CN111383072A

CN111383072A - User credit scoring method, storage medium and server

Info

Publication number: CN111383072A
Application number: CN201811653564.XA
Authority: CN
Inventors: 罗建平
Original assignee: TCL Research America Inc
Current assignee: TCL Corp; TCL Research America Inc
Priority date: 2018-12-29
Filing date: 2018-12-29
Publication date: 2020-07-07

Abstract

The invention discloses a user credit scoring method, a storage medium and a server, wherein the method comprises the following steps: providing user basic data; generating a fact label according to the data characteristics of the user basic data; generating at least one model label according to the plurality of fact labels; and scoring each preset prediction label by adopting a weighting model according to the plurality of fact labels and at least one model label. According to the method, the basic data of the user are collected through multiple channels, the basic data collected through each channel are integrated and analyzed to obtain the fact label and the model label, the preset prediction labels are graded in a weighting mode, and the credit score of the user is determined according to the grade of each prediction label, so that the grading accuracy is improved.

Description

User credit scoring method, storage medium and server

Technical Field

The invention relates to the technical field of data processing, in particular to a user credit scoring method, a storage medium and a server.

Background

And (3) the credit score of the user, namely the credit granting credibility of the user needing loan or credit card consumption is measured. With the progress of society, the living standard of people is continuously improved, and the consumption concept of people is advanced more and more. The trend of the subject consumption is gradually towards the younger, with the shift of mainstream consumption concepts. In a survey of young people, 57% of visitors expressed "dare for tomorrow's money" and 48% called themselves "do not worry about being consumed by debt". In the back hands of the young adult 90, essentially on average each person has at least one credit card. Consumption by credit can help young people travel abroad and deeply build. It is also gradually entering the era of advanced consumption at the intermediate-stage. They not only buy houses and cars, but also buy various consumer goods by using installment payment. The current consumption concept leads to blind consumption of partial consumers, and particularly after 90 and 00, the consumption without considering the economic bearing capacity of the consumers is also large. Therefore, aiming at the consumption behavior of the user, the bank or the e-commerce shopping platform trusts reasonable credit line to the consumer, and the feasible credit mechanism is researched by utilizing technologies such as big data processing and the like, which is a topic worthy of research.

The existing user credit scoring method only considers a bank credit card consumption and use record data source directly related to a user, and the main source of the bank credit card consumption and use record data source is offline data, so that the problems of slow updating of the data source and lack of timeliness exist, and the problems of hysteresis and low accuracy exist in user credit scoring.

Disclosure of Invention

The technical problem to be solved by the present invention is to provide a user credit scoring method, a storage medium and a server aiming at the defects of the prior art, so as to solve the problem that the existing user credit scoring has low accuracy.

In order to solve the technical problems, the technical scheme adopted by the invention is as follows:

a user credit scoring method, comprising:

providing user basic data, and generating a fact label according to data characteristics of the user basic data;

generating at least one model tag from the fact tags;

and scoring each preset prediction label in a weighting mode according to the fact label and at least one model label, and obtaining the user credit score of the user according to the score of each prediction label.

The user credit scoring method, wherein the providing of the user basic data and the generating of the fact label according to the data characteristics of the user basic data specifically include:

respectively acquiring the characteristic attribute and the characteristic keyword of each basic data in the user basic data;

grouping the basic data according to the feature keywords, and taking the feature keywords as corresponding grouped fact labels;

determining the label attribute of each fact label according to the characteristic attribute of the basic data contained in each fact label, and adding each label attribute into the corresponding fact label, wherein the label attribute comprises a dynamic attribute and a static attribute.

The user credit scoring method specifically includes the following steps of respectively acquiring the feature attributes and the feature keywords of each basic data in the user basic data:

respectively acquiring feature keywords and data types of each basic data of user basic data;

and determining the characteristic attribute of each basic data according to the data type to obtain the characteristic key words and the characteristic attribute of each basic data.

The user credit scoring method, wherein the generating at least one model tag according to the plurality of fact tags specifically includes:

acquiring the correlation among the fact labels in the plurality of fact labels;

and modeling and analyzing all the fact labels according to the correlation to generate at least one model label, wherein each model label corresponds to at least one fact label.

The user credit scoring method, wherein scoring preset prediction labels in a weighting manner according to the fact labels and the at least one model label, and obtaining the user credit score of the user according to the score of the prediction labels specifically includes:

for each preset prediction label, searching a corresponding weight list in a preset database, wherein the weight list comprises weight influence values of all fact labels and assignment probability values of all model labels;

determining a weighting model of each prediction label according to the obtained weight list, and grading each prediction label through the weighting model;

and obtaining the user credit score of the user according to the scores of the prediction labels.

