CN113554227A - Training method of user loss rate prediction model and user loss rate prediction method - Google Patents
Training method of user loss rate prediction model and user loss rate prediction method Download PDFInfo
- Publication number
- CN113554227A CN113554227A CN202110833599.7A CN202110833599A CN113554227A CN 113554227 A CN113554227 A CN 113554227A CN 202110833599 A CN202110833599 A CN 202110833599A CN 113554227 A CN113554227 A CN 113554227A
- Authority
- CN
- China
- Prior art keywords
- user
- loss rate
- contact
- rate prediction
- training
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000012549 training Methods 0.000 title claims abstract description 91
- 238000000034 method Methods 0.000 title claims abstract description 67
- 238000010801 machine learning Methods 0.000 claims abstract description 13
- 238000012795 verification Methods 0.000 claims description 21
- 239000013598 vector Substances 0.000 claims description 15
- 230000002159 abnormal effect Effects 0.000 claims description 6
- 238000010219 correlation analysis Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 5
- 238000012545 processing Methods 0.000 claims description 5
- 230000000694 effects Effects 0.000 abstract description 17
- 230000006870 function Effects 0.000 description 10
- 238000010586 diagram Methods 0.000 description 6
- 230000003287 optical effect Effects 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000001133 acceleration Effects 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 230000001186 cumulative effect Effects 0.000 description 2
- 238000005034 decoration Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000011022 operating instruction Methods 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 235000019800 disodium phosphate Nutrition 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/03—Credit; Loans; Processing thereof
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Accounting & Taxation (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Development Economics (AREA)
- Finance (AREA)
- Software Systems (AREA)
- Tourism & Hospitality (AREA)
- Data Mining & Analysis (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Technology Law (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Entrepreneurship & Innovation (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Game Theory and Decision Science (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the application provides a training method of a user loss rate prediction model and a user loss rate prediction method. The training method of the user loss rate prediction model comprises the following steps: determining a target user in a specified service state; determining a training set, wherein the training set comprises a contact label and user characteristics of a target user in a first time period, and the contact label is used for representing the contact collection condition of the target user in the first time period; and training the user loss rate prediction model through a machine learning algorithm according to the training set. The user loss rate prediction model trained based on the scheme can accurately predict the loss rate of overdue users due, and can provide a basis for effectively improving the effect of collection and reducing overdue bad account rate.
Description
Technical Field
The application relates to the technical field of data processing, in particular to a training method of a user loss rate prediction model and a user loss rate prediction method.
Background
In the loan transaction of the financial institution, when the user is overdue, the user needs to be urged to receive overdue debt.
At present, when overdue debt is charged, whether the contact with the overdue user can be successfully established can directly concern the charging effect of the debt. Therefore, if the loss rate of the overdue user can be predicted, a basis can be provided for effectively improving the effect of collection and reducing the overdue bad account rate.
Disclosure of Invention
The present application aims to solve at least one of the above technical drawbacks. The technical scheme adopted by the application is as follows:
in a first aspect, an embodiment of the present application provides a training method for a user loss rate prediction model, where the method includes:
determining a target user in a specified service state;
determining a training set, wherein the training set comprises a contact label and user characteristics of a target user in a first time period, and the contact label is used for representing the contact collection condition of the target user in the first time period;
and training the user loss rate prediction model through a machine learning algorithm according to the training set.
Optionally, determining the target user in the designated service state includes:
and determining the users with the overdue within the target deadline as the target users of the specified business state.
Optionally, the obtaining of the contact tag of the target user's collection contact condition in the first time period includes:
and determining a contact label associated with the prompt receiving contact condition of the target user based on the preset association relation.
Optionally, determining a training set comprises:
constructing a feature vector of each target user based on the contact tag and the user features;
and constructing a training set based on the feature vectors of the target users.
Optionally, before constructing the feature vector of each target user based on the contact tag and the user feature, the method further includes:
and processing the abnormal condition existing in the user characteristic.
Optionally, the obtaining the user characteristics of the target user in the first period includes:
acquiring user related information of a target user in a first period;
extracting initial user characteristics based on the user related information;
determining user characteristics from the initial user characteristics based on the results of the correlation analysis of the initial user characteristics with the contact tags.
