Intelligent case division system and method for personal loan unfavorable asset cases

Abstract

The invention provides an intelligent case division system and method for personal loan unfavorable asset cases. The system comprises a system initialization and module loading module, a case receiving and preprocessing module, a case division mode selection and execution stage module, an automatic outsourcing and distribution module, an AB test stage and strategy optimization module and a monitoring and strategy iteration module. Through integration of a multi-mode case allocation engine, a dynamic strategy platform and a collection management module, efficient case allocation and intelligent strategy optimization are realized. The system supports manual / automatic / outsourcing hybrid divisions, a preset strategy library is built in to reduce the cold start cost, the strategy iteration efficiency is improved through Java hot upload and AB test technologies, and the system is suitable for large-scale non-performing asset disposal scenes.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of financial technology and asset management, and particularly relates to an intelligent case distribution system and method for individual loan non-performing asset cases. BACKGROUND

[0002] At present, the existing intelligent case distribution system for individual loan non-performing asset cases has problems such as unsmooth manual / automatic case distribution switching, imperfect automatic outsourcing and distribution functions, and limited AB test functions. SUMMARY

[0003] In view of this, the present application aims to provide an intelligent case distribution system and method for individual loan non-performing asset cases, which realizes efficient case distribution and intelligent strategy optimization through the integration of a multi-mode case distribution engine, a dynamic strategy platform and a collection management module. The system supports mixed manual / automatic / outsourcing case distribution, reduces the cold start cost through the built-in preset strategy library, and improves the strategy iteration efficiency through the Java hot upload and AB test technology, and is suitable for large-scale non-performing asset disposal scenarios.

[0004] In a first aspect, embodiments of the present application provide a personal loan non-performing asset case intelligent distribution system, which comprises: a system initialization and module loading module, a case receiving and preprocessing module, a distribution mode selection and execution stage module, an automatic outsourcing and distribution module, an AB test stage and strategy optimization module, a monitoring and strategy iteration module; the system initialization and module loading module is used to load the core module through a specific startup program when the system starts; according to the preset configuration file or user input, the system parameters are initialized and set; the related information of the personal loan non-performing asset case is read from the specified data source; the case receiving and preprocessing module is used to detect whether there is new case information input through real-time monitoring of the data interface in a polling or event-driven manner; if there is new case information input, the data is preprocessed by using data cleaning, conversion and standardization technology; the distribution mode selection and execution stage module is used to determine the distribution mode by analyzing the current business scenario and matching the preset rules according to the multiple preset rules and business scenario models built in the system; after determining the distribution mode, the asset characteristics of the case are analyzed in depth by using data mining and machine learning algorithms, and the optimal disposal method is automatically matched by combining the strategy rules, the distribution result is generated and displayed to the user through the visual interface; the automatic outsourcing and distribution module is used to trigger the automatic outsourcing process when the case meets the preset automatic outsourcing rules; according to the characteristics of the case and the requirements of the collection strategy, the matching algorithm is used to select the outsourcing agency from the cooperative outsourcing agency library; the related information of the outsourcing case is distributed to the selected outsourcing agency through the data transmission interface; the AB test stage and strategy optimization module is used to set the AB test parameters through the visual strategy configuration interface; according to the AB test parameters; the case samples are divided into different groups, and different strategies are applied for disposal; during the test process, the disposal result data of each group of cases is collected in real time; after the test is completed, the data is analyzed in depth by using statistical analysis method, the disposal effect difference of different strategy groups is compared, and the collection strategy is optimized and adjusted according to the analysis result; the monitoring and strategy iteration module is used to set the early warning index to monitor the collection effect after distribution in real time; when the index is abnormal, the early warning information is sent out; the feedback information of each aspect is collected, the current strategy is comprehensively evaluated by using the effect evaluation model combined with the collection effect data; the strategy is iteratively optimized according to the evaluation result.

[0005] In a second aspect, the embodiments of the present application also provide a personal loan non-performing asset case intelligent distribution method, which comprises: system initialization and module loading, in which the core module is loaded by a specific starting program when the system starts; system parameters are initialized and set according to a preset configuration file or user input; relevant information of the personal loan non-performing asset case is read from a specified data source; a case receiving and preprocessing module detects whether there is new case information input through a real-time monitoring data interface by using a polling or event-driven mode; if there is new case information input, data preprocessing is performed by using data cleaning, conversion and standardization technologies; a distribution mode selection and execution stage module determines the distribution mode to be used by analyzing the current business scenario and matching the preset rules; after the distribution mode is determined, the asset characteristics of the case are analyzed in depth by using data mining and machine learning algorithms, and the optimal disposal mode is automatically matched by combining the strategy rules, and the distribution result is generated and displayed to the user through a visual interface; an automatic outsourcing and distribution module triggers the automatic outsourcing process when the case meets the preset automatic outsourcing rules; the outsourcing agency is selected from the cooperative outsourcing agency library according to the characteristics of the case and the requirements of the collection strategy; the relevant information of the outsourcing case is distributed to the selected outsourcing agency through a data transmission interface; an AB test stage and strategy optimization module sets the AB test parameters through a visual strategy configuration interface; the case samples are divided into different groups according to the AB test parameters, and different strategies are applied for disposal; in the test process, the disposal result data of each group of cases are collected in real time; after the test is completed, the data are analyzed in depth by using statistical analysis methods, the disposal effect differences of different strategy groups are compared, and the collection strategy is optimized and adjusted according to the analysis result; a monitoring and strategy iteration module sets early warning indicators to monitor the collection effect after the distribution in real time; when the indicators are abnormal, early warning information is sent; feedback information from all aspects is collected, the current strategy is comprehensively evaluated by using an effect evaluation model in combination with the collection effect data; and the strategy is iteratively optimized according to the evaluation result.

[0006] The embodiments of the present application bring the following beneficial effects: The embodiments of the present application provide a personal loan non-performing asset case intelligent distribution system and method, which realizes efficient distribution of cases and intelligent optimization of strategies by integrating a multi-mode distribution engine, a dynamic strategy platform and a collection management module. The system supports manual / automatic / outsourcing mixed distribution, has a built-in preset strategy library to reduce the cold start cost, and improves the strategy iteration efficiency by using Java hot uploading and AB testing technology, and is suitable for large-scale non-performing asset disposal scenarios.

[0007] Other features and advantages of the present disclosure will be described in the following description, or can be inferred or determined without doubt from the description, or can be known by implementing the above-mentioned technologies of the present disclosure.

[0008] To make the above-mentioned objects, features and advantages of this disclosure more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0009] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0010] Figure 1 This is a schematic diagram of the structure of an intelligent case allocation system for non-performing loan cases provided in an embodiment of the present invention; Figure 2 A flowchart illustrating an intelligent case allocation method for non-performing loan cases provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of an intelligent case allocation method for non-performing loan cases provided by an embodiment of the present invention; Figure 4 This is a schematic diagram of an intelligent case allocation method for non-performing loan cases provided in an embodiment of the present invention. Detailed Implementation

[0011] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0012] Currently, the existing intelligent case assignment system for non-performing personal loan assets mainly suffers from the following technical problems: 1. Inefficient Switching Between Manual and Automatic Case Assignment: In practice, it may be necessary to flexibly switch between manual and automatic case assignment modes depending on the situation. However, some similar products may experience inconvenience and untimely data synchronization during the switching process, affecting case assignment efficiency. For example, when financial institutions need to perform manual case assignment temporarily due to special circumstances, the system may not be able to quickly switch to manual mode, or the relevant data for automatic case assignment may not be updated in a timely manner after the switch, resulting in inconsistent case assignment results.