The user credit scoring method may further include, before scoring each preset prediction tag in a weighting manner according to the fact tags and the at least one model tag, obtaining a user credit score of the user according to the score of each prediction tag:

and classifying all model labels by adopting a clustering analysis method to generate a plurality of prediction labels.

The user credit scoring method includes the following specific steps that the user basic data are acquired through a plurality of preset acquisition channels:

acquiring identity information of a user;

and respectively acquiring basic data of the user through a credit investigation channel, a terminal equipment channel and a third-party application channel according to the user identity information.

The user credit scoring method, wherein the method comprises the following steps:

and obtaining the score value of each prediction label, and setting a credit granting grade corresponding to each prediction label according to the score value, wherein the credit granting grade is in direct proportion to the score value.

A computer readable storage medium, storing one or more programs, the one or more programs being executable by one or more processors to perform the steps of the user credit scoring method as recited in any of the above.

An application server, comprising: a processor, a memory, and a communication bus; the memory has stored thereon a computer readable program executable by the processor;

the communication bus realizes connection communication between the processor and the memory;

the processor, when executing the computer readable program, implements the steps in the user credit scoring method as described in any of the above.

Has the advantages that: compared with the prior art, the invention provides a user credit scoring method, a storage medium and a server, wherein the method comprises the following steps: providing user basic data; generating a fact label according to the data characteristics of the user basic data; generating at least one model label according to the plurality of fact labels; and scoring each preset prediction label by adopting a weighting model according to the plurality of fact labels and at least one model label. According to the method, the basic data of the user are collected through multiple channels, the basic data collected through each channel are integrated and analyzed to obtain the fact label and the model label, the preset prediction labels are graded in a weighting mode, and the credit score of the user is determined according to the grade of each prediction label, so that the grading accuracy is improved.

Drawings

Fig. 1 is a flowchart of an embodiment of a user credit scoring method provided by the present invention.

Fig. 2 is a flowchart of step S10 in an embodiment of the user credit scoring method provided in the present invention.

Fig. 3 is a flowchart of step S11 in an embodiment of the user credit scoring method provided in the present invention.

Fig. 4 is a diagram illustrating a relationship between a fact tag and a model tag in an embodiment of the user credit scoring method according to the present invention.

Fig. 5 is a flowchart of step S30 in an embodiment of the user credit scoring method provided in the present invention.

Fig. 6 is a schematic structural diagram of an embodiment of an application server provided in the present invention.

Detailed Description

The present invention provides a user credit scoring method, a storage medium and a server, and in order to make the purpose, technical scheme and effect of the present invention clearer and clearer, the present invention will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the 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.

It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the prior art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.

The invention will be further explained by the description of the embodiments with reference to the drawings.

Referring to fig. 1, fig. 1 is a flowchart illustrating a control method for application self-starting according to a preferred embodiment of the present invention. The method comprises the following steps:

and S10, providing user basic data, and generating a fact label according to the data characteristics of the user basic data.

Specifically, the user basic data is acquired through a plurality of acquisition channels, and the user basic data may include bank credit investigation report data of the user, behavior data of the television end user, and third-party data. Correspondingly, the acquisition channels can comprise a credit investigation channel, a terminal equipment channel and a third-party application channel. The credit investigation channel can collect credit investigation report data of a user, and the credit investigation report data can include basic residence and occupation information, credit information, loan information, guarantee information, external guarantee information, non-seller quasi credit card information, non-seller credit card information, credit examination and approval query records, pension insurance deposit information, overdue (overdraft) information, housing public accumulation fund payment information, quasi credit card information, overdue records, spouse tables and other information of the user. The terminal equipment channel collects user behavior data of the terminal equipment, and the user behavior data of the terminal equipment can be user shopping information, routing information, television equipment information and the like. The third-party application channel is used for collecting third-party user behavior data, and the third-party user behavior data can include various directly or indirectly acquired information such as the model of a mobile phone used by a user, data of various apps used in the mobile phone of the user, whether the user purchases a car or drives a company, and the like.

In addition, in order to improve the timeliness and comprehensiveness of data, different acquisition intervals are configured for different acquisition channels. For example, the collection interval time of the terminal device channel and the third-party application channel can be 1 day, and the collection interval time of the credit investigation channel can be one month. In addition, the original log data of the data collected each time are synchronized to a file system such as an HDFS in an automatic synchronization mode for storage. Therefore, when the data acquired and updated each time is missing, the data can be searched and recovered in a file system such as HDFS.