Optionally, the method further includes:
determining a verification set, wherein the verification set comprises contact tags of the collection contact condition of the target user and user characteristics in a second time period;
and verifying the user loss rate prediction model based on the verification set.
In a second aspect, an embodiment of the present application provides a method for predicting a user loss rate, where the method includes:
acquiring user characteristics of a user to be predicted;
inputting the user characteristics into a pre-trained user loss rate prediction model to obtain a user loss rate prediction result of the user to be predicted, wherein the user loss rate prediction model is obtained by training according to the training method of the user loss rate prediction model shown in any one of the embodiments of the first aspect.
Optionally, the method further includes:
and determining a hasty collection strategy based on the user loss rate prediction result.
In a third aspect, an embodiment of the present application provides a training apparatus for a user loss rate prediction model, where the apparatus includes:
the target user determining module is used for determining a target user in a specified service state;
the training set determining module is used for determining a training set, wherein the training set comprises a contact label and user characteristics of a target user in a first time period, and the contact label is used for representing the contact collection situation of the target user in the first time period;
and the model training module is used for training the user loss rate prediction model through a machine learning algorithm according to the training set.
Optionally, the target user determination module is specifically configured to:
and determining the users with the overdue within the target deadline as the target users of the specified business state.
Optionally, when the contact tag of the urge to receive contact condition of the target user in the first time period is acquired, the training set determining module is specifically configured to:
and determining a contact label associated with the prompt receiving contact condition of the target user based on the preset association relation.
Optionally, the training set determining module is specifically configured to:
constructing a feature vector of each target user based on the contact tag and the user features;
and constructing a training set based on the feature vectors of the target users.
Optionally, the apparatus further comprises:
and the exception handling module is used for handling the exception condition existing in the user characteristics before the characteristic vector of each target user is constructed based on the contact tag and the user characteristics.
Optionally, when obtaining the user characteristics of the target user in the first period, the training set determining module is specifically configured to:
acquiring user related information of a target user in a first period;
extracting initial user characteristics based on the user related information;
determining user characteristics from the initial user characteristics based on the results of the correlation analysis of the initial user characteristics with the contact tags.
Optionally, the apparatus further includes a verification module, where the verification module is configured to:
determining a verification set, wherein the verification set comprises contact tags of the collection contact condition of the target user and user characteristics in a second time period;
and verifying the user loss rate prediction model based on the verification set.
In a fourth aspect, an embodiment of the present application provides an apparatus for predicting a user loss rate, where the apparatus includes:
the user characteristic acquisition module is used for acquiring the user characteristics of the user to be predicted;
and the loss of contact rate prediction module is used for inputting the user characteristics into a pre-trained user loss of contact rate prediction model to obtain a user loss of contact rate prediction result of the user to be predicted, wherein the user loss of contact rate prediction model is obtained by training according to the training method of the user loss of contact rate prediction model shown in any one of the above-mentioned embodiments of the first aspect.
Optionally, the apparatus further comprises:
and the collection urging strategy determining module is used for determining a collection urging strategy based on the prediction result of the user loss rate.
In a fifth aspect, an embodiment of the present application provides an electronic device, including: a processor and a memory;
a memory for storing operating instructions;
a processor configured to perform the method as shown in any implementation of the first aspect or any implementation of the second aspect of the present application by calling an operation instruction.
In a sixth aspect, embodiments of the present application provide a computer-readable storage medium on which a computer program is stored, which when executed by a processor, implements the method shown in any of the embodiments of the first aspect or any of the embodiments of the second aspect of the present application.
The technical scheme provided by the embodiment of the application has the following beneficial effects:
according to the scheme provided by the embodiment of the application, the target user in the designated service state is determined, and the training set is determined according to the contact label and the user characteristics of the target user in the first time period, so that the user loss rate prediction model is trained through a machine learning algorithm according to the training set. The user loss rate prediction model trained based on the scheme can accurately predict the loss rate of overdue users due, and can provide a basis for effectively improving the effect of collection and reducing overdue bad account rate.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings used in the description of the embodiments of the present application will be briefly described below.