[0013] 2. Incomplete automatic outsourcing and allocation functions: During the automatic outsourcing process, the qualifications and capabilities of third-party collection agencies may not be accurately assessed, leading to the allocation of cases to unsuitable collection agencies and affecting the collection results. At the same time, there may be deficiencies in monitoring and management during the outsourcing process, which cannot timely grasp the collection progress and problems. For example, the system may not fully consider the professional capabilities, regional coverage, and other factors of the collection agency when outsourcing non-performing asset cases to the collection agency, leading to the allocation of cases to unsuitable collection agencies and affecting the collection results. In addition, there may be a lack of effective monitoring mechanisms during the allocation process, which cannot timely discover and handle problems during the allocation process.

[0014] 3. Limited AB testing function: AB testing is an important means of optimizing case allocation strategies, but the AB testing function of some similar products may not be perfect, such as being unable to test multiple variables simultaneously, having a long testing period, and inaccurate result analysis, which limits the effectiveness of strategy optimization.

[0015] Based on this, the embodiment of the present application provides an intelligent case allocation system and method for individual loan non-performing asset cases, which is suitable for the automatic allocation, strategy management, and collection optimization of non-performing asset cases by banks, asset management companies, and other financial institutions.

[0016] In order to facilitate the understanding of the present embodiment, first, a detailed introduction is made to the individual loan non-performing asset case intelligent allocation system disclosed by the present embodiment.

[0017] Embodiment one: The embodiment of the present application provides an individual loan non-performing asset case intelligent allocation system, referring to the structure schematic diagram of an individual loan non-performing asset case intelligent allocation system shown in Figure 1 The individual loan non-performing asset case intelligent allocation system includes: a system initialization and module loading module, a case receiving and preprocessing module, a case allocation mode selection and execution stage module, an automatic outsourcing and allocation module, an AB testing stage and strategy optimization module, and a monitoring and strategy iteration module.

[0018] I. The system initialization and module loading module is used to load the core module through a specific startup program when the system starts; according to the preset configuration file or user input, the system parameters are initialized and set; the related information of individual loan non-performing asset cases is read from the specified data source.

[0019] The principle of system initialization and module loading in this embodiment can be: when the system starts, load the core modules through a specific startup program, and these core modules contain various functional codes and algorithms required for system operation. At the same time, according to the preset configuration file or user input, the system parameters are initialized and set, such as case filing rules parameters, AB test parameters, etc. Importing case data is to read the relevant information of individual loan non-performing asset cases from a specified data source (such as a database, a file, etc.), providing a data basis for subsequent processing.

[0020] In some embodiments, the system initialization and module loading module includes: starting the system core framework submodule, configuring the system parameter submodule, and connecting the data source submodule; starting the system core framework submodule for running the computer program to start the individual loan non-performing asset case intelligent case filing system and loading the core module; the system parameter submodule is configured to configure the basic parameters and system permissions; the data source submodule is used to import case basic information, import third-party collection agency information, and load other system parameters.

[0021] 1. Start the system core framework submodule: run the computer program, start the individual loan non-performing asset case intelligent case filing system, load the core module, including the case filing mode management module, the collection strategy management module, the strategy platform access module, the visual strategy configuration module, the AB test module, etc. Initialize the database connection, including the case information library, the strategy rule library, the user permission library and the historical case filing result library.

[0022] 2. Configure the system parameter submodule: (1) Configure basic parameters: according to the business needs and regulatory requirements of financial institutions, set the time period of case filing (such as daily, weekly case filing), the number limit of case allocation, the triggering condition of automatic outsourcing (such as overdue for a certain period of time, overdue amount exceeding a certain amount) and other parameters.

[0023] (2) Configure system permissions: configure the permissions of the visual strategy configuration interface, and determine the operation permissions of different user roles (such as business personnel, technical personnel, management personnel) for strategy configuration.

[0024] 3. Connect the data source submodule: establish connection with the individual loan non-performing asset case database, collection agency information database, customer information database and other related data sources to ensure that the system can obtain case information, collection resource information and customer detailed data in real time. Load the following basic data: (1) Import case basic information: read the basic information of individual loan non-performing asset cases from the database of financial institutions, including but not limited to borrower name, contact information, loan amount, overdue period, collateral information, etc. Access individual loan non-performing asset case data from bank core systems, credit management systems, etc. through API or batch import, including: 1) Case Basis Information: Borrower ID, Loan Amount, Overdue Days, Collateral Status, etc.

[0025] 2) Historical Collection Records: Collection Frequency, Response Rate, Repayment Amount, etc.

[0026] 3) External Data: Borrower Credit Information, Social Behavior Data, Judicial Enforcement Information, etc.

[0027] (2) Import Third-Party Collection Agency Information: Import relevant data of third-party collection agencies, such as agency name, qualification certificate, professional field, geographical coverage, historical collection performance, etc.

[0028] (3) Load System Other Parameters: Load system default case allocation rules, collection strategy templates, and initial parameters for AB testing.

[0029] II. Case Reception and Preprocessing Module, used to monitor data interfaces in real time, detect whether there is new case information through polling or event-driven methods; if there is new case information, use data cleaning, conversion and standardization techniques for data preprocessing.

[0030] The principle of the case reception and preprocessing module in this embodiment can be: the system monitors the data interface in real time, detects whether there is new case information through polling or event-driven methods. Once the case data is received, use data cleaning, conversion and standardization techniques for data preprocessing. For example, remove duplicate data, fill in missing values, and unify data formats to ensure the quality and consistency of case data, providing accurate and reliable data for subsequent case allocation processing.

[0031] In some embodiments, the case reception and preprocessing module includes a new case reception sub-module and a case data preprocessing sub-module. The new case reception sub-module is used for real-time data import and initial information verification. The case data preprocessing sub-module is used for data cleaning and standardization and extraction of key case features.

[0032] 1. New Case Reception Sub-Module: The system monitors the data interface in real time and automatically receives case information when a new individual loan non-performing asset case enters, including case number, borrower information, loan amount, overdue period, loan contract details, etc.

[0033] (1) Real-time Data Import: The system monitors the financial institution's business system in real time and automatically captures and imports case information into the intelligent case allocation system when a new individual loan non-performing asset case is generated.

[0034] (2) Preliminary verification of information: The received case information is preliminarily checked for completeness and accuracy. For cases with missing key information or incorrect data format, such as checking whether the borrower's ID number is legal and the loan amount is a positive number, etc. Generate error reports and notify relevant personnel to correct.

[0035] 2. Case data preprocessing submodule: The received case data is cleaned, missing values and outliers are cleaned, data format is unified (such as time standardization, classification label coding), and data verification is performed to ensure the accuracy and completeness of the case information.

[0036] (1) Data cleaning and standardization: 1) Data cleaning: Fill or correct missing values and outliers (such as replacing missing overdue days with industry average).

[0037] 2) Unified data format: Date standardized to YYYY-MM-DD, amount unit converted to yuan.

[0038] 3) Data standardization: Generate case feature vectors (such as overdue stage coding, loan product type one-hot coding).