Further, when the user basic data is adopted through each acquisition channel, a user identifier needs to be configured for each acquisition channel, and each acquisition channel acquires the user basic data according to the user identifier. Correspondingly, the step of acquiring the user basic data through a plurality of preset acquisition channels specifically comprises the following steps: acquiring identity information of a user; and respectively acquiring basic data of the user through a credit investigation channel, a terminal equipment channel and a third-party application channel according to the user identity information. The identity information is a unique identity tag of the user, and the identity information can be used for distinguishing the user. In this embodiment, the identity information may be a telephone number, an identification number, and the like.

Further, in this embodiment, the data features include attribute features and feature keywords, and the feature keywords express the use of the basic data, for example, the feature keywords of the identity number may be identity information. That is, each data is configured with one feature keyword, and the user basic data collected through each acquisition channel may include user basic data having the same feature keyword. Therefore, basic data with the same characteristic keywords can be used as a group of data, the characteristic keywords can be used as the characteristic keywords of the group of data, and data of repeated fact labels can be reduced. Correspondingly, as shown in fig. 2, the providing of the user basic data and the generating of the fact label according to the data characteristics of the user basic data specifically include:

s11, acquiring the user basic data through a plurality of acquisition channels, wherein the acquisition channels comprise a credit investigation channel, a terminal equipment channel and a third-party application channel.

S12, respectively obtaining the characteristic attribute and the characteristic keyword of each basic data in the user basic data;

s13, grouping the basic data according to the feature keywords, and taking the feature keywords as corresponding grouped fact labels;

s14, determining the label attribute of each fact label according to the characteristic attribute of the basic data contained in each fact label, and adding each label attribute into the corresponding fact label, wherein the label attribute comprises a dynamic attribute and a static attribute.

Specifically, the characteristic attribute refers to the persistence of user basic data, for example, basic data such as population attribute, equipment attribute and other attributes of the user, the persistence of the basic data from generation to change is generally long; and the behaviors such as browsing, searching, viewing and the like generated by the channel APP and the like are relatively short in duration, so that the basic data can be divided into dynamic data and static data according to the characteristic attributes of the basic data. Accordingly, the fact tag can be divided into a dynamic tag and a static tag according to the basic data contained in the fact tag, that is, the tag attributes of the fact tag can be a dynamic attribute and a static attribute. In addition, when the fact label is corresponding to a plurality of basic data, the attribute features contained in each basic data can be respectively obtained, when the attribute features of all the basic data are consistent, the attribute features are used as the label attributes of the fact label, and when the attribute features of all the basic data are inconsistent, the label attributes of the fact label can be determined by adopting a principle that a small number of the basic data are subject to a majority. Of course, in order to improve the accuracy of the tag attribute, when the attribute features of all the basic data included in the fact tag are inconsistent, the basic data may be divided into two groups of basic data according to the attribute features, and the fact tag is divided into two fact tags, where the two fact tags correspond to the two groups of basic data, and the tag attributes are configured for the two fact tags respectively. Wherein two fact tags can be distinguished in a 1 and 2 manner.

Further, as shown in fig. 3, the respectively obtaining the feature attribute and the feature keyword of each basic data in the user basic data specifically includes:

s121, respectively obtaining feature keywords and data types of each basic data in the user basic data;

and S122, determining the characteristic attribute of each basic data according to the data type to obtain the characteristic key words and the characteristic attribute of each basic data.

Specifically, the data types include an attribute type and a behavior type, and the attribute type may include fixed data such as a user attribute and a device attribute. The user attributes may include gender, age, city, academic calendar, marital status, and the like, and the device attributes may include an operator, a network environment, a model, a version number, and the like. The behavior type may be user behavior data formed by the terminal device or the third application, and may include behavior data, contact websites/APPs, travel trip, and the like. The behavior data can comprise browsing data, collecting data, handling cards, opening cards and the like, the contact website/APP can comprise a Jingdong shopping mall, a Taobao, a microblog and the like, and the business trip can comprise a map, a designated driving, a public transport, a hotel and the like. In this embodiment, the characteristic attribute of the attribute-type data is set as static data, and the characteristic attribute of the behavior-type data is set as dynamic data. In the implementation, after configuring the tag attribute for the real tag, when updating the basic data of the user, the update period of the static data is set to be long, and the update period of the dynamic data is set to be short, so that the real-time performance of the basic data can be ensured, and the data updating time can be reduced.

Further, in an embodiment of the present invention, before the data features of each piece of basic data included in the user basic data are respectively obtained, the bases acquired by each acquisition channel may be respectively encrypted, so as to protect the sensitive data (e.g., information such as an identity card number, a mobile phone mac address, a home address, etc.) acquired by each acquisition channel. In which an algorithm for desensitizing data, such as MD5 encryption, may be used for encryption. Therefore, the basic data collected by each acquisition channel can be prevented from being leaked, and the privacy of the user is protected.

