing step if the j-th processing step is not completed at time t. Let k be the processing time limit for the j-th processing step, and k0, k1, a, and b are adjustable parameters.
[0152] The embodiments of this application provide a nonlinear risk assessment device, which is based on real-time nonlinear quantitative modeling of the responsibility risk of each processing step in each processing flow. It considers the correlation of responsibility between different processing steps, as well as the impact of changes in business processes and completion times on the risk values of each processing step. This enables the risk assessment model to perform risk assessment based on nonlinear time-varying risk values, which has strong adaptability and timely and accurate risk assessment and early warning.
[0153] The present application has been described in detail above with reference to specific embodiments and exemplary examples; however, these descriptions should not be construed as limiting the present application. Those skilled in the art will understand that various equivalent substitutions, modifications, or improvements can be made to the technical solutions and implementation methods of the present application without departing from the spirit and scope of the present application, and all such modifications and improvements fall within the scope of the present application. The scope of protection of the present application is determined by the appended claims.
Claims
1. A nonlinear risk assessment method, characterized in that, The method includes: Based on each processing flow and processing steps in the business processing system, parameters of the risk assessment model are generated. The parameters of the risk assessment model include the risk normalization coefficient corresponding to each processing flow, the associated liability risk coefficient of any processing step in the processing flow that is not completed and the liability risk base of the incomplete processing step to its own responsible person. The base amount of liability risk arising from the incomplete processing steps to the responsible party is generated in the following way: ； in, Let this be the base amount of liability risk for the person responsible for the j-th processing step if the j-th processing step is not completed at time t. Let k0, k1, a, and b be the processing time limit for the j-th processing step, and k0, k1, a, and b be adjustable parameters; when at time t, and t < t / 2, ... Once the j-th processing step is completed, that is, if the j-th processing step is completed ahead of schedule, the sum of all risks and liabilities arising from the j-th processing step and related risks and liabilities will be zero. Within a preset data observation period, the nonlinear time-varying risk value of each processing step in each processing flow of the business processing system is obtained based on the risk assessment model. Based on each nonlinear time-varying risk value, the dynamic comprehensive risk value of the business processing system is determined; When the dynamic comprehensive risk value of the business processing system exceeds the system risk warning threshold, a system risk warning message is generated. The method further includes: After the data observation period ends, the dynamic comprehensive risk value of the business processing system is determined as the fitness function of the genetic algorithm, and each risk normalization coefficient and each associated liability risk coefficient are determined as the genetic variation population optimization variables. With the goal of globally minimizing the fitness function, intelligent optimization iteration is performed based on a preset intelligent optimization algorithm; When the minimum value of the fitness function is found, the coefficient values of the treatment type risk coefficient vector and the associated liability risk coefficient matrix in the genetic variation population are extracted, and the parameters of the risk assessment model are updated to obtain the updated risk assessment model; the treatment type risk coefficient vector includes the risk normalization coefficient corresponding to each type; In the next data observation period, the nonlinear time-varying risk value of each processing step of each processing flow in the business processing system is obtained based on the updated risk assessment model.
2. The method according to claim 1, characterized in that, The risk normalization coefficient for each processing step is generated in the following way: Based on the type of each processing flow, all processing flows in the business processing system are classified. Quantitative modeling of various processing procedures is performed according to their importance and priority to obtain a processing type risk coefficient vector; The risk normalization coefficients for each type are determined by assigning values according to the importance of the type; The risk normalization coefficient corresponding to the type to which the processing flow belongs is determined as the risk normalization coefficient corresponding to the processing flow.
3. The method according to claim 2, characterized in that, The risk coefficient vector for the treatment type is represented by the following formula: ； Among them, R C This is the risk coefficient vector for the processing type. Let be the risk normalization coefficient corresponding to the i-th type of processing flow, and The integer i is greater than or equal to 0 and less than or equal to 1, where i is an integer greater than or equal to 1 and less than or equal to N, and N is the total number of processing flow types.
4. The method according to claim 1, characterized in that, The associated liability risk coefficient for the responsible parties at each step of the processing flow if any processing step is not completed is generated in the following manner: ； in, This is the associated liability risk coefficient matrix corresponding to the i-th type of processing flow. The associated liability risk coefficient for the person responsible for the j-th processing step in the i-th processing flow if the j-th processing step is not completed. The risk coefficient representing the associated liability of the person responsible for the (j-1)th processing step in the (j-1)th processing step due to the failure of the j-th processing step in the i-th processing flow. M represents the associated liability risk coefficient for the person responsible for the (j+1)th processing step in the i-th processing flow if the j-th processing step is not completed, and M is the total number of processing steps in the i-th processing flow.