[0039] (2) Extract key features of the case: 1) Static features: loan type, initial amount, interest rate, borrower age, occupation, credit score, etc.

[0040] 2) Dynamic features: overdue stage, collection history, repayment willingness assessment (such as call recording sentiment analysis).

[0041] 3) External data: court execution information, third-party credit data, economic environment indicators (such as regional unemployment rate).

[0042] Three, case selection and execution phase module, according to the system built-in multiple preset rules and business scene model, through the analysis of the current business scene and the matching of the preset rules, judge the adopted case mode; After determining the case mode, use data mining and machine learning algorithms to analyze the asset characteristics of the case in depth, and combine with the strategy rules to automatically match the optimal disposal mode, generate the case result and display it to the user through the visual interface.

[0043] The principle of the split mode selection and execution stage module in this embodiment can be that the system is built-in with multiple preset rules and business scenario models, and determines which split mode to use through analysis of the current business scenario (such as the number of cases, the distribution of asset types, etc.) and matching of the preset rules. After determining the split mode, the system uses data mining and machine learning algorithms to conduct in-depth analysis on the asset characteristics of the cases (such as overdue duration, amount of arrears, credit status of the borrower, etc.), and automatically matches the optimal disposition method in combination with strategy rules (such as disposition priority corresponding to different asset characteristics, disposition methods, etc.), generates a split result, and displays the result to the user through a visual interface.

[0044] In some embodiments, the split mode selection and execution stage module includes a split mode determination sub-module and an intelligent split execution sub-module. The split mode determination sub-module is configured to determine whether to use manual split or automatic split mode according to preset rules and the current business scenario. The intelligent split execution sub-module is configured to automatically match the optimal disposition method according to asset characteristics and strategy rules.

[0045] 1. Split mode determination sub-module: The system determines whether to use manual split or automatic split mode according to preset rules and the current business scenario. For example, when encountering special cases or situations requiring human intervention, the system prompts the user to perform manual split; in normal circumstances, the system executes according to the automatic split process.

[0046] 2. Intelligent split execution sub-module: The system automatically matches the optimal disposition method (such as self-collection, outsourcing, litigation) according to asset characteristics (such as type, amount, risk level) and strategy rules; through strategy platform access, multiple mature split strategies (such as according to asset type, risk level, regional distribution, etc.) are built-in, and the user can select or customize strategies according to needs.

[0047] (1) Manual split module: Supports business personnel to manually allocate cases through a visual interface, and manually selects case allocation objects (such as collectors, outsourcing agencies) by human intervention. The specific operation of manual split is as follows: 1) User selection: If the user selects the manual split mode, the business personnel logs in to the system and views the list of cases to be allocated and the information of collectible personnel or agencies through a visual interface.

[0048] 2) Manual screening and allocation: The business personnel manually allocates cases to suitable collectors or agencies according to the characteristics of the cases and the situation of the collection resources. During the allocation process, the system displays the allocation results and related statistical information in real time, such as the number of cases and the distribution of priority levels for each collector or agency.

[0049] 3) Case status adjustment: After the business staff completes the assignment, the system records and backs up the assignment results and updates the case status to "assigned".

[0050] (2) Automatic allocation module: When the user selects the automatic allocation mode or the system's preset automatic allocation conditions are met (such as reaching a certain number of cases to be allocated), the system starts the automatic allocation process. The system will automatically match allocation rules based on case attributes (loan amount, overdue duration, borrower region, debtor portrait, etc.), and the system will automatically match cases and resources according to the preset strategy. The specific operation of automatic allocation is as follows: 1) Strategy matching: The system matches the most suitable collection strategy in the collection strategy library according to the case classification, priority, borrower credit evaluation report, etc. For example, for high-priority cases with poor borrower credit, a more aggressive collection strategy is selected.

[0051] 2) Resource evaluation: In combination with the requirements of the collection strategy, evaluate the available collection resources, including the workload of internal collection personnel, skill level, and the qualifications, capabilities, and current business volume of third-party collection agencies, etc.

[0052] 3) Intelligent allocation: Use intelligent algorithms (such as rule-based reasoning, machine learning algorithms, etc.) to allocate cases to the most suitable collection personnel or collection agencies. In the allocation process, consider factors such as geographical factors and professional field matching degree to improve collection effectiveness.

[0053] 4) Result recording and updating: The system records the results of automatic allocation in the database and updates the case status to "assigned". At the same time, generate an allocation report detailing the allocation of each case, allocation basis, etc. for subsequent queries and analysis.

[0054] Specifically, to solve the "manual / automatic allocation switching is not smooth" problem in the prior art, the present application proposes an innovative system design scheme, aiming to realize quick and seamless switching between manual and automatic allocation modes, and ensure real-time synchronization of data during the switching process, thereby significantly improving allocation efficiency and accuracy. The specific solution is as follows: 1) Optimize system architecture design: By restructuring the system architecture, introduce an efficient state management mechanism, so that manual and automatic allocation modes can be flexibly switched within the same system framework, without complex interface jumps or step operations, realizing one-key switching, greatly simplifying the operation process.

[0055] 2) Real-time data synchronization technology: Advanced data synchronization algorithms are used to ensure that all relevant case data is immediately updated and synchronized to all modules of the system at the moment of mode switching. Whether switching from manual to automatic or vice versa, data consistency and real-time performance are guaranteed, avoiding case errors or confusion caused by data synchronization.

[0056] 3) Intelligent preloading and caching mechanism: Intelligent preloading and caching strategies are integrated into the system design to preload possible case data, reducing waiting time during switching. At the same time, through reasonable cache management, data is quickly accessed and accurately presented during switching, further improving user experience.

[0057] 4) User-friendly interface design: Focus on the intuitiveness and ease of use of the user interface, with clear instructions and feedback mechanisms to help users easily understand the current case mode and quickly complete mode switching operations. At the same time, provide detailed operation logs and status prompts to enhance user control and trust in system behavior.

[0058] (3) Automatic outsourcing module: Connect with third-party collection agencies, dynamically allocate outsourcing cases based on cost, compliance, historical recovery rate, etc. Automatically trigger external agency allocation processes for cases that meet outsourcing conditions (such as overdue duration, amount threshold).

[0059] (4) Case allocation result generation: The system generates detailed case allocation results for each case, including assigned collection agency or collection personnel information, collection strategy description, expected recovery target, etc. At the same time, update the case status to "allocated" and record the allocation results in the database.

[0060] (5) Case allocation result display: The system provides a case allocation result display interface to visually display each case's allocation in list, chart, etc. Form, including case number, borrower information, assigned collection agency or collection personnel, collection strategy, etc.

[0061] Four, automatic outsourcing and allocation module, used when the case meets the preset automatic outsourcing rules, triggers the automatic outsourcing process; according to the characteristics of the case and the requirements of the collection strategy, use matching algorithms to select the most suitable outsourcing agency from the cooperative outsourcing agency library; through the data transmission interface, the relevant information of the outsourcing case is allocated to the selected outsourcing agency.

