A multi-level hybrid risk management method, device and electronic equipment

Abstract

The present invention discloses a multi-level hybrid risk management method, including: constructing a first-level risk assessment model, the first-level risk assessment model is composed of a first risk assessment layer and a second risk assessment layer, the first risk assessment layer includes two first-level risk assessment links; obtaining initial user information, substituting it into the first-level risk assessment model to obtain the first-level risk estimate of the initial user; performing a first screening of the initial users according to the first-level risk estimate to form mid-term users; constructing a second-level risk assessment model, performing a second-level risk assessment on mid-term users who meet the second-level risk assessment conditions to obtain a second-level risk estimate, wherein the second-level risk assessment model includes three second-level risk assessment links divided according to user types; performing risk processing on users according to the second-level risk estimate. The present invention adopts different processing measures at different stages before and during the loan to achieve accurate and gentle post-loan management warnings, and ultimately optimizes the overall risk level and expands the business benefits of promoting post-loan conversion.

Description

Technical Field

[0001] The present invention relates to the field of computer information processing, and in particular to a multi-level hybrid risk management method, device, electronic equipment and computer-readable medium. Background Art

[0002] As an emerging industry, Internet finance attracts investors with its lucrative profits and huge potential. In recent years, with the entry of various capitals, the entire Internet finance industry has been full of vitality, and new Internet finance platforms and organizations have emerged continuously. However, the industry's rapid "growth" has also brought about problems such as "P2P platform explosion", "malicious loan fraud", and "violent debt collection". How to protect the legitimate rights and interests of lenders within the scope of the law and eliminate the phenomenon of "malicious loan fraud" has become one of the hot topics in public opinion.

[0003] In order to avoid violent debt collection and the inability of borrowers to recover their funds, a more stringent and cautious risk control plan is needed. Pre-screening and judging of users who are difficult to recover and prevention and control at the entrance and in the process can effectively reduce risks and thus achieve the problem of low overall user recovery rate. In the prior art, the A-card scoring model is usually used to assess the risk of applicants, but the A-card scoring model is relatively limited to the assessment of users' financial performance, such as fraud, and often does not take into account the performance of users in post-loan management. Summary of the invention

[0004] The purpose of the present invention is to provide a multi-level hybrid risk management method, device and electronic equipment, which adopts a layered hybrid post-loan management risk warning scheme by drawing on the circuit to separate and process the current to achieve the goal of different stage functions, forming a complete post-loan management warning system, and adopting different processing measures at different stages before and during the loan to achieve accurate and gentle post-loan management warnings, ultimately optimizing the overall risk level and expanding the business benefits of promoting post-loan conversion.

[0005] Other features and advantages of the present disclosure will become apparent from the following detailed description, or may be learned in part by the practice of the present disclosure.

[0006] To achieve the above object, one aspect of the present invention provides a multi-level hybrid risk management method, comprising:

[0007] Constructing a primary risk assessment model, wherein the primary risk assessment model is composed of a first risk assessment layer and a second risk assessment layer, wherein the first risk assessment layer includes two primary risk assessment links;

[0008] Acquire initial user information, and substitute it into the first-level risk assessment model to obtain the first-level risk estimate of the initial user;

[0009] Perform a first screening of the initial users according to the first-level risk estimation to form mid-term users;

[0010] Constructing a secondary risk assessment model, and conducting secondary risk assessment on mid-term users who meet the secondary risk assessment conditions to obtain a secondary risk estimate, wherein the secondary risk assessment model includes three secondary risk assessment links divided according to user types;

[0011] The user is risk-handled based on the secondary risk estimate.

[0012] According to a preferred embodiment of the present invention, constructing the two primary risk assessment links further includes: constructing a user pre-encouragement model link based on full-process performance and constructing a stage pre-encouragement model link based on staged performance.

[0013] According to a preferred embodiment of the present invention, constructing a user collection pre-model link further includes: obtaining historical user feature data, the feature data including time feature data and post-loan performance data in the post-loan management team; based on the historical user feature data and post-loan performance data, using a machine learning method to train and generate a user post-loan management pre-model to predict the post-loan performance value of a new user; setting a first risk threshold; and providing feedback warnings to users whose post-loan performance values ​​predicted by the user post-loan management pre-model exceed the first risk threshold.

[0014] According to a preferred embodiment of the present invention, constructing the stage collection pre-model link further includes: obtaining historical user characteristic data, the characteristic data including IVR stage characteristic data, manual stage characteristic data and post-loan performance data; based on the historical user characteristic data, using a machine learning method to train and generate an IVR post-loan management pre-model and a manual post-loan management pre-model to predict the IVR post-loan performance value and the manual post-loan performance value of the information user; setting an IVR risk threshold and a manual risk threshold respectively; and providing feedback warnings to users whose IVR post-loan performance value and manual post-loan performance value exceed the corresponding thresholds.

[0015] According to a preferred embodiment of the present invention, constructing the second risk assessment layer further includes: constructing a user risk prediction model for predicting user risk values, wherein the user risk prediction model includes an application scorecard model.