Fig. 1 is a schematic flowchart of a training method of a user loss rate prediction model according to an embodiment of the present disclosure;
fig. 2 is a schematic flowchart of a method for predicting a user loss rate according to an embodiment of the present disclosure;
fig. 3 is a schematic structural diagram of a training apparatus of a user loss rate prediction model according to an embodiment of the present application;
fig. 4 is a schematic structural diagram of a device for predicting a user loss rate according to an embodiment of the present disclosure;
fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
Detailed Description
Reference will now be made in detail to embodiments of the present application, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are exemplary only for the purpose of explaining the present application and are not to be construed as limiting the present invention.
As used herein, the singular forms "a", "an", "the" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and/or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof. It will be understood that when an element is referred to as being "connected" or "coupled" to another element, it can be directly connected or coupled to the other element or intervening elements may also be present. Further, "connected" or "coupled" as used herein may include wirelessly connected or wirelessly coupled. As used herein, the term "and/or" includes all or any element and all combinations of one or more of the associated listed items.
To make the objects, technical solutions and advantages of the present application more clear, embodiments of the present application will be described in further detail below with reference to the accompanying drawings.
At present, when overdue debt is charged, a plurality of charging strategies generally exist, and different charging strategies often have different charging effects for different users. If the loss rate of the overdue user can be predicted, corresponding collection acceleration strategies can be adopted for users with different loss rates, so that the collection acceleration effect is improved, and the overdue bad account rate is reduced.
The training method of the user loss rate prediction model and the prediction method of the user loss rate provided by the embodiment of the application aim to solve at least one of the above technical problems in the prior art.
The following describes the technical solutions of the present application and how to solve the above technical problems with specific embodiments. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments. Embodiments of the present application will be described below with reference to the accompanying drawings.
Fig. 1 shows a schematic flowchart of a training method for a user loss rate prediction model according to an embodiment of the present application, and as shown in fig. 1, the method mainly includes:
step S110: determining a target user in a specified service state;
step S120: determining a training set, wherein the training set comprises a contact label and user characteristics of a target user in a first time period, and the contact label is used for representing the contact collection condition of the target user in the first time period;
step S130: and training the user loss rate prediction model through a machine learning algorithm according to the training set.
In the embodiment of the application, the user loss rate can be predicted by training the user loss rate prediction model. In order to ensure the prediction accuracy of the user loss rate prediction model, the customer base can be subdivided according to the service state of the target user, and modeling is performed according to the data of the target user in the specified service state.
In the embodiment of the application, details of the contact collection situation of the target user in the first time period can be obtained, so that the contact tag is determined. The first period of time may be a specified length of time after the overdue occurs, such as one month after the overdue occurs.
As one example, contact tags may include connected, disconnected, and difficult to connect.
In the embodiment of the application, the user characteristics can be extracted by acquiring the relevant information of the user from multiple directions, wherein the relevant information of the user is as follows: the method comprises the following steps of obtaining transaction information, credit urging information, card information, APP clicking behavior, external multi-head debit and credit, debit card asset and debt information, public accumulation fund, address comparison information and the like, so that the known related information is fully utilized to search user characteristics from multiple directions, and the prediction capability of a finally trained model is improved.
In the embodiment of the application, a machine learning algorithm (such as an Xgboost algorithm) may be used for modeling, so as to predict the probability of losing contact of a future user according to the relationship (i.e. the prior probability) between the user characteristics of the target user and the contact tag. In actual use, the loss probability can be converted into scores, and the corresponding receiving-urging strategy is determined according to the scores to urge receiving, so that the receiving-urging effect is ensured.
According to the method provided by the embodiment of the application, the target user in the designated service state is determined, and the training set is determined according to the contact label and the user characteristics of the target user in the first time period, so that the user loss rate prediction model is trained through a machine learning algorithm according to the training set. The user loss rate prediction model trained based on the scheme can accurately predict the loss rate of overdue users due, and can provide a basis for effectively improving the effect of collection and reducing overdue bad account rate.
In an optional manner of the embodiment of the present application, determining a target user in a specified service state includes:
and determining the users with the overdue within the target deadline as the target users of the specified business state.
According to the embodiment of the application, the users can be divided into different business states according to the overdue time period of the debt of the users. The target user is the user with the debt overdue time period within the target time limit.
As an example, a user whose business status is overdue within one month may be designated as M1 overdue, and in this example, a user whose M1 is overdue may be designated as a target user.
In an optional manner of the embodiment of the application, acquiring a contact tag of a contact condition of a target user in a first time period includes:
and determining a contact label associated with the prompt receiving contact condition of the target user based on the preset association relation.