And S20, generating at least one model label according to the plurality of fact labels.

Specifically, the model tags are generated according to fact tags, and each model tag may correspond to one obtaining multiple fact tags. For example, as shown in FIG. 4, modeling model tag L1 uses fact tag L1 through fact tag L4; fact tag L2 and fact tag L3 were used when modeling model tag L2. In addition, in this embodiment, the data collected by all the acquisition channels may be modeled to obtain model tags, or the fact tags included in multiple channels may be modeled to obtain each model tag, and each model tag may carry its corresponding acquisition channel tag. Of course, when the fact labels of the acquisition channels are separately divided, the fact labels need to be configured with the acquisition channel identifiers when being established, and one fact label can carry a plurality of acquisition channel identifiers.

Further, when modeling analysis is performed based on fact tags, modeling may be performed based on correlation between fact tags. Correspondingly, the modeling analysis of the fact labels to obtain the model labels specifically comprises obtaining the correlation among the labels in the fact labels; and modeling and analyzing all the fact labels according to the correlation to obtain at least one model label, wherein each model label corresponds to at least one fact label. Wherein the correlation may be behavioral similarity, for example, if the fact label includes a plurality of fact labels for purchasing infant products, the maternal model label may be generated according to the plurality of fact labels for purchasing infant products. Of course, modeling and analyzing all the fact labels to obtain at least one model label may be to classify all the fact labels by using a clustering method and to set one model label for each classification. The model tags may include consumption level tags, consumption cycle tags, channel (or app) preference tags, relationship map tags, and professional status tags, among others. The consumption grade label can be suggested according to the fact labels such as the use of shopping APP or shopping channels by the user provided by the third party, the brand information of the watched television, the shopping commodity information and the consumption amount of the television end user and the like; the consumption period label can be established according to the fact labels of the user's income and expense condition, consumption payment ability, repayment ability, consumption times in a fixed period and the like; the channel (or APP) preference tag may be established according to the preference APP type and the preference channel tag; the relationship map label can be established according to the fact label of the social attribute of the user; the job status label can be established according to the fact label of the user to the use condition of job hunting APP such as 51 jobs, hook net, collar English, 58 city and the like.

And S30, scoring each preset prediction label in a weighting mode according to the fact labels and the at least one model label, and obtaining the user credit score of the user according to the score of each prediction label.

Specifically, the prediction tag may be preset, or may be obtained by performing cluster analysis on all model tags. And the score corresponding to the prediction label is obtained according to the model label related to the prediction label. For example, the score corresponding to the prediction label of the consumer level is obtained by the model label of all (related) consumers of the user. The influence of each fact label and each model label on different prediction labels is different, so that a weight list is configured for each prediction label in advance, and the weight lists of all the prediction labels are stored in a weight database so as to be convenient for searching. Correspondingly, as shown in fig. 5, the scoring the preset prediction tags in a weighting manner according to the fact tags and the at least one model tag, and obtaining the user credit score of the user according to the score of each prediction tag specifically includes:

s31, for each prediction label, searching a corresponding weight list in a preset database, wherein the weight list comprises weight influence values of all fact labels and assignment probability values of all model labels;

s32, determining a weighting model of each prediction label according to the obtained weight list, and scoring each prediction label through the weighting model;

and S33, obtaining the user credit score of the user according to the scores of the prediction labels.

Specifically, the weight influence value of each fact label and the assignment probability value of each model label may be configured in advance, and a weighted model of a prediction label may be determined according to the weight influence value of each fact label and the assignment probability value of each model label, and the prediction label is scored by weighting, so that a model obtained by only using several fact labels or a model label generated by only using a fact label in a single channel is avoided from scoring the prediction label, thereby avoiding the limitation that the user credit scoring may have.

Also in this embodiment, the weighting model may be adopted

p₁+p₂+...+p_i＝1；p₁+p₂+...+p_j＝1

Wherein { Xi ═ { event (i) } } represents whether or not an event i has occurred, and assigns a value of 1 if the event Xi tag has occurred, and assigns a value of 0 otherwise; p is a radical of_iWhen the event Xi occurs, the weight influence degree of the event Xi on a certain prediction label is generated; p is a radical of_jBased on the weighted combination of the plurality of model tags, the assigned probability value before each model tag, α is the weight value of the total (which is the total score criterion, such as 100 or 1000 points), and W is the score value of the predicted tag.