[0016] According to a preferred embodiment of the present invention, the step of obtaining initial user information and substituting it into the first-level risk assessment model to obtain the first-level risk estimate value of the initial user further includes: obtaining initial user information; randomly substituting the initial user information into any one of the two first-level risk assessment links in the first risk assessment layer; pushing the information of the user evaluated by the first-level risk assessment link to the second risk assessment layer to obtain the first-level risk estimate value of the initial user.

[0017] According to a preferred embodiment of the present invention, the step of constructing the secondary risk assessment link further includes: obtaining historical mid-term user feature data, the feature data including attribute features, behavioral features, financial features and recovery performance features; using a machine learning method to train and generate a mid-loan recovery rate model to predict the recovery rate of mid-term users who meet the secondary risk assessment conditions; setting a mid-loan recovery rate risk threshold; and providing feedback warnings to users whose mid-loan recovery rates exceed the mid-loan recovery rate risk threshold.

[0018] According to a preferred embodiment of the present invention, it further includes training different loan recovery rate models according to user types.

[0019] According to a preferred implementation manner of the present invention, the user types include new users, second-new users, and new and dormant users.

[0020] According to a preferred embodiment of the present invention, the step of performing a secondary risk assessment on mid-term users who meet the secondary risk assessment conditions to obtain a secondary risk estimate value further includes: obtaining mid-term user information that meets the secondary risk assessment conditions; substituting the mid-term user information into the secondary risk assessment link corresponding to the user type based on the user type; and obtaining the secondary risk estimate value of the user based on the evaluation of the secondary branch line assessment link.

[0021] According to a preferred embodiment of the present invention, the secondary risk assessment condition includes the user's spending behavior.

[0022] According to a preferred embodiment of the present invention, the machine learning method includes an xgboost algorithm.

[0023] The second aspect of the present invention provides a multi-level mixed risk management device, characterized in that it includes: a primary risk assessment model construction module, which is used to construct a primary risk assessment model, and the primary risk assessment model construction module includes a first and a second risk assessment layer construction unit, which is used to construct a first risk assessment layer and a second risk assessment layer constituting the primary risk assessment model, wherein the first risk assessment layer construction unit includes a first and a second primary risk assessment link construction component, which is used to construct a first and a second primary risk assessment link constituting the first risk assessment layer; an initial user information acquisition module, which is used to acquire initial user information; a user first screening module, which is used to substitute the initial user information into the primary risk assessment model to obtain the primary risk estimate of the initial user, and perform a first screening on the initial user according to the first-stage risk estimate; a secondary risk assessment model construction module, which is used to construct a secondary risk assessment model, and perform a secondary risk assessment on a mid-term user who meets the secondary risk assessment conditions to obtain a secondary risk estimate, wherein the secondary risk assessment model construction module includes a first, a second and a third secondary risk assessment link construction component, which is used to construct the first, the second and the third secondary risk assessment links constituting the second risk assessment layer; a risk processing module, which is used to perform risk processing on the user according to the secondary risk estimate.

[0024] According to a preferred embodiment of the present invention, the first-level risk assessment link construction component is used to construct a user pre-encouragement model link based on the full-process performance; the second-level risk assessment link construction component is used to construct a stage pre-encouragement model link based on the staged performance.

[0025] According to a preferred embodiment of the present invention, the first-level risk assessment link construction component includes: a historical user feature data component, which is used to obtain historical user feature data, and the feature data includes time feature data and post-loan performance data in the post-loan management team; a user post-loan management pre-model training component, which is used to train and generate a user post-loan management pre-model based on the historical user feature data and post-loan performance data using a machine learning method to predict the post-loan performance value of a new user; a first risk feedback warning component, which is used to set a first risk threshold, and provide feedback and warning to users whose post-loan performance value predicted by the user post-loan management pre-model exceeds the first risk threshold.

[0026] According to a preferred embodiment of the present invention, the second-level risk assessment link construction component includes: a historical user feature data acquisition component, used to acquire historical user feature data, the feature data includes IVR stage feature data, manual stage feature data and post-loan performance data; a stage reminder pre-model training component, used to train and generate an IVR post-loan management pre-model and a manual post-loan management pre-model based on the historical user feature data using a machine learning method to predict the IVR post-loan performance value and the manual post-loan performance value of the information user; a second risk feedback warning component, used to set the IVR risk threshold and the manual risk threshold, and provide feedback and warning to users whose post-loan performance value predicted by the user's post-loan management pre-model exceeds the first risk threshold.

[0027] According to a preferred embodiment of the present invention, the second risk assessment layer construction unit further comprises: constructing a user risk prediction model for predicting user risk values, wherein the user risk prediction model comprises an application score card model.

[0028] According to a preferred embodiment of the present invention, it further includes: an initial user information acquisition module acquires initial user information; randomly substitutes the initial user information into any one of the two first-level risk assessment links in the first risk assessment layer; and pushes the information of the user evaluated by the first-level risk assessment link to the second risk assessment layer to obtain the initial user's first-level risk estimate.