In the embodiment of the application, the association relationship between the collection prompting contact condition and the contact tag can be preset, so that the contact tag is determined based on the contact condition.
Specifically, when the contact tag is determined, the details of the contact collection situation of the target user may be collected, and the collection records may be classified into a connected collection record, a non-connected collection record, and a difficult-to-connect collection record.
The online collection urging record can be a collection urging record which the user has committed to repayment and contacted with the user.
The no-connection collection prompting record can be a collection prompting record of cancelled mobile phone of the debt, an empty number or shutdown of a contact phone and the like.
The difficult-to-connect collection prompting record can be a collection prompting record of a user mobile phone occupation line or a mobile phone without listening.
When the contact tag is determined, the contact tag of the target user who has at least one contact collection record in the month can be determined as contact; determining the contact label of the target user without any connected collection record and at least one connected collection record in the current month as no connection; and determining the contact tags of the target users of the collection prompting records except the connected collection prompting record and the non-connected collection prompting record as non-difficult connection.
In an optional manner of the embodiment of the present application, determining a training set includes:
constructing a feature vector of each target user based on the contact tag and the user features;
and constructing a training set based on the feature vectors of the target users.
In the embodiment of the application, the characteristic variables can be constructed according to the contact labels and the user characteristics and used for model training.
In an optional manner of the embodiment of the present application, before constructing the feature vector of each target user based on the contact tag and the user feature, the method further includes:
and processing the abnormal condition existing in the user characteristic.
In the embodiment of the application, when the user features are extracted, some abnormal conditions may exist in the user features, such as missing values or abnormal values, and the like, and at this time, the abnormal conditions in the user features may be processed first, so that the subsequent model training process is prevented from being influenced.
In an optional manner of the embodiment of the present application, obtaining user characteristics of a target user in a first period includes:
acquiring user related information of a target user in a first period;
extracting initial user characteristics based on the user related information;
determining user characteristics from the initial user characteristics based on the results of the correlation analysis of the initial user characteristics with the contact tags.
In the embodiment of the application, the types of the related information of the user are possibly more, so that the initial user features extracted based on the related information of the user are possibly more in dimensionality, in actual use, the relevance between the initial user features and the contact tags can be analyzed, and the user features are screened from the initial user features according to the relevance analysis result.
Specifically, the feature information value IV analysis, the correlation analysis, the cross-time PSI analysis, and the like can be performed, and the validity of the extracted user feature is ensured later. The magnitude of the IV value of the user characteristic represents the strength of the predicting capability of the user characteristic, and the magnitude of the PSI of the user characteristic represents the stability of the user characteristic. In the model development process, the user characteristics with large IV value and small PSI value should be selected as much as possible.
In an optional manner of the embodiment of the present application, the method further includes:
determining a verification set, wherein the verification set comprises contact tags of the collection contact condition of the target user and user characteristics in a second time period;
and verifying the user loss rate prediction model based on the verification set.
In the embodiment of the application, a verification set can be constructed according to the contact tags and the user characteristics of the urging contact condition of the target user in the second time period, and is used for verifying the trained user loss rate prediction model.
In actual use, the effect verification can be performed on the user loss rate prediction model through indexes such as model distinguishing effect, grading stability, model pressure test, distinguishing capability of the model on other customer groups, and the like:
(1) and distinguishing effect and grading stability of the model. The KS value is a difference value for measuring the cumulative distribution of good and bad samples, the greater the cumulative difference of the good and bad samples is, the greater the KS index is, the stronger the risk distinguishing capability of the model is, and the better the risk ranking performance of the model is. The stronger the score stability, the better the model.
(2) And (5) testing the model pressure. And (4) carrying out pressure test on the model, and assuming the model distinguishing effect and the stability of the model score under the condition that partial variable data are missing.
(3) The ability of the model to differentiate among other groups of hastened guests. The sample data developed by the user loss rate prediction model is that M1 is overdue. After the model development is completed, the scoring logic of the loss of connection rate model is applied to the guest groups of the M2 overdue users, and a relatively obvious distinguishing effect can be still obtained.