[0062] The principle of the automatic outsourcing and distribution module in this embodiment can be: the system pre-sets automatic outsourcing rules, when the case meets these rules (such as overdue time exceeding a certain period, internal collection effect is poor, etc.), triggers the automatic outsourcing process. The system selects the most suitable outsourcing agency from the cooperative outsourcing agency library according to the characteristics of the case (such as the amount of arrears, the region of the borrower, the type of assets, etc.) and the requirements of the collection strategy (such as the specialized field of the outsourcing agency, the charging standard, the historical collection effect, etc.), and uses the matching algorithm. Then, through the data transmission interface, the relevant information of the outsourcing case is distributed to the selected outsourcing agency.

[0063] In some embodiments, the automatic outsourcing and distribution module includes: an outsourcing condition judgment submodule, a collection agency selection submodule, and an outsourcing distribution submodule; the collection agency selection submodule is used to screen eligible agencies, generate a candidate agency list, and intelligently select the most suitable agency; the outsourcing distribution submodule is used for outsourcing distribution and notification, and updating the case status.

[0064] 1. Outsourcing condition judgment submodule: the system monitors the overdue situation and the amount of arrears of the case in real time, and when the case meets the pre-set automatic outsourcing conditions, such as the case overdue time exceeding a certain period, the internal collection effect being poor, etc., it automatically judges which cases need to be handled by outsourcing, and triggers the automatic outsourcing process.

[0065] 2. Collection agency selection submodule: (1) Screening eligible agencies: for cases that need to be outsourced, according to the characteristics of the case (such as region, amount of arrears, type of collateral, etc.) and the requirements of the collection strategy, the system screens eligible collection agencies from the third-party collection agency library according to the qualifications, historical cooperation records, recovery rate, charging standard, etc. of the outsourcing agency.

[0066] (2) Generating a candidate agency list: comprehensively evaluating the screened collection agencies, considering factors such as the qualifications, professional ability, historical collection performance, charging standard, etc. of the agency, generating a candidate collection agency list for each case.

[0067] (3) Intelligent selection of the most suitable agency: using intelligent distribution algorithms to select the most suitable agency from the candidate collection agency list to undertake the case. In the distribution process, factors such as the business load of the agency and the proximity to the borrower's region are considered to ensure the collection effect.

[0068] 3. Outsourcing distribution submodule: (1) Outsourcing distribution and notification: the system distributes the outsourcing case to the selected outsourcing agency, and notifies the outsourcing agency to receive the case information through system interfaces or emails, SMS, etc. At the same time, records the relevant information of the outsourcing distribution, such as the outsourcing time, the name of the outsourcing agency, the outsourcing fee, etc.

[0069] (2) Update the case status: the system automatically generates a outsourcing contract or a commission agreement, and sends the case information and collection requirements to the selected collection agency. At the same time, update the case status to "outsourced", and record the outsourcing time, collection agency information and other related data.

[0070] In some embodiments, the intelligent case distribution system for individual loan non-performing asset cases is constructed with an intelligent third-party collection agency evaluation system and a whole-process outsourcing monitoring and management mechanism.

[0071] Specifically, in view of the fact that the prior art cannot accurately evaluate the qualifications and capabilities of third-party collection agencies in the automatic outsourcing process, and that the cases are allocated to unsuitable collection agencies, which affects the collection effect, and the outsourcing process is not monitored and managed, and the collection progress and problems cannot be grasped in time, the present application builds an intelligent third-party collection agency evaluation system and a whole-process outsourcing monitoring and management mechanism, and the specific innovations are as follows: (1) Third-party collection agency evaluation: 1) Qualification audit module: Collect the basic information of the third-party collection agency, including business license, tax registration certificate, relevant industry qualification certificate, etc., and strictly audit the legality and compliance of the agency. At the same time, verify the personnel scale, office site and other hardware conditions of the agency to ensure that it has the basic ability to carry out collection business.

[0072] 2) Professional ability evaluation module: According to different types of debt cases (such as credit cards, mortgages, car loans, etc.), design corresponding professional ability evaluation indexes. For example, for credit card collection cases, the evaluation indexes can include the familiarity of the collection team with the rules of credit card business, the past success rate of credit card collection, the cooperation experience with card issuing banks, etc. Through comprehensive analysis of these indexes, the professional ability of the collection agency in a specific business field is accurately evaluated.

[0073] 3) Regional coverage evaluation module: Considering the regional distribution of debtors, evaluate the service regional range of the collection agency. By collecting data such as the branch information of the collection agency in different regions, the distribution of business teams and the relationship with local partners, determine the regional range that can be effectively covered by the agency, and ensure that the cases can be allocated to the collection agencies with corresponding regional service capabilities.

[0074] 4) Reputation evaluation module: Establish a credit file of the collection agency, collect information such as its reputation in the industry, past evaluations from entrusting parties, whether there are records of violations, etc. Through analysis and evaluation of these credit data, generate a credit score for each collection agency, which serves as an important reference for case allocation.

[0075] (2) Case reasonable distribution module: 1) According to the evaluation results generated by the third-party collection agency evaluation system, combined with the specific characteristics of the case (such as debt type, debt amount, debtor region, etc.), intelligent algorithms are used to achieve reasonable allocation of cases. For example, for large overdue housing loan cases, preferentially allocate to collection agencies with rich experience and high professional ability in the field of real estate and service network in the debtor's region.

[0076] 2) During the allocation process, the case carrying capacity of each collection agency is fully considered to avoid over-allocation leading to a decline in collection quality. At the same time, a dynamic adjustment mechanism is established to flexibly adjust the allocation of subsequent cases according to the real-time work progress and performance of the collection agency, ensuring optimal allocation of resources.

[0077] (3) Outsourcing process monitoring and management mechanism: 1) Real-time data acquisition module: Through system integration with collection agencies or requiring them to report data regularly, real-time acquisition of key information during the collection process is achieved, including communication records with debtors, repayment negotiation, collection action plans and execution, etc. At the same time, technical means (such as voice recognition, text analysis, etc.) are used to automatically analyze and organize voice calls and text information during the collection process, extracting valuable information.

[0078] 2) Progress monitoring module: Set key nodes and expected time progress for collection cases, and monitor the collection progress of each case in real time. By comparing with the preset progress, timely find out the cases with progress lag, and issue warning information. At the same time, statistical analysis is made on the overall collection progress to provide decision support for management.

[0079] 3) Problem handling module: Establish a problem feedback and handling mechanism. When problems such as irregular operations, negative work or disputes with debtors are found in the monitoring process, the problems are fed back to the responsible person in time, and the progress of problem handling is tracked. For serious violations, take appropriate measures such as suspension of case allocation, deduction of commission, etc. to ensure the standardization and compliance of outsourcing process.

[0080] 4) Performance evaluation module: According to the case allocation quantity, collection success rate, repayment amount, customer satisfaction of the collection agency, etc., a scientific and reasonable performance evaluation system is established. The performance of the collection agency is evaluated and ranked regularly to provide basis for subsequent case allocation and cooperation decision-making. At the same time, the performance evaluation results are linked with the commission settlement of the collection agency to encourage them to improve the efficiency and quality of collection.

[0081] V. AB test stage and strategy optimization module, used for setting AB test parameters through a visual strategy configuration interface; dividing case samples into different groups according to AB test parameters; applying different strategies to dispose of the case samples respectively; collecting disposal result data of the case samples in each group in real time during the test; and performing in-depth analysis on the data after the test, comparing disposal effect differences of different strategy groups, and optimizing and adjusting the collection strategy according to the analysis result.