Further, the weight influence values and the assignment probability values of the model labels can be updated according to the prediction scores and the corresponding behavior scores. Correspondingly, for each prediction tag, searching the corresponding weight list in the preset database specifically includes:

for each prediction label, searching a corresponding weight list in a preset database;

and reading the fact label of each dynamic attribute, and updating the influence value of the dynamically read fact label according to the previous prediction score and the behavior score.

Specifically, when the weight influence value of each fact label is updated, the label attribute of each fact label may be determined, when the label attribute is a static attribute, the weight influence value of the fact label is not changed, and when the label attribute is a dynamic attribute, the weight influence value of the fact label is updated according to the relationship between the prediction score and the behavior score corresponding thereto. Further, in one embodiment of the present invention, the method comprises: and obtaining the score value of each prediction label, and setting a credit granting grade corresponding to each prediction label according to the score value, wherein the credit granting grade is in direct proportion to the score value. The credit corresponding to the prediction tag can include credit in channels such as an installment loan amount, a credit card credible consumption amount per month, a consumable credit amount in a shopping mall, and the like.

Further, after the score value of each prediction label is determined, scoring is carried out on the user information according to the score value of each prediction label. Wherein the user credit score may be an average of scores of the respective predictive tags. In practical applications, in order to improve the accuracy of the user credit score, when determining the user credit score, weights may be configured for the prediction tags, and then the user credit score may be determined in a weighting manner, where the weights of the prediction tags may be configured in advance, and may be set according to importance levels of the prediction tags.

Based on the user credit scoring method, the invention further provides a computer-readable storage medium storing one or more programs, which can be executed by one or more processors to implement the steps of the user credit scoring method as described in any one of the above

Based on the above user credit scoring method, the present invention further provides an application server, as shown in fig. 6, which includes at least one processor (processor) 20; a display screen 21; and a memory (memory)22, and may further include a communication Interface (Communications Interface)23 and a bus 24. The processor 20, the display 21, the memory 22 and the communication interface 23 can communicate with each other through the bus 24. The display screen 21 is configured to display a user guidance interface preset in the initial setting mode. The communication interface 23 may transmit information. The processor 20 may call logic instructions in the memory 22 to perform the methods in the embodiments described above.

Furthermore, the logic instructions in the memory 22 may be implemented in software functional units and stored in a computer readable storage medium when sold or used as a stand-alone product.

The memory 22, which is a computer-readable storage medium, may be configured to store a software program, a computer-executable program, such as program instructions or modules corresponding to the methods in the embodiments of the present disclosure. The processor 20 executes the functional application and data processing, i.e. implements the method in the above-described embodiments, by executing the software program, instructions or modules stored in the memory 22.

The memory 22 may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created according to the use of the terminal device, and the like. Further, the memory 22 may include a high speed random access memory and may also include a non-volatile memory. For example, a variety of media that can store program codes, such as a usb disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk, may also be transient storage media.

In addition, the specific processes loaded and executed by the instruction processors in the storage medium and the application server are described in detail in the method, and are not stated herein.

Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.

Claims

1. A method for scoring a user's credit, comprising:

generating at least one model tag from the fact tags;

2. The user credit scoring method according to claim 1, wherein the providing user basic data and the generating of the fact label according to the data characteristics of the user basic data specifically comprise:

acquiring user basic data through a plurality of acquisition channels, wherein the acquisition channels comprise a credit investigation channel, a terminal equipment channel and a third-party application channel;

3. The user credit scoring method according to claim 2, wherein the respectively obtaining the feature attribute and the feature keyword of each basic data in the user basic data specifically comprises:

4. The user credit scoring method of claim 1, wherein the generating at least one model tag from the fact tags specifically comprises:

acquiring the correlation among the fact labels;

5. The method of claim 1, wherein the scoring the preset prediction tags in a weighting manner according to the fact tag and the at least one model tag, and obtaining the user credit score of the user according to the score of each prediction tag specifically comprises:

6. The method according to claim 1 or 5, wherein before scoring each preset prediction tag in a weighting manner according to the fact tag and at least one model tag and obtaining the user credit score of the user according to the score of each prediction tag, the method further comprises:

7. The method as claimed in claim 2, wherein the step of obtaining the user basic data through a plurality of predetermined obtaining channels comprises:

acquiring identity information of a user;

8. The method of claim 1, wherein the method comprises:

9. A computer readable storage medium, storing one or more programs, which are executable by one or more processors to perform the steps of the method of any one of claims 1 to 8.

10. An application server, comprising: a processor, a memory, and a communication bus; the memory has stored thereon a computer readable program executable by the processor;

the processor, when executing the computer readable program, implements the steps in the user credit scoring method of any one of claims 1-8.