Fig. 2 shows a schematic flow chart of a user loss rate prediction method provided in an embodiment of the present application, and as shown in fig. 2, the method mainly includes:
step S210: acquiring user characteristics of a user to be predicted;
step S220: and inputting the user characteristics into a pre-trained user loss rate prediction model to obtain a user loss rate prediction result of the user to be predicted, wherein the user loss rate prediction model is obtained by training according to the training method of the user loss rate prediction model.
In the embodiment of the application, the user to be predicted can be an overdue debt user, the loss rate of the user can be predicted through a pre-trained user loss rate prediction model, and the user characteristics can be extracted according to relevant information of the user to be predicted, such as transaction information, credit urging information, card information, APP clicking behavior, external multi-head debit, debit card asset and debt information, public accumulation fund, address comparison information and the like. And inputting the user characteristics into a pre-trained user loss rate prediction model, and outputting a user loss rate prediction result.
According to the method, the user characteristics of the user to be predicted can be obtained, the user characteristics are input into the pre-trained user loss rate prediction model, the user loss rate prediction result of the user to be predicted is obtained, based on the scheme, the loss rate of the overdue user can be accurately predicted, the effect of urging collection can be effectively improved, and the overdue bad-account rate can be reduced.
In an optional manner of the embodiment of the present application, the method further includes:
and determining a hasty collection strategy based on the user loss rate prediction result.
In the embodiment of the application, the predicted loss rate of the user can be obtained according to the loss rate prediction result.
In the embodiment of the application, different collection urging strategies can be set according to different user loss rates, so that after the user loss rate of the user is predicted, the collection urging strategies are determined according to the prediction result of the user loss rate.
As one example, different incentives may be set for different frequency of contacts to overdue users.
Based on the same principle as the method shown in fig. 1, fig. 3 shows a schematic structural diagram of a training apparatus of a user loss rate prediction model provided in an embodiment of the present application, and as shown in fig. 3, the training apparatus 30 of the user loss rate prediction model may include:
a target user determining module 310, configured to determine a target user in a specified service state;
a training set determining module 320, configured to determine a training set, where the training set includes a contact tag of a target user and user characteristics in a first time period, and the contact tag is used to represent a contact collection situation of the target user in the first time period;
and the model training module 330 is configured to train the user loss rate prediction model through a machine learning algorithm according to the training set.
According to the device provided by the embodiment of the application, the target user in the designated service state is determined, and the training set is determined according to the contact label and the user characteristics of the target user in the first time period, so that the user loss rate prediction model is trained through a machine learning algorithm according to the training set. The user loss rate prediction model trained based on the scheme can accurately predict the loss rate of overdue users due, and can provide a basis for effectively improving the effect of collection and reducing overdue bad account rate.
Optionally, the target user determination module is specifically configured to:
and determining the users with the overdue within the target deadline as the target users of the specified business state.
Optionally, when the contact tag of the urge to receive contact condition of the target user in the first time period is acquired, the training set determining module is specifically configured to:
and determining a contact label associated with the prompt receiving contact condition of the target user based on the preset association relation.
Optionally, the training set determining module is specifically configured to:
constructing a feature vector of each target user based on the contact tag and the user features;
and constructing a training set based on the feature vectors of the target users.
Optionally, the apparatus further comprises:
and the exception handling module is used for handling the exception condition existing in the user characteristics before the characteristic vector of each target user is constructed based on the contact tag and the user characteristics.
Optionally, when obtaining the user characteristics of the target user in the first period, the training set determining module is specifically configured to:
acquiring user related information of a target user in a first period;
extracting initial user characteristics based on the user related information;
determining user characteristics from the initial user characteristics based on the results of the correlation analysis of the initial user characteristics with the contact tags.
Optionally, the apparatus further includes a verification module, where the verification module is configured to:
determining a verification set, wherein the verification set comprises contact tags of the collection contact condition of the target user and user characteristics in a second time period;
and verifying the user loss rate prediction model based on the verification set.
Optionally, the apparatus further comprises:
and the collection urging strategy determining module is used for determining a collection urging strategy based on the prediction result of the user loss rate.
It can be understood that the above modules of the training apparatus for the user loss rate prediction model in the embodiment have functions of implementing corresponding steps of the training method for the user loss rate prediction model in the embodiment shown in fig. 1. The function can be realized by hardware, and can also be realized by executing corresponding software by hardware. The hardware or software includes one or more modules corresponding to the functions described above. The modules can be software and/or hardware, and each module can be implemented independently or by integrating a plurality of modules. For the functional description of each module of the training device of the user loss rate prediction model, reference may be specifically made to the corresponding description of the training method of the user loss rate prediction model in the embodiment shown in fig. 1, and details are not repeated here.