[0029] According to a preferred embodiment of the present invention, the secondary risk assessment link construction component further includes: a historical mid-term user feature data acquisition component, used to obtain historical mid-term user feature data, the feature data including attribute features, behavioral features, financial features and recovery performance features; a mid-loan recovery rate model construction component, used to train and generate a mid-loan recovery rate model using a machine learning method to predict the recovery rate of mid-term users who meet the secondary risk assessment conditions; a secondary risk assessment component, used to set a mid-loan recovery rate risk threshold, and provide feedback warning to users whose mid-loan recovery rates exceed the mid-loan recovery rate risk threshold.

[0030] According to a preferred embodiment of the present invention, the secondary risk assessment link constituent component trains different loan recovery rate models according to user types to construct the first, second and third secondary risk assessment links.

[0031] According to a preferred implementation manner of the present invention, the user types include new users, second-new users, and new and dormant users.

[0032] According to a preferred embodiment of the present invention, it further includes: obtaining mid-term user information that meets the secondary risk assessment conditions; substituting the mid-term user information into the secondary risk assessment link corresponding to the user type based on the user type; and obtaining the secondary risk estimate of the user based on the evaluation of the secondary branch line assessment link.

[0033] According to a preferred embodiment of the present invention, the secondary risk assessment condition includes the user's spending behavior.

[0034] According to a preferred embodiment of the present invention, the machine learning method includes an xgboost algorithm.

[0035] A third aspect of the present invention provides an electronic device, wherein the electronic device comprises:

[0036] A processor; and a memory storing computer executable instructions, wherein when the executable instructions are executed, the processor executes the multi-level hybrid risk management method.

[0037] A fourth aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores one or more programs, and when the one or more programs are executed by a processor, the above-mentioned multi-level hybrid risk management method is implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to make the technical problems solved by the present invention, the technical means adopted and the technical effects achieved more clearly, the specific embodiments of the present invention will be described in detail with reference to the accompanying drawings. However, it should be noted that the drawings described below are only drawings of exemplary embodiments of the present invention. For those skilled in the art, drawings of other embodiments can be obtained based on these drawings without creative work.

[0039] Figure 1 It is a main flow chart of a multi-level hybrid risk management method according to an exemplary embodiment.

[0040] Figure 2 It is a main flow chart of a method for constructing a primary risk assessment model according to an exemplary embodiment.

[0041] Figure 3 It is a main flow chart of a method for constructing a second-level risk link according to an exemplary embodiment.

[0042] Figure 4 The present invention is a specific example flow chart showing a multi-level hybrid risk management method according to an exemplary embodiment.

[0043] Figure 5It is a block diagram of a risk management device of a multi-level hybrid connection according to an exemplary embodiment.

[0044] Figure 6 It is a block diagram of a primary risk assessment model construction module according to an exemplary embodiment.

[0045] Figure 7 It is a block diagram of a first risk assessment layer construction unit according to an exemplary embodiment.

[0046] Figure 8 The figure is a model block diagram of a multi-stage hybrid risk management device according to an exemplary embodiment.

[0047] Fig. 9 is a structural block diagram of an exemplary embodiment of an electronic device according to the present invention.

[0048] Fig.10 It is a block diagram of a computer-readable medium according to an exemplary embodiment. DETAILED DESCRIPTION

[0049] Exemplary embodiments of the present invention will now be described more fully with reference to the accompanying drawings. However, exemplary embodiments can be implemented in a variety of forms, and should not be construed as limiting the present invention to the embodiments set forth herein. On the contrary, providing these exemplary embodiments enables the present invention to be more comprehensive and complete, and is more convenient for fully conveying the inventive concept to those skilled in the art. The same reference numerals in the figures represent the same or similar elements, components or parts, and thus their repeated description will be omitted.

[0050] Under the premise of being consistent with the technical concept of the present invention, the features, structures, characteristics or other details described in a specific embodiment do not exclude that they can be combined in one or more other embodiments in a suitable manner.

[0051] In the description of specific embodiments, the features, structures, characteristics or other details described in the present invention are intended to enable those skilled in the art to fully understand the embodiments. However, it does not exclude that those skilled in the art can practice the technical solutions of the present invention without one or more of the specific features, structures, characteristics or other details.

[0052] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0053] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0054] It should be understood that although the terms "first", "second", "third", etc. may be used herein to describe various devices, elements, components or parts, this should not be limited by these terms. These terms are used to distinguish one from another. For example, a first device may also be called a second device without departing from the essential technical solution of the present invention.

[0055] The term "and / or" or "and / or" includes any one and all combinations of one or more of the associated listed items.

[0056] Figure 1 The main flow chart of a multi-stage hybrid risk management method according to an exemplary embodiment is shown. A multi-stage hybrid risk management method at least includes steps S101 to S106.

[0057] In step S101, a first-level risk assessment model is constructed.

[0058] Construct a first-level risk assessment model to conduct first-level risk assessment on users. Figure 2 As shown, more specifically, constructing a first-level risk assessment model includes step S201 constructing a first risk assessment layer and step S202 constructing a second risk assessment layer. The execution of the first risk assessment layer and the second risk assessment layer are connected in series, so as to first evaluate the user using the first risk assessment layer and then using the second risk assessment layer.

[0059] The step of constructing the first risk assessment layer in step S201 further includes steps S301 of constructing a first level risk assessment link and S302 of constructing a second level risk assessment link.