[0082] The principle of the AB test stage and strategy optimization module in the embodiment can be that a user sets AB test parameters through a visual strategy configuration interface, including test variables (such as collection scripts, collection frequencies, etc.), variable value ranges, test sample grouping manners, test periods, etc. The system divides case samples into different groups according to the settings, and applies different strategies to dispose of the case samples respectively. The system collects disposal result data (such as recovery amounts, recovery rates, customer complaint rates, etc.) of the case samples in each group in real time during the test. After the test, statistical analysis methods are used to perform in-depth analysis on the data, compare disposal effect differences of different strategy groups, and optimize and adjust the collection strategy according to the analysis result.

[0083] In some embodiments, the AB test stage and strategy optimization module includes an AB test setting submodule, a test execution submodule, and a result analysis and strategy optimization submodule. The AB test setting submodule is used to create test groups and control groups and an AB test engine. The test execution submodule is used to execute collection strategies and collect processing data. The result analysis and strategy optimization submodule is used for result analysis and strategy optimization. The AB test stage and strategy optimization module supports testing multiple variables at the same time.

[0084] 1. AB test setting submodule: (1) Create test groups and control groups: business personnel or technical personnel set AB test parameters and grouping rules through a visual strategy configuration interface. For example, divide the collection strategies to be tested into group A and group B, and each group uses different strategy parameters or collection manners. In the case division process, the system uses the AB test module to randomly divide cases into test groups and control groups. The test groups apply new collection strategies, and the control groups apply original collection strategies. The sample size and test period to be determined are the focus, to ensure that the test result has statistical significance.

[0085] (2) AB test engine: run multiple groups of strategies in parallel, compare disposal effects through recovery rates, costs, and other indicators, and automatically evaluate the optimal solution. 2. Test execution submodule: (1) Execute collection strategies: the system randomly allocates cases to group A and group B according to the AB test settings, and applies corresponding collection strategies for collection respectively.

[0086] (2) Collecting and processing data: During the test process, the system collects case handling data of the test group and the control group in real time, including recovery rate, collection cost, customer complaint rate and other indicators.

[0087] 3. Result analysis and strategy optimization submodule: (1) Result analysis: After the test period ends, the system analyzes the results of AB testing, compares the differences in various indicators between the two groups of cases using statistical analysis methods, and determines which collection strategy is more effective.

[0088] (2) Strategy optimization: Based on the test results, the collection strategy is optimized and adjusted. If the B group strategy performs better than the A group, the system applies some or all parameters of the B group strategy to the overall collection strategy and deploys the new strategy to the system through dynamic loading function.

[0089] Specifically, due to the imperfect AB testing function of some similar products, there are problems such as inability to test multiple variables simultaneously, long test period, inaccurate result analysis, etc., which limit the effect of strategy optimization. The present application supports simultaneous testing of multiple variables, shortens the test period, and improves the accuracy of result analysis, thereby helping business personnel to develop and adjust disposal strategies more scientifically and improve the efficiency of non-performing asset disposal and recovery rate.

[0090] (1) Simultaneous testing of multiple variables: The present system supports simultaneous testing of multiple variables. In the non-performing asset disposal strategy, users can combine multiple key variables (such as collection method, repayment plan, communication frequency, etc.) into different strategy versions for simultaneous testing according to actual needs. In this way, the influence of multiple variables and their interaction on the effect of the strategy can be comprehensively and efficiently evaluated, greatly shortening the time and number of tests required for comprehensive evaluation of the strategy.

[0091] (2) Shorten the test period: Advanced sample allocation algorithms and efficient data collection and analysis techniques are used to significantly shorten the test period. In terms of sample allocation, the system intelligently allocates samples to different strategy version groups according to pre-set rules and targets, ensuring that each group of samples is representative and balanced. At the same time, using real-time data acquisition and rapid analysis algorithms, data can be acquired and analyzed in real time during the test process, and it can be quickly judged whether the differences between different strategy versions are significant, so as to end the ineffective test in advance, saving time and resources.

[0092] (3) Improve the accuracy of result analysis: use various data analysis methods and models to conduct in-depth and accurate analysis of test results. The system can not only handle common effectiveness evaluation indicators such as recovery rate and conversion rate, but also accurately quantify and analyze complex indicators such as long-term recovery rate and customer satisfaction in the field of non-performing asset disposal. By eliminating sample bias and reducing data noise and other interference factors, it ensures that the test results can truly reflect the differences between different strategy versions and provide reliable decision-making basis for strategy optimization.

[0093] Six, monitoring and strategy iteration module, for setting early warning indicators to monitor the collection effect after the division; when the index is abnormal, send an early warning message; collect feedback information from all aspects, combine the collection effect data, and use the effect evaluation model to comprehensively evaluate the current strategy; according to the evaluation result, the strategy is iteratively optimized.

[0094] The principle of the monitoring and strategy iteration module in this embodiment can be: the system sets a series of early warning indicators (such as recovery rate below the threshold value, customer complaint rate exceeding the threshold value, etc.), and monitors the collection effect after the division in real time. When the index is abnormal, the system immediately sends an early warning message. At the same time, the system collects feedback information from all aspects (such as feedback from external agencies, customer feedback, etc.), combines the collection effect data, and uses the effect evaluation model to comprehensively evaluate the current strategy. According to the evaluation result, the system automatically or prompts the user to iteratively optimize the strategy, forming a continuous improvement cycle.

[0095] In some embodiments, the monitoring and strategy iteration module includes: a real-time monitoring submodule, a feedback and adjustment submodule; the real-time monitoring submodule is used for visual display, monitoring effect, and early warning disposal; the feedback and adjustment submodule is used for effect evaluation and strategy iteration, subsequent tracking and feedback.

[0096] 1. Real-time monitoring submodule: (1) Visual display: the system monitors the case division process and collection progress in real time, and displays key indicators such as case allocation rate, recovery rate, collection progress, collection cost, strategy hit rate, and customer complaint rate through a visual dashboard. Generate a division effect board through the visualization module to display: 1) Strategy recovery rate comparison: recovery rate comparison curve of different strategy groups.

[0097] 2) Performance ranking: performance ranking of collectors / external agencies (such as recovery amount and case handling timeliness).

[0098] 3) AB test effect: AB test result heat map (such as optimal strategy distribution under different overdue stages).

[0099] 4) Report generation and statistics: According to user needs, the system generates various case allocation reports, such as daily case allocation statistics, collection agency workload reports, case recovery rate reports, etc. Reports can be exported to Excel, PDF, etc. Format, convenient for users to analyze data and make decisions.

[0100] (2) Monitoring effect: Monitor the collection effect after case allocation (such as 24-hour response rate, 7-day recovery rate): 1) AB test effect analysis: Through AB test, compare the recovery rate, cost and other indicators of different strategies. If the recovery rate of a certain strategy group continues to be lower than the benchmark value, trigger strategy rollback or retraining, and automatically optimize strategy parameters (such as adjusting outsourcing threshold, collection frequency).

[0101] 2) Outsourcing agency adjustment: If the recovery rate of the outsourcing agency decreases, automatically reduce its case allocation and mark it as "need to re-evaluate".

[0102] (3) Early warning and disposal: Daily / weekly case allocation effect report (recovery rate, cost, team productivity) is generated, and early warning mechanism is set. When the key indicators exceed the preset threshold, the system sends early warning information in time to inform relevant personnel to handle.

[0103] 2, Feedback and adjustment sub-module: (1) Effect evaluation and strategy iteration: The system analyzes and organizes feedback information, adjusts and optimizes case allocation strategies and collection strategies according to the analysis results to improve non-performing asset disposal efficiency and recovery rate. At the same time, the adjusted strategy and feedback processing result are fed back to the relevant personnel to form a closed-loop management.

[0104] (2) Subsequent tracking and feedback: The system tracks the collection progress of the case in real time and collects feedback information in the collection process, such as the borrower's repayment willingness and problems encountered in the collection process. According to the feedback information, the system can adjust the collection strategy or reassign the case to improve the efficiency of non-performing asset disposal and recovery rate.

[0105] In summary, the embodiment of the present application provides an individual loan non-performing asset case intelligent allocation system, which realizes efficient case allocation and intelligent optimization of strategies through the integration of multi-mode allocation engine, dynamic strategy platform and collection management module. The system supports manual / automatic / outsourcing mixed allocation, has a built-in preset strategy library to reduce cold start cost, and improves strategy iteration efficiency through Java hot upload and AB test technology, and is suitable for large-scale non-performing asset disposal scenarios.

[0106] Compared with the prior art, the individual loan non-performing asset case intelligent allocation system provided in the embodiment mainly solves the following technical problems: I. Solve the problem of smooth switching between manual / automatic allocation: The deficiencies of the prior art: In the actual operation of the prior art for the same kind of product, when manually and automatically switching the case allocation mode, there are inconvenient operation, data synchronization is not timely, and the case allocation efficiency is affected, such as the system cannot quickly switch the mode when the financial institution temporarily manually allocates cases, or the automatic case allocation data is not updated in time after switching, resulting in chaotic case allocation results.

[0107] Solution: The present application realizes the quick and seamless switching of manual / automatic case allocation mode by optimizing system design, and ensures the real-time synchronization of data during switching process, effectively improving the case allocation efficiency.

[0108] II. Improve the automatic outsourcing and allocation functions: (1) Deficiency of the prior art: In view of the prior art, in the automatic outsourcing process, the quality and ability of the third-party collection agency cannot be accurately evaluated, and the case is allocated to an unsuitable collection agency, which affects the collection effect, and the outsourcing process monitoring and management are insufficient, and the collection progress and problems cannot be grasped in time, such as the system does not fully consider the allocation of cases according to the professional ability, regional coverage and other factors of the collection agency, and the allocation process lacks effective monitoring mechanism.

[0109] (2) Solution: The present application constructs a perfect third-party collection agency evaluation system, which can accurately evaluate the quality and ability of the collection agency, and realize the reasonable allocation of cases; at the same time, a comprehensive outsourcing process monitoring and management mechanism is established, the collection progress is grasped in real time, the problems are handled in time, and the collection effect is improved.

[0110] III. Enhance the AB testing function: (1) Deficiency of the prior art: Due to the imperfect AB testing function of some similar products, there are problems such as unable to test multiple variables at the same time, long test period, inaccurate result analysis, etc., which limits the optimization effect of strategy.

[0111] (2) Solution: The present application enhances the AB testing function, supports testing multiple variables at the same time, shortens the test period, and improves the accuracy of result analysis, thereby providing stronger support for the optimization of case allocation strategy, and improving the efficiency and recovery rate of non-performing asset disposal.

[0112] Compared with the prior art, the individual loan non-performing asset case intelligent allocation system provided in the embodiment mainly has the following advantages: (1) Comprehensive intelligent case allocation processing: The prior art may rely on manual experience or simple rules when allocating individual loan non-performing asset cases, and lacks intelligent analysis of the overall characteristics of the case. The system of the present application can comprehensively consider asset characteristics and various strategy rules, automatically match the optimal disposal mode, improve the accuracy and scientificity of case allocation, and make each case more suitable for processing, thereby improving the overall disposal efficiency.

[0113] (2) Efficient automatic outsourcing mechanism: the traditional way may be arbitrary or lack of pertinence in the selection of outsourcing agencies, resulting in poor outsourcing effect. When the case triggers the automatic outsourcing rule, the present application can accurately select the most suitable outsourcing agency according to the characteristics of the case and the requirements of the collection strategy, ensure that the outsourcing case is handled professionally and effectively, and improve the success rate of outsourcing collection.

[0114] (3) Flexible and accurate AB testing and strategy optimization: the AB testing function of existing similar products is imperfect, and there are problems such as inability to test multiple variables at the same time, long test period, inaccurate result analysis, etc., which limit the effect of strategy optimization. The present application flexibly configures AB testing parameters through a visual strategy configuration interface, supports testing multiple variables at the same time, shortens the test period, and can optimize and adjust the collection strategy in time according to accurate analysis results, so that the strategy always adapts to the changing market environment and case characteristics, improving the efficiency and recovery rate of non-performing asset disposal.

[0115] Embodiment two, The present embodiment provides an intelligent case allocation method for individual loan non-performing asset cases, which is implemented on the basis of the above-mentioned embodiment and applied to the intelligent case allocation system for individual loan non-performing asset cases provided by the above-mentioned embodiment. Referring to the flowchart of an intelligent case allocation method for individual loan non-performing asset cases shown in Figure 2 The intelligent case allocation method for individual loan non-performing asset cases comprises the following steps: Step S202, system initialization and module loading: the core modules are loaded by a specific startup program when the system starts; the system parameters are initialized and set according to the preset configuration file or user input; and the relevant information of individual loan non-performing asset cases is read from the specified data source.

[0116] Referring to the schematic diagram of an intelligent case allocation method for individual loan non-performing asset cases shown in Figure 3 System initialization and module loading can be performed: 1. Start the intelligent case allocation system for individual loan non-performing asset cases, and load the core modules such as the data interface module, the case allocation algorithm module, the AB testing module, etc.

[0117] 2. Configure system parameters, such as setting the priority rules of case allocation, the sample allocation ratio of AB testing, etc.

[0118] Step S204, case receiving and preprocessing module: through real-time monitoring of the data interface, the polling or event-driven method is used to detect whether there is new case information input; if there is new case information input, the data is preprocessed by using data cleaning, conversion and standardization techniques.

[0119] As shown in Figure 3 Case receiving and preprocessing can be performed: 1. The system monitors the data interface in real time. When new case information is transmitted, it is immediately received and stored in the temporary data buffer.

[0120] 2. The received case data is pre-processed, including data cleaning (removing invalid characters, duplicate data, etc.), data conversion (unifying different formats of data into system standard format), and data standardization (normalizing numerical data, etc.).

[0121] Step S206, the case distribution mode selection and execution module determines the distribution mode based on the built-in preset rules and business scenario models in the system, analyzes the current business scenario, and matches the preset rules to determine the distribution mode. After determining the distribution mode, the data mining and machine learning algorithm is used to analyze the asset characteristics of the case in depth, and the optimal disposal method is automatically matched based on the strategy rules, and the distribution result is generated and displayed to the user through the visual interface.

[0122] As shown in Figure 3 , the case distribution mode selection and execution can be performed: 1. The system determines the distribution mode based on the preset rules and the current business scenario (such as the current case backlog, the proportion of different types of assets, etc.), such as distribution by asset type, distribution by overdue period, etc.

[0123] 2. Analyze the asset characteristics of the case using the distribution algorithm, and automatically match the optimal disposal method based on the strategy rules, such as legal litigation disposal for high-risk large overdue cases, and telephone collection disposal for low-risk small overdue cases.

[0124] 3. Generate the distribution result, including the disposal method, disposal personnel or institution information for each case, and display it to the user through the visual interface, and store the distribution result in the database.

[0125] Step S208, the automatic outsourcing and distribution module triggers the automatic outsourcing process when the case meets the preset automatic outsourcing rules; based on the characteristics of the case and the requirements of the collection strategy, it uses the matching algorithm to select the most suitable outsourcing institution from the cooperative outsourcing institution library; through the data transmission interface, the relevant information of the outsourcing case is distributed to the selected outsourcing institution.

[0126] As shown in Figure 3 , automatic outsourcing and distribution can be performed: 1. The system monitors the case status in real time. When the case triggers the automatic outsourcing rule (such as overdue for more than 90 days and no progress in internal collection), it enters the automatic outsourcing process.

[0127] 2. According to the characteristics of the case (such as the amount of arrears between 1-5 million yuan, borrowers in certain areas, etc.) and the requirements of the collection strategy (such as the desire for external agencies to have expertise in financial dispute mediation), the qualified external agencies are selected from the external agency library.

[0128] 3. The most suitable external agency is selected by using a matching algorithm (such as considering the expertise, fees, and historical performance of the external agency for weighted scoring), and the relevant information of the external case (such as case details, collection requirements, etc.) is assigned to the external agency, and the assignment information is recorded to the database.

[0129] Step S210, the AB test phase and the strategy optimization module set AB test parameters through the visual strategy configuration interface; according to the AB test parameters; the case samples are divided into different groups, and different strategies are applied respectively; during the test process, the disposal result data of each group of cases are collected in real time; after the test is completed, statistical analysis methods are used to analyze the data in depth, compare the disposal effect differences of different strategy groups, and optimize and adjust the collection strategy according to the analysis results.

[0130] As shown in Figure 3 , the AB test phase and the strategy optimization can be carried out: 1. The user sets the AB test parameters through the visual strategy configuration interface, such as setting two collection script variables (mild and tough) for testing, randomly dividing the case samples into two groups, applying different collection scripts respectively, and the test period is 30 days.

[0131] 2. The system executes the strategy according to the settings, and collects the data of the two groups of cases in real time during the test process, such as the recovery amount, recovery rate, and customer complaint rate.

[0132] 3. After the test is completed, the system compares the significance of the differences between the two groups of data by using statistical analysis methods (such as t-test, variance analysis, etc.), determines which collection script is more effective according to the analysis results, optimizes and adjusts the collection strategy, such as using the mild collection script as the recommended script for subsequent collection.

[0133] Step S212, the monitoring and strategy iteration module sets early warning indicators to monitor the collection effect after the case is divided; when the indicators are abnormal, it sends early warning information; collects feedback information from all aspects, combines with the collection effect data, uses the effect evaluation model to comprehensively evaluate the current strategy; according to the evaluation results, the strategy is iteratively optimized.

[0134] As shown in Figure 3 , monitoring and strategy iteration can be carried out: 1. The system sets early warning indicators, such as the recovery rate being lower than 5% for 3 consecutive days, the customer complaint rate being more than 2%, etc., to monitor the collection effect after the case is divided in real time.

[0135] 2、When the early warning index is monitored, the system sends early warning information to relevant personnel through email, short message, etc.

[0136] 3、The system collects the feedback information of the collection progress of the outsourcing agency and the satisfaction of the customers, combines the collection effect data, and uses the effect evaluation model (such as the analytic hierarchy process, fuzzy comprehensive evaluation method, etc.) to evaluate the current strategy.

[0137] 4、According to the evaluation result, the system automatically generates strategy optimization suggestions or prompts the user to adjust the strategy, such as adjusting the case allocation rules, optimizing the collection process, etc., to realize the continuous iteration and optimization of the strategy.

[0138] Also see Figure 4 The schematic diagram of the individual loan non-performing asset case intelligent allocation method is shown in FIG. 1, mainly including: Step 1, system initialization and module loading: starting the individual loan non-performing asset case intelligent allocation system, loading the core module, configuring the system parameters, and importing the case data.

[0139] Step 2, case receiving and preprocessing: the system monitors the data interface to receive case information in real time, and pre-processes the received case data.

[0140] Step 3, allocation mode selection and execution: the system determines the allocation mode according to the preset rules and the current business scenario, and automatically matches the optimal disposal mode according to the asset characteristics and strategy rules to generate and display the allocation result.

[0141] Step 4, automatic outsourcing and distribution: when the case is sent, the automatic outsourcing rule is selected, the most suitable outsourcing agency is selected according to the characteristics of the case and the requirements of the collection strategy, and the case is distributed to the selected outsourcing agency.

[0142] Step 5, AB test phase and strategy optimization: configure the AB test parameters through the visual strategy configuration interface, the system executes the strategy according to the AB test settings, and optimizes and adjusts the collection strategy according to the test results.

[0143] Step 6, according to the test results, the collection strategy is optimized and adjusted, the collection effect after the allocation is monitored and the early warning is set, and the system collects the feedback information, evaluates the effect and iterates the strategy.

[0144] In summary, the embodiment of the present application provides an individual loan non-performing asset case intelligent allocation method, which mainly has the following advantages: (1) Intelligent subcase and automatic matching of optimal disposal method: Unlike traditional manual subcase methods, traditional methods rely on human experience and are prone to inaccurate subcase and unreasonable disposal methods. The intelligent subcase system of the present application automatically matches the optimal disposal method according to preset rules and asset characteristics, improving the accuracy and scientificity of subcase, and enabling each case to find the most suitable processing path, thereby improving case disposal efficiency and recovery rate.

[0145] (2) Real-time monitoring and automatic outsourcing allocation: In existing technologies, case outsourcing allocation may be lagging and lack precise matching of case characteristics. The system of the present application monitors data interfaces to receive case information in real time, and when the automatic outsourcing rule is triggered, it automatically selects the most suitable outsourcing agency for allocation according to case characteristics and collection strategy requirements. This real-time and precise matching improvement enables cases to be timely processed by professionals, improving the effectiveness and efficiency of outsourcing collection.

[0146] (3) AB testing and strategy dynamic optimization mechanism: Traditional collection strategies are often fixed and cannot adapt to changing market environments and borrower behavior. The present application introduces AB testing and strategy optimization mechanisms, which can easily configure test parameters through a visual interface. The system adjusts collection strategies in real time based on test results and continuously monitors effects for iterative optimization. This improvement enables the system to quickly respond to market changes, always maintain optimal collection strategies, and improve the flexibility and effectiveness of collection work.

[0147] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the individual loan non-performing asset case intelligent subcase method described above can refer to the corresponding process in the foregoing embodiments, which will not be repeated here.

[0148] Finally, it should be noted that: the above embodiments are only specific implementations of the present application, used to illustrate the technical solutions of the present application, but not to limit it. The protection scope of the present application is not limited to this. Although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can modify or easily think of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed by the present application, or make equivalent substitutions for some technical features; and these modifications, changes or substitutions do not cause the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. An intelligent case assignment system for non-performing loan cases, characterized in that, The intelligent case allocation system for non-performing loan cases includes: system initialization and module loading module, case receiving and preprocessing module, case allocation mode selection and execution phase module, automatic outsourcing and allocation module, A / B testing phase and strategy optimization module, and monitoring and strategy iteration module. The system initialization and module loading module is used to load the core module through a specific startup program when the system starts; to initialize the system parameters according to the preset configuration file or user input; and to read relevant information on individual loan non-performing asset cases from a specified data source. The case receiving and preprocessing module is used to detect whether new case information is being received by real-time monitoring data interface using polling or event-driven methods; if new case information is received, the data is preprocessed using data cleaning, transformation and standardization techniques. The case allocation mode selection and execution phase module is used to determine the case allocation mode to be adopted based on the system's built-in preset rules and business scenario models, by analyzing the current business scenario and matching the preset rules. After determining the case allocation mode, data mining and machine learning algorithms are used to conduct in-depth analysis of the asset characteristics of the case, and combined with strategy rules, the optimal handling method is automatically matched, the case allocation result is generated, and it is displayed to the user through a visual interface. The automatic outsourcing and allocation module is used to trigger the automatic outsourcing process when a case meets the preset automatic outsourcing rules; based on the characteristics of the case and the requirements of the collection strategy, it uses a matching algorithm to select outsourcing agencies from the pool of cooperating outsourcing agencies; and it allocates the relevant information of the outsourced case to the selected outsourcing agency through the data transmission interface. The A / B testing phase and strategy optimization module is used to set A / B test parameters through a visual strategy configuration interface; divide case samples into different groups according to the A / B test parameters, and apply different strategies to handle them respectively; collect the handling result data of each group of cases in real time during the test; and use statistical analysis methods to conduct in-depth analysis of the data after the test, compare the differences in handling effects of different strategy groups, and optimize and adjust the collection strategy based on the analysis results. The monitoring and strategy iteration module is used to set early warning indicators to monitor the collection effect after case allocation in real time; when abnormal indicators are detected, an early warning message is issued; feedback information from various aspects is collected, and combined with collection effect data, the current strategy is comprehensively evaluated using an effect evaluation model; and the strategy is iteratively optimized based on the evaluation results.

2. The intelligent case assignment system for non-performing loan cases according to claim 1, characterized in that, The system initialization and module loading module includes: a system core framework startup submodule, a system parameter configuration submodule, and a data source connection submodule; The startup system core framework sub-module is used to run a computer program to start the intelligent case allocation system for individual loan non-performing asset cases and load the core module. The configuration system parameter submodule is used to configure basic parameters and system permissions; The data source connection submodule is used to import basic case information, import third-party debt collection agency information, and load other system parameters.

3. The intelligent case assignment system for non-performing loan cases according to claim 1, characterized in that, The case receiving and preprocessing module includes: a new case receiving submodule and a case data preprocessing submodule; The new case receiving submodule is used for real-time data import and preliminary information verification; The case data preprocessing submodule is used for data cleaning and standardization, and extraction of key case features.

4. The intelligent case assignment system for non-performing loan cases according to claim 1, characterized in that, The case allocation mode selection and execution phase module includes: a case allocation mode judgment submodule and an intelligent case allocation execution submodule; The case allocation mode determination submodule is used to determine whether to use manual or automatic case allocation mode based on preset rules and the current business scenario. The intelligent case allocation and execution submodule is used to automatically match the optimal disposal method based on asset characteristics and strategy rules.

5. The intelligent case assignment system for non-performing loan cases according to claim 4, characterized in that, The intelligent case allocation execution submodule includes: a manual case allocation unit, an automatic case allocation unit, an automatic outsourcing unit, a case allocation result generation unit, and a case allocation result display unit; The intelligent case assignment system for non-performing loan cases is used to switch between manual and automatic case assignment modes by optimizing system architecture design, real-time data synchronization technology, intelligent preloading and caching mechanism, and user-friendly interface design.

6. The intelligent case assignment system for non-performing loan cases according to claim 1, characterized in that, The automatic outsourcing and allocation module includes: an outsourcing condition judgment submodule, a collection agency selection submodule, and an outsourcing allocation submodule; The collection agency selection submodule is used to filter eligible agencies, generate a list of candidate agencies, and intelligently select agencies. The outsourcing allocation submodule is used for outsourcing allocation and notification, and updating case status.

7. The intelligent case assignment system for non-performing loan cases according to claim 6, characterized in that, The intelligent case assignment system for non-performing loan cases has an intelligent third-party collection agency evaluation system and a full-process outsourcing monitoring and management mechanism.

8. The intelligent case assignment system for non-performing loan cases according to claim 1, characterized in that, The A / B testing phase and strategy optimization module includes: an A / B test setup submodule, a test execution submodule, and a result analysis and strategy optimization submodule; The AB test setup submodule is used to create test groups and control groups, and the AB test engine; The test execution submodule is used to execute collection strategies and collect and process data; The result analysis and strategy optimization submodule is used for result analysis and strategy optimization. The A / B testing phase and strategy optimization module support testing multiple variables simultaneously.

9. The intelligent case assignment system for non-performing loan cases according to claim 1, characterized in that, The monitoring and strategy iteration module includes: a real-time monitoring submodule and a feedback and adjustment submodule; The real-time monitoring submodule is used for visualization, monitoring results, and early warning handling. The feedback and adjustment submodule is used for effect evaluation and strategy iteration, follow-up tracking and feedback.

10. A method for intelligent case allocation in individual loan non-performing asset cases, characterized in that, The method includes: The system initialization and module loading module loads the core module through a specific startup program when the system starts; it initializes and sets system parameters according to preset configuration files or user input; and it reads relevant information on individual loan non-performing asset cases from a specified data source. The case receiving and preprocessing module detects whether new case information is being received by monitoring the data interface in real time, using polling or event-driven methods. If new case information is received, the data is preprocessed using data cleaning, transformation, and standardization techniques. The case allocation mode selection and execution module determines the appropriate case allocation mode based on the system's built-in preset rules and business scenario models, through analysis of the current business scenario and matching of preset rules. After determining the case allocation mode, it uses data mining and machine learning algorithms to conduct in-depth analysis of the case's asset characteristics, and combines them with strategy rules to automatically match the optimal handling method, generate case allocation results, and display them to the user through a visual interface. When a case meets the preset automatic outsourcing rules, the automatic outsourcing module triggers the automatic outsourcing process; based on the characteristics of the case and the requirements of the collection strategy, it uses a matching algorithm to select outsourcing agencies from the pool of cooperating outsourcing agencies; and it distributes the relevant information of the outsourced case to the selected outsourcing agency through the data transmission interface. The A / B testing phase and strategy optimization module allows users to set A / B test parameters through a visual strategy configuration interface; according to the A / B test parameters, case samples are divided into different groups, and different strategies are applied to each group for handling; during the test, the handling results data of each group of cases are collected in real time; after the test, statistical analysis methods are used to conduct in-depth analysis of the data, compare the differences in handling effects of different strategy groups, and optimize and adjust the collection strategy based on the analysis results. The monitoring and strategy iteration module sets early warning indicators to monitor the collection effect after case allocation in real time; when abnormal indicators are detected, an early warning message is issued; feedback information from all parties is collected, and combined with collection effect data, the current strategy is comprehensively evaluated using an effect evaluation model; the strategy is iteratively optimized based on the evaluation results.

Images