Based on the same principle as the method shown in fig. 2, fig. 4 shows a schematic structural diagram of a device for predicting a user loss rate provided in an embodiment of the present application, and as shown in fig. 4, the device 40 for predicting a user loss rate may include:
a user characteristic obtaining module 410, configured to obtain a user characteristic of a user to be predicted;
the loss of contact rate prediction module 420 is configured to input the user characteristics into a pre-trained user loss of contact rate prediction model to obtain a user loss of contact rate prediction result of the user to be predicted, where the user loss of contact rate prediction model is obtained by training according to the training method of the user loss of contact rate prediction model shown in any one of the above embodiments.
The device provided by the embodiment of the application can acquire the user characteristics of the user to be predicted, inputs the user characteristics into the pre-trained user loss rate prediction model, and obtains the user loss rate prediction result of the user to be predicted.
Optionally, the apparatus further comprises:
and the collection urging strategy determining module is used for determining a collection urging strategy based on the prediction result of the user loss rate.
It can be understood that the above modules of the device for predicting the user loss rate in the embodiment have functions of implementing corresponding steps of the method for predicting the user loss rate in the embodiment shown in fig. 2. The function can be realized by hardware, and can also be realized by executing corresponding software by hardware. The hardware or software includes one or more modules corresponding to the functions described above. The modules can be software and/or hardware, and each module can be implemented independently or by integrating a plurality of modules. For the function description of each module of the device for predicting the user loss rate, reference may be specifically made to the corresponding description of the method for predicting the user loss rate in the embodiment shown in fig. 2, and details are not repeated here.
The embodiment of the application provides an electronic device, which comprises a processor and a memory;
a memory for storing operating instructions;
and the processor is used for executing the method provided by any embodiment of the application by calling the operation instruction.
As an example, fig. 5 shows a schematic structural diagram of an electronic device to which an embodiment of the present application is applicable, and as shown in fig. 5, the electronic device 2000 includes: a processor 2001 and a memory 2003. Wherein the processor 2001 is coupled to a memory 2003, such as via a bus 2002. Optionally, the electronic device 2000 may also include a transceiver 2004. It should be noted that the transceiver 2004 is not limited to one in practical applications, and the structure of the electronic device 2000 is not limited to the embodiment of the present application.
The processor 2001 is applied to the embodiment of the present application to implement the method shown in the above method embodiment. The transceiver 2004 may include a receiver and a transmitter, and the transceiver 2004 is applied to the embodiments of the present application to implement the functions of the electronic device of the embodiments of the present application to communicate with other devices when executed.
The Processor 2001 may be a CPU (Central Processing Unit), general Processor, DSP (Digital Signal Processor), ASIC (Application Specific Integrated Circuit), FPGA (Field Programmable Gate Array) or other Programmable logic device, transistor logic device, hardware component, or any combination thereof. Which may implement or perform the various illustrative logical blocks, modules, and circuits described in connection with the disclosure. The processor 2001 may also be a combination of computing functions, e.g., comprising one or more microprocessors, DSPs and microprocessors, and the like.
The Memory 2003 may be a ROM (Read Only Memory) or other type of static storage device that can store static information and instructions, a RAM (Random Access Memory) or other type of dynamic storage device that can store information and instructions, an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical Disc storage, optical Disc storage (including Compact Disc, laser Disc, optical Disc, digital versatile Disc, blu-ray Disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited to these.
Optionally, the memory 2003 is used for storing application program code for performing the disclosed aspects, and is controlled in execution by the processor 2001. The processor 2001 is used to execute the application program code stored in the memory 2003 to implement the methods provided in any of the embodiments of the present application.
The electronic device provided by the embodiment of the application is applicable to any embodiment of the method, and is not described herein again.
Compared with the prior art, the embodiment of the application provides the electronic equipment, a training set is determined according to the contact label and the user characteristics of the target user in the specified service state, and therefore the user loss rate prediction model is trained through a machine learning algorithm according to the training set. The user loss rate prediction model trained based on the scheme can accurately predict the loss rate of overdue users due, and can provide a basis for effectively improving the effect of collection and reducing overdue bad account rate.