[0060] In one embodiment, step S301, the construction of the first level risk assessment link is to construct a user reminder pre-model link based on the whole process performance. The specific steps may include:

[0061] Acquire historical user characteristic data, the characteristic data including time characteristic data in the post-loan management team and post-loan performance data;

[0062] Based on the historical user feature data and post-loan performance data, a machine learning method is used to train and generate a user post-loan management pre-model to predict the post-loan performance value of new users;

[0063] Setting a first risk threshold;

[0064] Feedback warning is provided to users whose post-loan performance values ​​predicted by the user post-loan management ex ante model exceed the first risk threshold.

[0065] In one embodiment, step S302, the second level risk assessment link is constructed by constructing a stage entry pre-model link based on stage performance. The specific steps may include:

[0066] Acquire historical user characteristic data, including IVR stage characteristic data, manual stage characteristic data and post-loan performance data;

[0067] Based on the historical user feature data, a machine learning method is used to train and generate an IVR post-loan management pre-model and an artificial post-loan management pre-model to predict the IVR post-loan performance value and the artificial post-loan performance value of the information user;

[0068] Set IVR risk threshold and manual risk threshold separately;

[0069] Feedback warning is provided to users whose IVR post-loan performance value and manual post-loan performance value exceed corresponding thresholds.

[0070] The execution of the above-mentioned IVR post-loan management pre-model and the manual post-loan management pre-model are connected in series, and are used to evaluate the IVR post-loan management pre-model and the manual post-loan management pre-model on users.

[0071] The execution connection mode of the above-mentioned first level risk link and the second level risk link are connected in a parallel manner, and the user randomly enters one of the two links to complete the first level risk link evaluation.

[0072] In step S202, constructing the second risk assessment layer may specifically include constructing a user risk prediction model for predicting user risk values, wherein the user risk prediction model includes an application scorecard model.

[0073] In step S102, information is obtained and a first-level evaluation is performed.

[0074] The initial user information is obtained and substituted into the first-level risk assessment model. More specifically, the first risk assessment layer is evaluated, and two first-level risk assessment links are randomly entered. Users who pass the evaluation of any first-level risk assessment link enter the second risk assessment layer to obtain the first-level risk assessment value of the initial user.

[0075] In step 103, mid-term users are formed.

[0076] Based on the primary risk value of the initial users, mid-term users are screened.

[0077] In one embodiment, a certain credit line is granted to the user based on the primary risk value of the initial user. The initial user who obtains the credit line is a mid-term user.

[0078] In step S104, a secondary risk assessment model is constructed.

[0079] A secondary risk assessment model is constructed to conduct secondary risk assessment on mid-term users who meet the secondary risk assessment conditions to obtain secondary risk estimates. The construction of the secondary risk assessment model includes constructing three secondary risk assessment links divided by user type.

[0080] In one embodiment, the step of constructing the secondary risk assessment link further includes:

[0081] Obtain historical mid-term user characteristic data, the characteristic data including attribute characteristics, behavioral characteristics, financial characteristics and recycling performance characteristics;

[0082] The machine learning method is used to train and generate a loan recovery rate model to predict the recovery rate of mid-term users who meet the secondary risk assessment conditions;

[0083] Set the risk threshold of loan recovery rate;

[0084] Feedback warning is provided to users whose loan recovery rate exceeds the loan recovery rate risk threshold.

[0085] Furthermore, the three secondary risk assessment links are divided based on user types, and the user types may specifically include new users, relatively new users, and new and dormant users.

[0086] In step S105, information is obtained and a secondary evaluation is performed.

[0087] Acquire mid-term user information that meets the secondary risk assessment conditions; substitute the mid-term user information into the secondary risk assessment link corresponding to the user type based on the user type; and obtain the secondary risk estimate of the user based on the assessment of the secondary branch assessment link.

[0088] Furthermore, mid-term users who meet the secondary risk assessment criteria include users who have made expenditures.

[0089] In step S106, risk management.

[0090] Prompt feedback is provided to users with higher risk values ​​in any risk assessment in steps S101-S105 to facilitate pre-loan or mid-loan management and reduce overall risk.

[0091] In addition, the machine learning algorithms used in model training are all commonly used algorithms in the field, and may specifically include the use of the xgboost algorithm, which will not be described in detail here.

[0092] Example:

[0093] Figure 4 This is a specific example flow chart of a multi-level hybrid risk management method based on the present invention.

[0094] like Figure 4 As described above, a first risk assessment model N1 and a second risk assessment model N2 are constructed.

[0095] In step S401, the initial user submits an application and obtains the user's application information;

[0096] Randomly send the user's application information to the first risk assessment link-step S404 or the second risk assessment link S402-S403 for risk assessment;

[0097] For the user who has passed any risk assessment in the above two links, risk prediction of the risk prediction model in step S405 is performed to obtain a first risk estimation value of the user.

[0098] The user's credit limit is determined based on the user's first risk estimate.

[0099] In step S406, the credit user makes a withdrawal, and based on the account type of the user, the user making the withdrawal is sent to the loan recovery rate model (step S408-1, step S408-2, step S4-8-3) for recovery rate evaluation.