5. The method according to claim 1, characterized in that, The process of obtaining the nonlinear time-varying risk value of each processing step in each processing flow of the business processing system based on the risk assessment model includes: The nonlinear time-varying risk value of each processing step in each processing flow of the business processing system is obtained using the following formula: ； in, Let be the nonlinear time-varying risk value of the nth processing step in the i-th processing flow at time t. This represents the risk normalization coefficient corresponding to the i-th type of processing procedure within the current data observation period. The associated liability risk coefficient is the risk factor for the responsible persons of each processing step in the i-th type of processing flow if the nth processing step is not completed within the current data observation period. The comprehensive risk value generated by the i-th type of processing flow at time t is... Determined by the following formula: ； in, The comprehensive risk liability value generated by the i-th type of processing flow at time t is determined based on the associated liability risk coefficient and the liability risk base. The associated liability risk coefficient for the person responsible for the l-th processing step in the i-th processing flow if the j-th processing step is not completed. M represents the base risk of liability for the person responsible for the j-th processing step if the j-th processing step is not completed at time t, and M represents the total number of processing steps in the i-th type of processing flow.
6. The method according to claim 1, characterized in that, The determination of the dynamic comprehensive risk value of the business processing system based on various nonlinear time-varying risk values includes: The dynamic comprehensive risk value of the business processing system is determined by the following formula: ； in, This represents the dynamic comprehensive risk value of the business processing system. To handle the type risk coefficient vector, The comprehensive risk liability value generated by the i-th type of processing flow at time t is determined based on the associated liability risk coefficient and the liability risk base, where i is an integer greater than or equal to 1 and less than or equal to N, and N is the total number of processing flow types. The risk normalization coefficient corresponding to the i-th type of processing flow is... The associated liability risk coefficient for the person responsible for the l-th processing step in the i-th processing flow if the j-th processing step is not completed. M represents the base risk of liability for the person responsible for the j-th processing step if the j-th processing step is not completed at time t, and M represents the total number of processing steps in the i-th type of processing flow.
7. A nonlinear risk assessment device, characterized in that, The device includes: The parameter generation unit is used to generate parameters for the risk assessment model based on each processing flow and processing steps in the business processing system. The parameters of the risk assessment model include the risk normalization coefficient corresponding to each processing flow, the associated liability risk coefficient of any processing step in the processing flow that is not completed, and the liability risk base of the incomplete processing step to its own responsible person. The base amount of liability risk arising from the incomplete processing steps to the responsible party is generated in the following way: ； in, Let this be the base amount of liability risk for the person responsible for the j-th processing step if the j-th processing step is not completed at time t. Let k0, k1, a, and b be the processing time limit for the j-th processing step, and k0, k1, a, and b be adjustable parameters; when at time t, and t < t / 2, ... Once the j-th processing step is completed, that is, if the j-th processing step is completed ahead of schedule, the sum of all risks and liabilities arising from the j-th processing step and related risks and liabilities will be zero. The risk monitoring unit is used to obtain the nonlinear time-varying risk value of each processing step of each processing flow in the business processing system based on the risk assessment model within a preset data observation period. The risk assessment unit is used to determine the dynamic comprehensive risk value of the business processing system based on various nonlinear time-varying risk values. The risk warning unit is used to generate system risk warning information when the dynamic comprehensive risk value of the business processing system exceeds the system risk warning threshold. The parameter update unit is used to determine the dynamic comprehensive risk value of the business processing system as the fitness function of the genetic algorithm after the data observation period ends, and to determine each risk normalization coefficient and each associated liability risk coefficient as the genetic variation population optimization variable. With the goal of globally minimizing the fitness function, intelligent optimization iteration is performed based on a preset intelligent optimization algorithm; When the minimum value of the fitness function is found, the coefficient values of the treatment type risk coefficient vector and the associated liability risk coefficient matrix in the genetic variation population are extracted, and the parameters of the risk assessment model are updated to obtain the updated risk assessment model; the treatment type risk coefficient vector includes the risk normalization coefficient corresponding to each type; In the next data observation period, the nonlinear time-varying risk value of each processing step of each processing flow in the business processing system is obtained based on the updated risk assessment model.