The present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the computer program implements the method shown in the above method embodiments.
The computer-readable storage medium provided in the embodiments of the present application is applicable to any of the embodiments of the foregoing method, and is not described herein again.
Compared with the prior art, the embodiment of the application provides a computer-readable storage medium, a target user in a specified service state is determined, a training set is determined according to a contact label and user characteristics of the target user in a first time period, and therefore a user loss rate prediction model is trained through a machine learning algorithm according to the training set. The user loss rate prediction model trained based on the scheme can accurately predict the loss rate of overdue users due, and can provide a basis for effectively improving the effect of collection and reducing overdue bad account rate.
It should be understood that, although the steps in the flowcharts of the figures are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and may be performed in other orders unless explicitly stated herein. Moreover, at least a portion of the steps in the flow chart of the figure may include multiple sub-steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed alternately or alternately with other steps or at least a portion of the sub-steps or stages of other steps.
The foregoing is only a partial embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.
Claims (13)
1. A training method of a user loss rate prediction model is characterized by comprising the following steps:
determining a target user in a specified service state;
determining a training set, wherein the training set comprises contact labels and user characteristics of the target user in a first time period, and the contact labels are used for representing the urge contact condition of the target user in the first time period;
and training a user loss rate prediction model through a machine learning algorithm according to the training set.
2. The method of claim 1, wherein the determining the target user in the specified traffic state comprises:
and determining the users with the overdue within the target deadline as the target users of the specified business state.
3. The method of claim 1, wherein obtaining a contact tag of the target user's proclaimed contact during a first time period comprises:
and determining a contact tag associated with the prompt contact condition of the target user based on a preset association relation.
4. The method of claim 1, wherein the determining the training set comprises:
constructing a feature vector of each target user based on the contact tag and the user features;
and constructing a training set based on the feature vectors of the target users.
5. The method of claim 4, wherein prior to constructing a feature vector for each of the target users based on the contact tags and the user features, the method further comprises:
and processing the abnormal condition existing in the user characteristic.
6. The method of claim 4, wherein obtaining the user characteristics of the target user in the first period comprises:
acquiring user related information of the target user in a first period;
extracting initial user features based on the user-related information;
determining user characteristics from the initial user characteristics based on results of the correlation analysis of the initial user characteristics with the contact tags.
7. The method according to any one of claims 1-6, further comprising:
determining a verification set, wherein the verification set comprises contact tags of the proclaimed contact condition of the target user and user characteristics in a second time period;
and verifying the user loss rate prediction model based on the verification set.
8. A method for predicting a user loss rate is characterized by comprising the following steps:
acquiring user characteristics of a user to be predicted;
inputting the user characteristics into a pre-trained user loss rate prediction model to obtain a user loss rate prediction result of the user to be predicted, wherein the user loss rate prediction model is obtained by training according to the training method of the user loss rate prediction model as claimed in any one of claims 1 to 7.
9. The method of claim 8, further comprising:
and determining a collection urging strategy based on the user loss rate prediction result.
10. A training device for a user loss rate prediction model is characterized by comprising:
the target user determining module is used for determining a target user in a specified service state;
a training set determination module, configured to determine a training set, where the training set includes a contact tag and a user characteristic of the target user in a first time period, and the contact tag is used to represent a contact collection situation of the target user in the first time period;
and the model training module is used for training the user loss rate prediction model through a machine learning algorithm according to the training set.
11. An apparatus for predicting a loss rate of a user, comprising:
the user characteristic acquisition module is used for acquiring the user characteristics of the user to be predicted;
and the loss of contact rate prediction module is used for inputting the user characteristics into a pre-trained user loss of contact rate prediction model to obtain a user loss of contact rate prediction result of the user to be predicted, wherein the user loss of contact rate prediction model is obtained by training according to the training method of the user loss of contact rate prediction model as claimed in any one of claims 1 to 7.
12. An electronic device comprising a processor and a memory;
the memory is used for storing operation instructions;
the processor is used for executing the method of any one of claims 1-9 by calling the operation instruction.