[0100] Based on the risk assessment result in any of the above steps, risk processing is performed on the user S409.

[0101] By using the method provided by the present invention, the post-loan management warning model is deployed in two stages, namely, the pre-loan and mid-loan stages, and a hybrid warning disposal scheme of the staged model and the full-process model in parallel is adopted for the pre-loan stage. For the mid-loan stage, since the users have entered the actual borrowing stage, a warning scheme of first classifying and then disposing is carried out for these users. There are three classifications in total, each corresponding to a model for disposal and warning. Finally, the pre-loan and mid-loan systems are formed into a series system, and the warning schemes in the pre-loan and mid-loan processes are hybrid, and finally a hybrid user post-loan management warning system is formed. The system is layered, graded, and multi-stage, and a comprehensive judgment is made on the users, realizing a collection warning scheme for a user in the life cycle of pre-loan and mid-loan, which can effectively reduce the proportion of users who are difficult to collect entering the actual borrowing stage, and achieve the goal of reducing risks.

[0102] Those skilled in the art will appreciate that all or part of the steps to implement the above embodiments are implemented as a program (computer program) executed by a computer data processing device. When the computer program is executed, the above method provided by the present invention can be implemented. Moreover, the computer program can be stored in a computer-readable storage medium, which can be a readable storage medium such as a disk, an optical disk, a ROM, a RAM, or a storage array composed of multiple storage media, such as a disk or a tape storage array. The storage medium is not limited to centralized storage, and it can also be distributed storage, such as cloud storage based on cloud computing.

[0103] The following describes an apparatus embodiment of the present invention, which can be used to execute a method embodiment of the present invention. The details described in the apparatus embodiment of the present invention should be regarded as supplementary to the above method embodiment; details not disclosed in the apparatus embodiment of the present invention can be implemented with reference to the above method embodiment.

[0104] Figure 5 It is a block diagram of a risk management device of a multi-level hybrid connection according to an exemplary embodiment.

[0105] like Figure 5-7 As shown, the multi-level mixed risk management device 500 may include: a first-level risk assessment model construction module 501, an initial user information acquisition module 502, a user first screening module 503, a second-level risk assessment model construction module 504, and a risk processing module 505.

[0106] The first-level risk assessment model construction module 501 is used to construct a first-level risk assessment model. The first-level risk assessment model construction module includes first and second risk assessment layer construction units 601 and 602, which are used to construct the first risk assessment layer and the second risk assessment layer constituting the first-level risk assessment model 501, wherein the first risk assessment layer construction unit 601 includes first and second first-level risk assessment link construction components 701 and 702, which are used to construct the first and second first-level risk assessment links constituting the first risk assessment layer.

[0107] In one embodiment, the first level risk assessment link building component 701 is used to build a user entry pre-model link based on the full process performance.

[0108] The first level risk assessment link construction component 701 includes: a historical user feature data component, which is used to obtain historical user feature data, and the feature data includes time feature data and post-loan performance data in the post-loan management team; a user post-loan management pre-model training component, which is used to train and generate a user post-loan management pre-model based on the historical user feature data and post-loan performance data using a machine learning method to predict the post-loan performance value of a new user; a first risk feedback warning component, which is used to set a first risk threshold, and provide feedback and warning to users whose post-loan performance value predicted by the user post-loan management pre-model exceeds the first risk threshold.

[0109] The second first-level risk assessment link building component 702 is used to build a stage entry pre-model link based on staged performance.

[0110] The second first-level risk assessment link construction component 702 includes: a historical user feature data acquisition component, which is used to acquire historical user feature data, and the feature data includes IVR stage feature data, manual stage feature data and post-loan performance data; a stage collection pre-model training component, which is used to train and generate an IVR post-loan management pre-model and a manual post-loan management pre-model based on the historical user feature data and adopts a machine learning method to predict the IVR post-loan performance value and the manual post-loan performance value of the information user; a second risk feedback warning component, which is used to set the IVR risk threshold and the manual risk threshold, and to provide feedback and warning to users whose post-loan performance value predicted by the user's post-loan management pre-model exceeds the first risk threshold.

[0111] In one embodiment, the second risk assessment layer construction unit 602 further includes constructing a user risk prediction model for predicting the user risk value, wherein the user risk prediction model includes an application score card model.

[0112] An initial user information acquisition module 502 is used to acquire initial user information;

[0113] The user initial screening module 503 is used to substitute the initial user information into the primary risk assessment model to obtain the primary risk estimate of the initial user, and perform the initial screening of the initial user according to the primary risk estimate;

[0114] The secondary risk assessment model construction module 504 is used to construct a secondary risk assessment model, and conduct secondary risk assessment on mid-term users who meet the secondary risk assessment conditions to obtain a secondary risk estimate, wherein the secondary risk assessment model construction module includes first, second and third secondary risk assessment link construction components, which are used to construct the first, second and third secondary risk assessment links that constitute the second risk assessment layer.

[0115] The secondary risk assessment link construction component further includes: a historical mid-term user feature data acquisition component, which is used to obtain historical mid-term user feature data, and the feature data includes attribute features, behavioral features, financial features and recovery performance features; a mid-loan recovery rate model construction component, which is used to train and generate a mid-loan recovery rate model using a machine learning method to predict the recovery rate of mid-term users who meet the secondary risk assessment conditions; a secondary risk assessment component, which is used to set a mid-loan recovery rate risk threshold, and provide feedback and warning to users whose mid-loan recovery rates exceed the mid-loan recovery rate risk threshold.

[0116] Furthermore, the secondary risk assessment link constituent components train different loan recovery rate models according to user types to construct first, second and third secondary risk assessment links, and the user types include new users, second-new users and new and dormant users.

[0117] Furthermore, the secondary risk assessment conditions include the user's spending behavior, and the machine learning method includes the xgboost algorithm.

[0118] The risk processing module 505 is used to perform risk processing on the user according to the secondary risk estimation value.

[0119] Example:

[0120] Figure 8 The figure is a model block diagram of a multi-stage hybrid risk management device according to an exemplary embodiment.

[0121] like Figure 8 As shown, the model block diagram of the multi-level hybrid risk management device may specifically include a user post-loan management ex-model 801, an IVR post-loan management ex-model 802, a manual post-loan management ex-model 803, a risk prediction model 804, a new user loan recovery rate model 806, a second-new user loan recovery rate model 807, a new and dormant account loan recovery rate model 808 and a decision system 810.

[0122] Among them, the user post-loan management pre-model 801, the IVR post-loan management pre-model 802, the manual post-loan management pre-model 803, and the risk prediction model 804 constitute a first-level risk assessment model.

[0123] Furthermore, the user post-loan management pre-model 801, the IVR post-loan management pre-model 802, and the manual post-loan management pre-model 803 constitute the first risk assessment layer, the risk prediction model 804 is the second risk assessment layer, and the first risk assessment layer and the second risk assessment layer are connected in series.

[0124] Furthermore, the user post-loan management pre-model 801 is the first level risk assessment link, and the IVR post-loan management pre-model 802 and the manual post-loan management pre-model 803 are connected in series to form the second level risk assessment link. The first level risk assessment link and the second level risk assessment link are connected in parallel to form the first risk assessment layer.

[0125] The new user loan recovery rate model 806, the second-new user loan recovery rate model 807, and the new dormant account loan recovery rate model 808 form a second risk assessment model in parallel and are connected in series with the first risk assessment model, and are used for post-loan management risk prediction during and before the loan, respectively.

[0126] Furthermore, each model may send an evaluation result to the decision system 810, and the decision system 810 determines and implements risk treatment measures based on the evaluation result.

[0127] Those skilled in the art will appreciate that the modules in the above device embodiments may be distributed in the device as described, or may be changed accordingly and distributed in one or more devices different from the above embodiments. The modules in the above embodiments may be combined into one module, or may be further split into multiple submodules.

[0128] The following describes an electronic device embodiment of the present invention, which can be regarded as a specific physical implementation of the method and device embodiments of the present invention. The details described in the electronic device embodiment of the present invention should be regarded as a supplement to the above method or device embodiments; details not disclosed in the electronic device embodiment of the present invention can be implemented with reference to the above method or device embodiments.

[0129] Fig. 9 1 is a block diagram of an exemplary embodiment of an electronic device according to the present invention. Fig. 9 The electronic device 200 according to this embodiment of the present invention is described. Fig. 9 The electronic device 200 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0130] like Fig. 9 As shown, the electronic device 200 is in the form of a general computing device. The components of the electronic device 200 may include, but are not limited to: at least one processing unit 210, at least one storage unit 220, a bus 230 connecting different system components (including the storage unit 220 and the processing unit 210), a display unit 240, etc.

[0131] The storage unit stores program codes, which can be executed by the processing unit 210, so that the processing unit 210 executes the steps according to various exemplary embodiments of the present invention described in the electronic prescription circulation processing method section of this specification. For example, the processing unit 210 can execute the following steps: Figure 1 Steps shown.

[0132] The storage unit 220 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 2201 and / or a cache memory unit 2202 , and may further include a read-only memory unit (ROM) 2203 .

[0133] The storage unit 220 may also include a program / utility 2204 having a set (at least one) of program modules 2205, such program modules 2205 including but not limited to: an operating system, one or more application programs, other program modules and program data, each of which or some combination may include the implementation of a network environment.

[0134] Bus 230 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0135] The electronic device 200 may also communicate with one or more external devices 300 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 200, and / or communicate with any device that enables the electronic device 200 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 250. Furthermore, the electronic device 200 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 260. The network adapter 260 may communicate with other modules of the electronic device 200 via the bus 230. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 200, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0136] Through the description of the above implementation mode, it is easy for those skilled in the art to understand that the exemplary embodiments described in the present invention can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the implementation mode of the present invention can be embodied in the form of a software product, which can be stored in a computer-readable storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, or a network device, etc.) to execute the above method according to the present invention. When the computer program is executed by a data processing device, the computer-readable medium can implement the above method of the present invention.

[0137] The computer program may be stored on one or more computer readable media. The computer readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination thereof. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0138] The computer readable storage medium may include a data signal propagated in a baseband or as part of a carrier wave, wherein a readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The readable storage medium may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by an instruction execution system, an apparatus, or a device or used in combination with it. The program code contained on the readable storage medium may be transmitted with any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the above.

[0139] Program code for performing the operations of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., through the Internet using an Internet service provider).

[0140] In summary, the present invention can be implemented in hardware, or in a software module running on one or more processors, or in a combination thereof. It should be understood by those skilled in the art that general data processing devices such as microprocessors or digital signal processors (DSPs) can be used in practice to implement some or all of the functions of some or all of the components in the embodiments of the present invention. The present invention can also be implemented as a device or apparatus program (e.g., a computer program and a computer program product) for executing part or all of the methods described herein. Such a program implementing the present invention can be stored on a computer-readable medium, or can have the form of one or more signals. Such a signal can be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0141] The specific embodiments described above further describe the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the present invention is not inherently related to any specific computer, virtual device or electronic device, and various general devices can also implement the present invention. The above description is only a specific embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A multi-level hybrid risk management method, characterized in that: include: Constructing a primary risk assessment model for performing primary risk assessment on a user, wherein the primary risk assessment model includes a first risk assessment layer and a second risk assessment layer, and the first risk assessment layer and the second risk assessment layer are connected in series to assess the user; The two first-level risk assessment links included in the construction of the first risk assessment layer include: constructing a user pre-collection model link based on the full-process performance and constructing a stage pre-collection model link based on the staged performance, and the execution of the user pre-collection model link and the stage pre-collection model link are connected in parallel to predict the user; wherein the user post-loan management pre-model trained and generated in the user pre-collection model link based on the full-process performance is used to predict the post-loan performance value of the new user; wherein the IVR post-loan management pre-model and the manual post-loan management pre-model trained and generated in the stage pre-collection model link based on the staged performance are connected in series to predict the IVR post-loan performance value and the manual post-loan performance value of the information user; Constructing the second risk assessment layer includes: constructing a user risk prediction model for predicting user risk values; Obtaining initial user information, and substituting it into the first-level risk assessment model to obtain the first-level risk estimate of the initial user, including: randomly substituting the initial user information into any one of the two first-level risk assessment links in the first risk assessment layer; pushing the information of the user evaluated by any one of the first-level risk assessment links to the second risk assessment layer to obtain the first-level risk estimate corresponding to the initial user information; Determine the credit limit of the initial user corresponding to the initial user information based on the primary risk estimation value and conduct a first screening of the initial user to form a mid-term user; A secondary risk assessment model is constructed to conduct secondary risk assessment on mid-term users who meet the secondary risk assessment conditions to obtain a secondary risk estimate, wherein the secondary risk assessment model includes three secondary risk assessment links divided according to user types; wherein the user types include new users, second-new users, and new dormant users, and a loan recovery rate model in the secondary risk assessment link is constructed to predict the recovery rate of mid-term users who meet the secondary risk assessment conditions; Obtaining mid-term user information of mid-term users who meet the secondary risk assessment conditions, substituting the mid-term user information of the mid-term users into the loan recovery rate model corresponding to the secondary risk assessment link corresponding to the user type to evaluate the recovery rate based on the user type, and obtaining the secondary risk estimate of the mid-term users according to the evaluation of the secondary risk assessment link, so as to form a classified and handled early warning hybrid structure and make a comprehensive judgment on the users; wherein, mid-term users who meet the secondary risk assessment conditions include mid-term users who have made expenditures; The user is risk-handled based on the secondary risk estimate.

2. The method according to claim 1, characterized in that The pre-model link for building user reminders includes: Acquire historical user characteristic data, the characteristic data including time characteristic data in the post-loan management team and post-loan performance data; Based on the historical user feature data and post-loan performance data, a machine learning method is used to train and generate a user post-loan management pre-model; Setting a first risk threshold; Feedback warning is provided to users whose post-loan performance values ​​predicted by the user post-loan management ex ante model exceed the first risk threshold.

3. The method according to claim 1, characterized in that The construction of the pre-entry model link in the stage includes: Acquire historical user characteristic data, including IVR stage characteristic data, manual stage characteristic data and post-loan performance data; Based on the historical user feature data, a machine learning method is used to train and generate an IVR post-loan management pre-model and an artificial post-loan management pre-model; Set IVR risk threshold and manual risk threshold separately; Feedback warning is provided to users whose IVR post-loan performance value and manual post-loan performance value exceed the corresponding risk threshold.

4. The method according to claim 1, characterized in that: The user risk prediction model includes an application scorecard model.

5. The method according to claim 1, characterized in that Constructing the secondary risk assessment link includes: Obtain historical mid-term user characteristic data, the characteristic data including attribute characteristics, behavioral characteristics, financial characteristics and recycling performance characteristics; Use machine learning methods to train and generate different loan recovery rate models according to user types; Set the risk threshold of loan recovery rate; Feedback warning is provided to users whose loan recovery rate exceeds the loan recovery rate risk threshold.

6. The method according to claim 2, 3 or 5, characterized in that: The machine learning methods include the xgboost algorithm.

7. A multi-stage hybrid risk management device, characterized in that: include: A first-level risk assessment model construction module is used to construct a first-level risk assessment model for performing a first-level risk assessment on a user. The first-level risk assessment model construction module includes first and second risk assessment layer construction units, which are used to construct a first risk assessment layer and a second risk assessment layer that constitute the first-level risk assessment model. The first risk assessment layer and the second risk assessment layer are executed in a serial manner to evaluate the user; wherein the first risk assessment layer construction unit includes first and second first-level risk assessment link construction components, the first first-level risk assessment link construction component is used to construct a user entry pre-model link based on the full-process performance, and the second first-level risk assessment link construction component is used to construct a user entry pre-model link based on the staged table. The present stage entry reminder pre-model link, and the execution of the user entry reminder pre-model link and the stage entry reminder pre-model link are connected in parallel to predict the user; wherein the user post-loan management pre-model trained and generated in the user entry reminder pre-model link based on the full process performance is used to predict the post-loan performance value of the new user; wherein the IVR post-loan management pre-model and the manual post-loan management pre-model trained and generated in the stage entry reminder pre-model link based on the staged performance are connected in series to predict the IVR post-loan performance value and the manual post-loan performance value of the information user; constructing the second risk assessment layer includes: constructing a user risk prediction model to predict the user risk value; An initial user information acquisition module is used to acquire initial user information; A user first screening module is used to randomly substitute the initial user information into any one of the first-level risk assessment links of the user pre-entry model link and the stage pre-entry model link in the first risk assessment layer, and to push the information of the user evaluated by any one of the first-level risk assessment links to the second risk assessment layer to obtain the first-level risk estimate corresponding to the initial user information through the first-level risk assessment model, and to determine the credit limit of the initial user corresponding to the initial user information according to the first-level risk estimate and to perform a first screening on the initial user to form a mid-term user; A secondary risk assessment model construction module is used to conduct secondary risk assessment on mid-term users who meet the secondary risk assessment conditions to obtain a secondary risk estimate, including first, second and third secondary risk assessment link construction components, which are used to construct the first, second and third secondary risk assessment links that constitute the second risk assessment layer according to user types; wherein user types include new users, second-new users and new dormant users, and construct a loan recovery rate model in the secondary risk assessment link to predict the recovery rate of mid-term users who meet the secondary risk assessment conditions, specifically including: obtaining mid-term user information of mid-term users who meet the secondary risk assessment conditions, substituting the mid-term user information of the mid-term users into the loan recovery rate model corresponding to the secondary risk assessment link corresponding to the user type based on the user type to evaluate the recovery rate, obtaining the secondary risk estimate of the mid-term users according to the evaluation of the secondary risk assessment link, so as to form a warning hybrid structure for classified disposal and make a comprehensive judgment on the users; wherein mid-term users who meet the secondary risk assessment conditions include mid-term users who have made expenditures; The risk processing module is used to perform risk processing on the user according to the secondary risk estimate.

8. The device according to claim 7, characterized in that The first level risk assessment link building components include: The historical user characteristic data component is used to obtain historical user characteristic data, wherein the characteristic data includes characteristic data of the duration in the post-loan management team and post-loan performance data; A user post-loan management pre-model training component is used to train and generate a user post-loan management pre-model based on the historical user feature data and post-loan performance data using a machine learning method; The first risk feedback warning component is used to set a first risk threshold and provide feedback warning to users whose post-loan performance values ​​predicted by the user post-loan management ex ante model exceed the first risk threshold.

9. The device according to claim 7, characterized in that The construction of the pre-entry model link in the stage includes: A historical user characteristic data acquisition component is used to acquire historical user characteristic data, wherein the characteristic data includes IVR stage characteristic data, manual stage characteristic data and post-loan performance data; A stage collection pre-model training component is used to train and generate an IVR post-loan management pre-model and a manual post-loan management pre-model based on the historical user feature data using a machine learning method; The second risk feedback warning component is used to set the IVR risk threshold and the manual risk threshold respectively, and to provide feedback warning to the users whose IVR post-loan performance value and the manual post-loan performance value exceed the corresponding risk threshold.

10. According to the device of claim 7, the user risk prediction model includes an application score card model.

11. The device according to claim 7, characterized in that The secondary risk assessment link building component includes: A historical mid-term user characteristic data acquisition component is used to acquire historical mid-term user characteristic data, wherein the characteristic data includes attribute characteristics, behavior characteristics, financial characteristics and recycling performance characteristics; The component for building a loan recovery rate model is used to train and generate different loan recovery rate models according to user types using a machine learning method; The secondary risk assessment component is used to set a risk threshold for the mid-loan recovery rate and to provide feedback and warning to users whose mid-loan recovery rate exceeds the risk threshold for the mid-loan recovery rate.

12. The device according to claim 8, 9 or 11, characterized in that The machine learning methods include the xgboost algorithm.

13. An electronic device, wherein: The electronic device includes: processor; and, A memory storing computer executable instructions which, when executed, cause the processor to perform the method according to any one of claims 1-6.

14. A computer-readable storage medium, wherein: The computer-readable storage medium stores one or more programs, and when the one or more programs are executed by a processor, the method of any one of claims 1 to 6 is implemented.

Images