13. A computer-readable storage medium, characterized in that the storage medium has stored thereon a computer program which, when being executed by a processor, carries out the method of any one of claims 1-9.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110833599.7A CN113554227A (en) | 2021-07-23 | 2021-07-23 | Training method of user loss rate prediction model and user loss rate prediction method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110833599.7A CN113554227A (en) | 2021-07-23 | 2021-07-23 | Training method of user loss rate prediction model and user loss rate prediction method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113554227A true CN113554227A (en) | 2021-10-26 |
Family
ID=78132576
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110833599.7A Pending CN113554227A (en) | 2021-07-23 | 2021-07-23 | Training method of user loss rate prediction model and user loss rate prediction method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113554227A (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110245985A (en) * | 2019-06-11 | 2019-09-17 | 深圳前海微众银行股份有限公司 | A kind of information processing method and device |
CN110533520A (en) * | 2019-06-06 | 2019-12-03 | 上海凯京信达科技集团有限公司 | A kind of ranking method of the individual customer overdue loan grade based on multi-model |
CN111383093A (en) * | 2020-03-06 | 2020-07-07 | 北京网众共创科技有限公司 | Intelligent overdue bill collection method and system |
KR20200139603A (en) * | 2019-06-03 | 2020-12-14 | 고려대학교 산학협력단 | Method of predicting the best time to contact with delinquent customer, recording medium and apparatus for performing the method |
CN113554228A (en) * | 2021-07-23 | 2021-10-26 | 中信银行股份有限公司 | Repayment rate prediction model training method and repayment rate prediction method |
-
2021
- 2021-07-23 CN CN202110833599.7A patent/CN113554227A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20200139603A (en) * | 2019-06-03 | 2020-12-14 | 고려대학교 산학협력단 | Method of predicting the best time to contact with delinquent customer, recording medium and apparatus for performing the method |
CN110533520A (en) * | 2019-06-06 | 2019-12-03 | 上海凯京信达科技集团有限公司 | A kind of ranking method of the individual customer overdue loan grade based on multi-model |
CN110245985A (en) * | 2019-06-11 | 2019-09-17 | 深圳前海微众银行股份有限公司 | A kind of information processing method and device |
CN111383093A (en) * | 2020-03-06 | 2020-07-07 | 北京网众共创科技有限公司 | Intelligent overdue bill collection method and system |
CN113554228A (en) * | 2021-07-23 | 2021-10-26 | 中信银行股份有限公司 | Repayment rate prediction model training method and repayment rate prediction method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109034209B (en) | Training method and device for active risk real-time recognition model | |
CN111275491B (en) | Data processing method and device | |
JP2019511037A (en) | Method and device for modeling machine learning model | |
CN111932269A (en) | Equipment information processing method and device | |
CN111028016A (en) | Sales data prediction method and device and related equipment | |
CN111260368A (en) | Account transaction risk judgment method and device and electronic equipment | |
CN113554228B (en) | Training method of repayment rate prediction model and repayment rate prediction method | |
CN108399482A (en) | Appraisal procedure, device and the electronic equipment of contract | |
CN112200402B (en) | Risk quantification method, device and equipment based on risk portrait | |
CN109635953A (en) | A kind of feature deriving method, device and electronic equipment | |
CN111882140A (en) | Risk evaluation method, model training method, device, equipment and storage medium | |
CN110334936B (en) | Method, device and equipment for constructing credit qualification scoring model | |
CN112529575A (en) | Risk early warning method, equipment, storage medium and device | |
CN116823428A (en) | Anti-fraud detection method, device, equipment and storage medium | |
CN109783381B (en) | Test data generation method, device and system | |
CN111784053A (en) | Transaction risk detection method, device and readable storage medium | |
CN111275071A (en) | Prediction model training method, prediction device and electronic equipment | |
CN113554227A (en) | Training method of user loss rate prediction model and user loss rate prediction method | |
CN113052604A (en) | Object detection method, device, equipment and storage medium | |
CN111324732B (en) | Model training method, text processing device and electronic equipment | |
CN113065944A (en) | Intelligent analysis and evaluation method and system for credit risk of letter and goods condition | |
CN115953248B (en) | Wind control method, device, equipment and medium based on saprolitic additivity interpretation | |
CN113837863B (en) | Business prediction model creation method and device and computer readable storage medium | |
CN114707990B (en) | User behavior pattern recognition method and device | |
CN110782342B (en) | Method and device for verifying correctness of new channel feature engineering based on binary classification model |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |