CN112270596A - Risk control system and method based on user portrait construction - Google Patents
Risk control system and method based on user portrait construction Download PDFInfo
- Publication number
- CN112270596A CN112270596A CN202011149392.XA CN202011149392A CN112270596A CN 112270596 A CN112270596 A CN 112270596A CN 202011149392 A CN202011149392 A CN 202011149392A CN 112270596 A CN112270596 A CN 112270596A
- Authority
- CN
- China
- Prior art keywords
- data
- user
- portrait
- classification
- risk
- 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.)
- Withdrawn
Links
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
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/03—Credit; Loans; Processing thereof
-
- 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
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
Abstract
The invention relates to the technical field of risk control, in particular to a risk control system and method based on user portrait construction. The system comprises: the system comprises a data processing module, a user portrait construction module and a risk control module; the data processing module is configured to acquire user data, perform data preprocessing on the acquired user data, perform data classification on a result of the user data preprocessing, and add a data class label to the user data of each class; the user data at least comprises: the system comprises user information and user historical behaviors, wherein the user historical behaviors are uniquely subordinate to one piece of user information. The user portrait is constructed by using the historical behavior data of the user, so that the accuracy and the scientification of risk control are realized; meanwhile, in the process of constructing the user portrait, the accuracy and the efficiency of constructing the user portrait are improved by adopting an innovative data classification algorithm.
Description
Technical Field
The invention belongs to the technical field of risk control, and particularly relates to a risk control system and method based on user portrait construction.
Background
At present, banks and financial institutions in China basically take early-stage financial investigation and audit work as the center of gravity of risk control, but neglect later-stage financial risks caused by insufficient audit of early-stage financial information or financial fraud information. China banks and financial institutions still find risks by means of manual work for post-loan management work, such as field investigation, telephone verification and the like, and the defects of the manual risk mainly include the following aspects:
personnel cost is high: both field investigation and phone verification require significant costs.
Lack of post-loan risk awareness: due to the influence of the domestic large environment, the cognition deficiency of the risk management after the loan is insufficient, and the management after the loan becomes a layout focusing on the pre-loan and the mid-loan.
Human competence: the post-loan managers have uneven ability to find and handle risks, which leads to untimely risk finding and avoidance.
Risk delay: risk delay avoiding difficulty caused by difficult contact of customers and the like due to a plurality of uncertain factors such as manual investigation, telephone verification and the like.
Lack of early warning data: sources of post-loan management data rely primarily on pre-loan audit data as a basis, while borrower data may not be constant enough to discover post-loan risks if only the data of the user's financial application is referenced.
Due to the reasons, the work of banks and financial institutions in China on risk management after loan is difficult and serious, and behaviors of fraud after loan, overdue finance, malicious delinquent and the like which disturb the order of the financial market still frequently occur.
Disclosure of Invention
In view of the above, the main objective of the present invention is to provide a risk control system and method based on user portrait construction, which constructs a user portrait by using historical behavior data of a user, and realizes accuracy and scientization of risk control; meanwhile, in the process of constructing the user portrait, the accuracy and the efficiency of constructing the user portrait are improved by adopting an innovative data classification algorithm.
In order to achieve the purpose, the technical scheme of the invention is realized as follows:
a risk control system constructed based on a user representation, the system comprising:
the data processing module is configured to acquire user data, perform data preprocessing on the acquired user data, perform data classification on a result of the user data preprocessing, and add a data classification label to the user data of each classification; the user data at least comprises: the method comprises the following steps of (1) user information and user historical behaviors, wherein the user historical behaviors are only subordinate to one user information; when the data classification is carried out, classifying a plurality of user historical behaviors corresponding to user information based on a preset classification model, and adding a data classification label to each classified user historical behavior; the data category label includes at least: low risk, medium risk and high risk categories;
the user portrait construction module is configured to construct a user portrait corresponding to each user by using a preset model based on a plurality of user historical behaviors which are added with data category labels and correspond to each user information;
the risk control module is configured for associating the constructed user portrait with preset seed users with different financial risk levels to obtain user portrait data with different financial risk levels; and grading the users according to the user image data of different financial risk grades.
Further, the data processing module comprises: a data acquisition unit configured to acquire user data; the data preprocessing unit is configured to perform data preprocessing on the acquired user data; the data preprocessing process specifically comprises: removing unique attributes, processing missing values, detecting and processing abnormal values, processing data specifications and processing data standardization; the data classification unit is configured to perform data classification on the result of the user data preprocessing and add a data classification label to the user data of each classification; the data classification process specifically includes: randomly dividing the result of the pre-processing of the user data into two sets T1And T2(ii) a According to said two sets T1And T2Classifying the result data of the user data preprocessing by using a classification model constructed by the following formula:
wherein λ ismClassifying items for a collectionThe value range of the weight is as follows: 1-5; g is a classification number, and the value is 3 or 6; the weight value of the v-bit correction term ranges from 0.1 to 0.4;
for gradient operator, H (T) is a correction function, which is set as a linear function, and the parameters of the linear function can be set;
as a set T1And T2Is estimated as a probability density function, pmAs a set T1And T2A probability density function of; delta is an intensity mean value adjusting parameter, and the value range is as follows: 3-5; i is the intensity mean function.
Further, the data preprocessing unit, a method for preprocessing the acquired user data, includes: removing the unique attribute, processing missing values and abnormal value detection and processing; and carrying out data specification processing, including: removing an average value, calculating a covariance matrix, calculating an eigenvalue and an eigenvector of the covariance matrix, sorting the eigenvalues from large to small, reserving the largest k eigenvector, and converting data into a new space constructed by the eigenvector; finally, new processed data are obtained, and the data are irrelevant pairwise, but the original information is kept. And carrying out data standardization processing, and scaling the data to enable the data to fall into a set interval.
Further, the risk control module includes: the risk association unit is configured for associating the constructed user portrait with preset seed users with different financial risk levels to obtain user portrait data with different financial risk levels; and the risk grade establishing unit is configured with a grading model used for training user portrait data based on different financial risk grades to obtain the financial risk grade, and the grading model is used for grading the user on a preset application program to be divided into different financial risk grades.
Further, the method for constructing the user portrait corresponding to the user by the user portrait construction module based on the plurality of user historical behaviors after the data category labels are added corresponding to each user information and by using a preset model includes: constructing an intermediate portrait according to the user historical behaviors and the categories of the user historical behaviors; constructing the user representation from the intermediate representation; the constructing an intermediate portrait according to the intermediate portrait and the type of the intermediate portrait specifically includes: calculating the average value of the vector of the intermediate portrait, and expressing the semantic meaning of the intermediate portrait through the average value of the vector; calculating the vector average value of the intermediate image of the same type according to the average value of the vectors of the intermediate image, and taking the vector average value as the intermediate image; when the intermediate representation includes one or more of the types, one or more of the intermediate representations are constructed.
A risk control method based on user profile construction, the method performing the steps of:
step 1: acquiring user data, performing data preprocessing on the acquired user data, performing data classification on the result of the user data preprocessing, and adding a data classification label to the user data of each classification; the user data at least comprises: the method comprises the following steps of (1) user information and user historical behaviors, wherein the user historical behaviors are only subordinate to one user information; when the data classification is carried out, classifying a plurality of user historical behaviors corresponding to user information based on a preset classification model, and adding a data classification label to each classified user historical behavior; the data category label includes at least: low risk, medium risk and high risk categories;
step 2: based on a plurality of user historical behaviors which are added with data category labels and correspond to each user information, constructing a user portrait corresponding to the user by using a preset model;
and step 3: associating the constructed user portrait with preset seed users with different financial risk levels to obtain user portrait data with different financial risk levels; and grading the users according to the user image data of different financial risk grades.
Further, in the step 1: the method for preprocessing the acquired user data comprises the following steps: removing the unique attribute, processing missing values and abnormal value detection and processing; and carrying out data specification processing, including: removing an average value, calculating a covariance matrix, calculating an eigenvalue and an eigenvector of the covariance matrix, sorting the eigenvalues from large to small, reserving the largest eigenvector, and converting data into a new space constructed by the eigenvector; finally, new processed data are obtained, and the data are irrelevant pairwise, but the original information is kept. And carrying out data standardization processing, and scaling the data to enable the data to fall into a set interval.
Further, in the step 1: the method for classifying the data of the result of the user data preprocessing and adding the data category label to the user data of each category comprises the following steps: randomly dividing the result of the pre-processing of the user data into two sets T1And T2(ii) a According to said two sets T1And T2Classifying the result data of the user data preprocessing by using a classification model constructed by the following formula:
wherein λ ismThe value range of the weight of the set classification item is as follows: 1-5; g is a classification number, and the value is 3 or 6; the weight value of the v-bit correction term ranges from 0.1 to 0.4;
for gradient operator, H (T) is a correction function, which is set as a linear function, and the parameters of the linear function can be set;
as a set T1And T2Is estimated as a probability density function, pmAs a set T1And T2A probability density function of; delta is an intensity mean value adjusting parameter, and the value range is as follows: 3-5; i isIs an intensity mean function; for the classified user data, if the classified user data still belongs to the set T1Then add a high risk label; if it is classified, it still belongs to the set T2Then add a low risk label; if classified, the group does not belong to the set T1(ii) a Nor to the set T2Then a tag of medium risk is added.
Further, the step 3 specifically includes: associating the constructed user portrait with preset seed users with different financial risk levels to obtain user portrait data with different financial risk levels; training user portrait data based on different financial risk levels to obtain a grading model of the financial risk levels, and grading users on a preset application program by using the grading model to obtain different financial risk levels.
Further, the step 2: the method for constructing the user portrait corresponding to the user by using the preset model based on the historical behaviors of the user after the data category labels are added to the plurality of pieces of user information corresponding to each piece of user information comprises the following steps: constructing an intermediate portrait according to the user historical behaviors and the categories of the user historical behaviors; constructing the user representation from the intermediate representation; the constructing an intermediate portrait according to the intermediate portrait and the type of the intermediate portrait specifically includes: calculating the average value of the vector of the intermediate portrait, and expressing the semantic meaning of the intermediate portrait through the average value of the vector; calculating the vector average value of the intermediate image of the same type according to the average value of the vectors of the intermediate image, and taking the vector average value as the intermediate image; when the intermediate representation includes one or more of the types, one or more of the intermediate representations are constructed.
The risk control system and method based on user portrait construction have the following beneficial effects: the user portrait is constructed by using the historical behavior data of the user, so that the accuracy and the scientification of risk control are realized; meanwhile, in the process of constructing the user portrait, an innovative data classification algorithm is adopted, so that the accuracy and the efficiency of constructing the user portrait are improved; the method is mainly realized by the following steps: 1. classification of user data: the inventionBefore constructing the user portrait, user data classification is carried out firstly, and the classification process can improve the efficiency of constructing the user portrait; meanwhile, the classification model used by the invention is based on the following formula:
compared with the data classification method in the prior art, the classification model and the formula consider more variables, and the classification uses a probability-based classification method, so that the classification result is more accurate; 2. constructing a user portrait: in the user portrait construction process, firstly, an intermediate portrait is constructed, the concept of the intermediate portrait is a novel concept originally created by the invention, and the semantics of the intermediate portrait is expressed by calculating the average value of the vector of the intermediate portrait and the average value of the vector; calculating the vector average value of the intermediate images of the same type according to the average value of the vectors of the intermediate images, and taking the vector average value as the intermediate images; constructing one or more intermediate representations when the intermediate representations include one or more of the types; the user portrait constructed is more accurate, and compared with the method for constructing the user portrait in the prior art, the construction efficiency is higher; 3. and (3) risk control: the method comprises the steps of associating a constructed user portrait with preset seed users with different financial risk levels to obtain user portrait data with different financial risk levels; therefore, the user portrait can be hooked with the financial risk level, so that the financial risk control is more scientific and accurate.
Drawings
FIG. 1 is a schematic system diagram of a risk control system constructed based on a user profile according to an embodiment of the present invention;
fig. 2 is a schematic flow chart of a method of risk control based on user profile construction according to an embodiment of the present invention.
Detailed Description
The method of the present invention will be described in further detail below with reference to the accompanying drawings and embodiments of the invention.
Example 1
As shown in fig. 1, a risk control system constructed based on a user profile, the system comprising:
the data processing module is configured to acquire user data, perform data preprocessing on the acquired user data, perform data classification on a result of the user data preprocessing, and add a data classification label to the user data of each classification; the user data at least comprises: the method comprises the following steps of (1) user information and user historical behaviors, wherein the user historical behaviors are only subordinate to one user information; when the data classification is carried out, classifying a plurality of user historical behaviors corresponding to user information based on a preset classification model, and adding a data classification label to each classified user historical behavior; the data category label includes at least: low risk, medium risk and high risk categories;
the user portrait construction module is configured to construct a user portrait corresponding to each user by using a preset model based on a plurality of user historical behaviors which are added with data category labels and correspond to each user information;
the risk control module is configured for associating the constructed user portrait with preset seed users with different financial risk levels to obtain user portrait data with different financial risk levels; and grading the users according to the user image data of different financial risk grades.
By adopting the technical scheme, the user portrait is constructed by utilizing the historical behavior data of the user, so that the accuracy and the scientificity of risk control are realized; meanwhile, in the process of constructing the user portrait, an innovative data classification algorithm is adopted, so that the accuracy and the efficiency of constructing the user portrait are improved; the method is mainly realized by the following steps: 1. classification of user data: according to the method, before the user portrait is constructed, user data classification is firstly carried out, and the efficiency of constructing the user portrait can be improved in the classification process; meanwhile, the classification model used by the invention is based on the following formula:
compared with the data classification method in the prior art, the classification model and the formula consider more variables, and the classification uses a probability-based classification method, so that the classification result is more accurate; 2. constructing a user portrait: in the user portrait construction process, firstly, an intermediate portrait is constructed, the concept of the intermediate portrait is a novel concept originally created by the invention, and the semantics of the intermediate portrait is expressed by calculating the average value of the vector of the intermediate portrait and the average value of the vector; calculating the vector average value of the intermediate images of the same type according to the average value of the vectors of the intermediate images, and taking the vector average value as the intermediate images; constructing one or more intermediate representations when the intermediate representations include one or more of the types; the user portrait constructed is more accurate, and compared with the method for constructing the user portrait in the prior art, the construction efficiency is higher; 3. and (3) risk control: the method comprises the steps of associating a constructed user portrait with preset seed users with different financial risk levels to obtain user portrait data with different financial risk levels; therefore, the user portrait can be hooked with the financial risk level, so that the financial risk control is more scientific and accurate.
Example 2
On the basis of the above embodiment, the data processing module includes: a data acquisition unit configured to acquire user data; the data preprocessing unit is configured to perform data preprocessing on the acquired user data; the data preprocessing process specifically comprises: removing unique attributes, processing missing values, detecting and processing abnormal values, processing data specifications and processing data standardization; the data classification unit is configured to perform data classification on the result of the user data preprocessing and add a data classification label to the user data of each classification; the above-mentionedThe data classification process specifically comprises the following steps: randomly dividing the result of the pre-processing of the user data into two sets T1And T2(ii) a According to said two sets T1And T2Classifying the result data of the user data preprocessing by using a classification model constructed by the following formula:
wherein λ ismThe value range of the weight of the set classification item is as follows: 1-5; g is a classification number, and the value is 3 or 6; the weight value of the v-bit correction term ranges from 0.1 to 0.4;
for gradient operator, H (T) is a correction function, which is set as a linear function, and the parameters of the linear function can be set;
as a set T1And T2Is estimated as a probability density function, pmAs a set T1And T2A probability density function of; delta is an intensity mean value adjusting parameter, and the value range is as follows: 3-5; i is the intensity mean function.
With the above technical solution, before data analysis, data is generally standardized (normalization), and the standardized data is used for data analysis. Data normalization is the indexing of statistical data. The data standardization processing mainly comprises two aspects of data chemotaxis processing and dimensionless processing. The data homochemotaxis processing mainly solves the problem of data with different properties, directly sums indexes with different properties and cannot correctly reflect the comprehensive results of different acting forces, and firstly considers changing the data properties of inverse indexes to ensure that all the indexes are homochemotactic for the acting forces of the evaluation scheme and then sum to obtain correct results. The data dimensionless process mainly addresses the comparability of data. There are many methods for data normalization, and the methods are commonly used, such as "min-max normalization", "Z-score normalization", and "normalization on a decimal scale". Through the standardization processing, the original data are all converted into non-dimensionalized index mapping evaluation values, namely, all index values are in the same quantity level, and comprehensive evaluation analysis can be carried out.
Example 3
On the basis of the above embodiment, the data preprocessing unit, the method for performing data preprocessing on the acquired user data, includes: removing the unique attribute, processing missing values and abnormal value detection and processing; and carrying out data specification processing, including: removing an average value, calculating a covariance matrix, calculating an eigenvalue and an eigenvector of the covariance matrix, sorting the eigenvalues from large to small, reserving the largest k eigenvector, and converting data into a new space constructed by the eigenvector; finally, new processed data are obtained, and the data are irrelevant pairwise, but the original information is kept. And carrying out data standardization processing, and scaling the data to enable the data to fall into a set interval.
Specifically, there are two main approaches to data reduction: attribute selection and data sampling, for attributes and records in the original dataset, respectively.
Assume that data is selected for analysis at the company's data warehouse. So that the data set will be very large. Complex data analysis buckle mining on massive data would take a long time, making such analysis impractical or infeasible.
Data reduction techniques may be used to obtain a reduced representation of a data set that, while small, substantially maintains the integrity of the original data. In this way, mining on the reduced data set will be more efficient and produce the same (or nearly the same) analysis results.
Example 4
On the basis of the above embodiment, the risk control module includes: the risk association unit is configured for associating the constructed user portrait with preset seed users with different financial risk levels to obtain user portrait data with different financial risk levels; and the risk grade establishing unit is configured with a grading model used for training user portrait data based on different financial risk grades to obtain the financial risk grade, and the grading model is used for grading the user on a preset application program to be divided into different financial risk grades.
Specifically, the method comprises the steps of associating seed users with different financial risk levels with user portrait to obtain user portrait data with different financial risk levels, training the user portrait data based on the different financial risk levels to obtain a grading model of the financial risk levels, and dividing users on a preset application program into different financial risk levels by using the grading model; and performing risk scoring according to different financial risk grades corresponding to the users, and only showing preset financial entries to the users with scores meeting preset requirements. The financial risk scoring can be performed on the users on the preset application program, and only the financial entries on the preset application program are displayed for the users with the scoring meeting the requirements, so that the financial success rate of popularizing financial services on the guided preset application program is improved, the situation that the users with the scoring not meeting the requirements, namely inferior customers, can handle finance through the financial entries at will is avoided, the risk of a financial company is reduced, and the conversion rate of financial users is improved.
Example 5
On the basis of the previous embodiment, the method for constructing the user portrait corresponding to the user by the user portrait construction module based on the plurality of user historical behaviors after the data category labels are added corresponding to each piece of user information and by using a preset model comprises the following steps: constructing an intermediate portrait according to the user historical behaviors and the categories of the user historical behaviors; constructing the user representation from the intermediate representation; the constructing an intermediate portrait according to the intermediate portrait and the type of the intermediate portrait specifically includes: calculating the average value of the vector of the intermediate portrait, and expressing the semantic meaning of the intermediate portrait through the average value of the vector; calculating the vector average value of the intermediate image of the same type according to the average value of the vectors of the intermediate image, and taking the vector average value as the intermediate image; when the intermediate representation includes one or more of the types, one or more of the intermediate representations are constructed.
Specifically, the user portrait is also called a user role, and is an effective tool for delineating a target user and connecting user appeal and design direction, and the user portrait is widely applied in various fields. In the practical operation process, the attributes and behaviors of the user are often combined with expected data conversion by the utterances with the most shallow and close to life. As a virtual representation of an actual user, the user roles formed by user portrayal are not constructed outside products and markets, and the formed user roles need to represent the main audience and target groups of the products.
Example 6
A risk control method based on user profile construction, the method performing the steps of:
step 1: acquiring user data, performing data preprocessing on the acquired user data, performing data classification on the result of the user data preprocessing, and adding a data classification label to the user data of each classification; the user data at least comprises: the method comprises the following steps of (1) user information and user historical behaviors, wherein the user historical behaviors are only subordinate to one user information; when the data classification is carried out, classifying a plurality of user historical behaviors corresponding to user information based on a preset classification model, and adding a data classification label to each classified user historical behavior; the data category label includes at least: low risk, medium risk and high risk categories;
step 2: based on a plurality of user historical behaviors which are added with data category labels and correspond to each user information, constructing a user portrait corresponding to the user by using a preset model;
and step 3: associating the constructed user portrait with preset seed users with different financial risk levels to obtain user portrait data with different financial risk levels; and grading the users according to the user image data of different financial risk grades.
Specifically, classification of user data: the invention firstly classifies the user data before constructing the user portrait, and the classification process can be improvedThe efficiency of user portrait construction is improved; meanwhile, the classification model used by the invention is based on the following formula:
compared with the data classification method in the prior art, the classification model and the formula consider more variables, and the classification uses a probability-based classification method, so that the classification result is more accurate.
Example 7
On the basis of the above embodiment, in the step 1: the method for preprocessing the acquired user data comprises the following steps: removing the unique attribute, processing missing values and abnormal value detection and processing; and carrying out data specification processing, including: removing an average value, calculating a covariance matrix, calculating an eigenvalue and an eigenvector of the covariance matrix, sorting the eigenvalues from large to small, reserving the largest eigenvector, and converting data into a new space constructed by the eigenvector; finally, new processed data are obtained, and the data are irrelevant pairwise, but the original information is kept. And carrying out data standardization processing, and scaling the data to enable the data to fall into a set interval.
In the user portrait construction process, firstly, an intermediate portrait is constructed, the concept of the intermediate portrait is a novel concept originally created by the invention, and the semantics of the intermediate portrait is expressed by calculating the average value of the vector of the intermediate portrait and the average value of the vector; calculating the vector average value of the intermediate images of the same type according to the average value of the vectors of the intermediate images, and taking the vector average value as the intermediate images; constructing one or more intermediate representations when the intermediate representations include one or more of the types; the user portrait constructed is more accurate, and compared with the method for constructing the user portrait in the prior art, the construction efficiency is higher.
Example 8
On the basis of the above embodiment, in the step 1: the method for classifying the data of the result of the user data preprocessing and adding the data category label to the user data of each category comprises the following steps: randomly dividing the result of the pre-processing of the user data into two sets T1And T2(ii) a According to said two sets T1And T2Classifying the result data of the user data preprocessing by using a classification model constructed by the following formula:
wherein λ ismThe value range of the weight of the set classification item is as follows: 1-5; g is a classification number, and the value is 3 or 6; the weight value of the v-bit correction term ranges from 0.1 to 0.4;
for gradient operator, H (T) is a correction function, which is set as a linear function, and the parameters of the linear function can be set;
as a set T1And T2Is estimated as a probability density function, pmAs a set T1And T2A probability density function of; delta is an intensity mean value adjusting parameter, and the value range is as follows: 3-5; i is an intensity mean function; for the classified user data, if the classified user data still belongs to the set T1Then add a high risk label; if it is classified, it still belongs to the set T2Then add a low risk label; if classified, the group does not belong to the set T1(ii) a Nor to the set T2Then a tag of medium risk is added.
The method comprises the steps of associating a constructed user portrait with preset seed users with different financial risk levels to obtain user portrait data with different financial risk levels; therefore, the user portrait can be hooked with the financial risk level, so that the financial risk control is more scientific and accurate.
Example 9
On the basis of the above embodiment, the step 3 specifically includes: associating the constructed user portrait with preset seed users with different financial risk levels to obtain user portrait data with different financial risk levels; training user portrait data based on different financial risk levels to obtain a grading model of the financial risk levels, and grading users on a preset application program by using the grading model to obtain different financial risk levels.
Example 10
On the basis of the above embodiment, the step 2: the method for constructing the user portrait corresponding to the user by using the preset model based on the historical behaviors of the user after the data category labels are added to the plurality of pieces of user information corresponding to each piece of user information comprises the following steps: constructing an intermediate portrait according to the user historical behaviors and the categories of the user historical behaviors; constructing the user representation from the intermediate representation; the constructing an intermediate portrait according to the intermediate portrait and the type of the intermediate portrait specifically includes: calculating the average value of the vector of the intermediate portrait, and expressing the semantic meaning of the intermediate portrait through the average value of the vector; calculating the vector average value of the intermediate image of the same type according to the average value of the vectors of the intermediate image, and taking the vector average value as the intermediate image; when the intermediate representation includes one or more of the types, one or more of the intermediate representations are constructed.
It can be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working process and related description of the system described above may refer to the corresponding process in the foregoing method embodiments, and will not be described herein again.
It should be noted that, the system provided in the foregoing embodiment is only illustrated by dividing the functional units, and in practical applications, the functions may be distributed by different functional units according to needs, that is, the units or steps in the embodiments of the present invention are further decomposed or combined, for example, the units in the foregoing embodiment may be combined into one unit, or may be further decomposed into multiple sub-units, so as to complete all or the functions of the units described above. The names of the units and steps involved in the embodiments of the present invention are only for distinguishing the units or steps, and are not to be construed as unduly limiting the present invention.
It can be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working processes and related descriptions of the storage device and the processing device described above may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
Those of skill in the art would appreciate that the various illustrative elements, method steps, described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that programs corresponding to the elements, method steps may be located in Random Access Memory (RAM), memory, Read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. To clearly illustrate this interchangeability of electronic hardware and software, various illustrative components and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as electronic hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
The terms "first," "second," and the like are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order.
The terms "comprises," "comprising," or any other similar term are intended to cover a non-exclusive inclusion, such that a process, method, article, or unit/apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or unit/apparatus.
So far, the technical solutions of the present invention have been described in connection with the preferred embodiments shown in the drawings, but it is easily understood by those skilled in the art that the scope of the present invention is obviously not limited to these specific embodiments. Equivalent modifications or substitutions of the related art marks may be made by those skilled in the art without departing from the principle of the present invention, and the technical solutions after such modifications or substitutions will fall within the protective scope of the present invention.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention.
Claims (10)
1. A risk control system constructed based on a user representation, the system comprising:
the data processing module is configured to acquire user data, perform data preprocessing on the acquired user data, perform data classification on a result of the user data preprocessing, and add a data classification label to the user data of each classification; the user data at least comprises: the method comprises the following steps of (1) user information and user historical behaviors, wherein the user historical behaviors are only subordinate to one user information; when the data classification is carried out, classifying a plurality of user historical behaviors corresponding to user information based on a preset classification model, and adding a data classification label to each classified user historical behavior; the data category label includes at least: low risk, medium risk and high risk categories;
the user portrait construction module is configured to construct a user portrait corresponding to each user by using a preset model based on a plurality of user historical behaviors which are added with data category labels and correspond to each user information;
the risk control module is configured for associating the constructed user portrait with preset seed users with different financial risk levels to obtain user portrait data with different financial risk levels; and grading the users according to the user image data of different financial risk grades.
2. The system of claim 1, wherein the data processing module comprises: a data acquisition unit configured to acquire user data; the data preprocessing unit is configured to perform data preprocessing on the acquired user data; the data preprocessing process specifically comprises: removing unique attributes, processing missing values, detecting and processing abnormal values, processing data specifications and processing data standardization; the data classification unit is configured to perform data classification on the result of the user data preprocessing and add a data classification label to the user data of each classification; the data classification process specifically includes: randomly dividing the result of the pre-processing of the user data into two sets T1And T2(ii) a According to said two sets T1And T2Classifying the result data of the user data preprocessing by using a classification model constructed by the following formula:
wherein λ ismThe value range of the weight of the set classification item is as follows: 1-5; g is a classification number, and the value is 3 or 6; the weight value of the v-bit correction term ranges from 0.1 to 0.4;
for gradient operator, H (T) is a correction function, which is set as a linear function, and the parameters of the linear function can be set;
as a set T1And T2Is estimated as a probability density function, pmAs a set T1And T2Function of probability densityCounting; delta is an intensity mean value adjusting parameter, and the value range is as follows: 3-5; i is the intensity mean function.
3. The system of claim 2, wherein the data pre-processing unit, the method of data pre-processing the acquired user data, comprises: removing the unique attribute, processing missing values and abnormal value detection and processing; and carrying out data specification processing, including: removing an average value, calculating a covariance matrix, calculating an eigenvalue and an eigenvector of the covariance matrix, sorting the eigenvalues from large to small, reserving the largest k eigenvector, and converting data into a new space constructed by the eigenvector; finally, new processed data are obtained, and the data are irrelevant pairwise, but the original information is kept. And carrying out data standardization processing, and scaling the data to enable the data to fall into a set interval.
4. The system of claim 3, wherein the risk control module comprises: the risk association unit is configured for associating the constructed user portrait with preset seed users with different financial risk levels to obtain user portrait data with different financial risk levels; and the risk grade establishing unit is configured with a grading model used for training user portrait data based on different financial risk grades to obtain the financial risk grade, and the grading model is used for grading the user on a preset application program to be divided into different financial risk grades.
5. The system of claim 4, wherein the user representation construction module, based on the historical behavior of the user after adding the data category labels to the plurality of data category labels corresponding to each user information, uses a preset model to construct the user representation corresponding to the user, and comprises: constructing an intermediate portrait according to the user historical behaviors and the categories of the user historical behaviors; constructing the user representation from the intermediate representation; the constructing an intermediate portrait according to the intermediate portrait and the type of the intermediate portrait specifically includes: calculating the average value of the vector of the intermediate portrait, and expressing the semantic meaning of the intermediate portrait through the average value of the vector; calculating the vector average value of the intermediate image of the same type according to the average value of the vectors of the intermediate image, and taking the vector average value as the intermediate image; when the intermediate representation includes one or more of the types, one or more of the intermediate representations are constructed.
6. A method for risk control based on user profile construction based on a system according to any of claims 1 to 5, characterized in that the method performs the following steps:
step 1: acquiring user data, performing data preprocessing on the acquired user data, performing data classification on the result of the user data preprocessing, and adding a data classification label to the user data of each classification; the user data at least comprises: the method comprises the following steps of (1) user information and user historical behaviors, wherein the user historical behaviors are only subordinate to one user information; when the data classification is carried out, classifying a plurality of user historical behaviors corresponding to user information based on a preset classification model, and adding a data classification label to each classified user historical behavior; the data category label includes at least: low risk, medium risk and high risk categories;
step 2: based on a plurality of user historical behaviors which are added with data category labels and correspond to each user information, constructing a user portrait corresponding to the user by using a preset model;
and step 3: associating the constructed user portrait with preset seed users with different financial risk levels to obtain user portrait data with different financial risk levels; and grading the users according to the user image data of different financial risk grades.
7. The method of claim 6, wherein in step 1: the method for preprocessing the acquired user data comprises the following steps: removing the unique attribute, processing missing values and abnormal value detection and processing; and carrying out data specification processing, including: removing an average value, calculating a covariance matrix, calculating an eigenvalue and an eigenvector of the covariance matrix, sorting the eigenvalues from large to small, reserving the largest eigenvector, and converting data into a new space constructed by the eigenvector; finally, new processed data are obtained, and the data are irrelevant pairwise, but the original information is kept. And carrying out data standardization processing, and scaling the data to enable the data to fall into a set interval.
8. The method of claim 7, wherein in step 1: the method for classifying the data of the result of the user data preprocessing and adding the data category label to the user data of each category comprises the following steps: randomly dividing the result of the pre-processing of the user data into two sets T1And T2(ii) a According to said two sets T1And T2Classifying the result data of the user data preprocessing by using a classification model constructed by the following formula:
wherein λ ismThe value range of the weight of the set classification item is as follows: 1-5; g is a classification number, and the value is 3 or 6; the weight value of the v-bit correction term ranges from 0.1 to 0.4;
for gradient operator, H (T) is a correction function, which is set as a linear function, and the parameters of the linear function can be set;
as a set T1And T2Is estimated as a probability density function, pmAs a set T1And T2A probability density function of; delta is an intensity mean value adjusting parameter, and the value range is as follows: 3-5; i is an intensity mean function; for the classified user data, if the classified user data still belongs to the set T1Then add a high risk label; if it is classified, thenBut belongs to the set T2Then add a low risk label; if classified, the group does not belong to the set T1(ii) a Nor to the set T2Then a tag of medium risk is added.
9. The method according to claim 8, wherein the step 3 specifically comprises: associating the constructed user portrait with preset seed users with different financial risk levels to obtain user portrait data with different financial risk levels; training user portrait data based on different financial risk levels to obtain a grading model of the financial risk levels, and grading users on a preset application program by using the grading model to obtain different financial risk levels.
10. The method of claim 9, wherein the step 2: the method for constructing the user portrait corresponding to the user by using the preset model based on the historical behaviors of the user after the data category labels are added to the plurality of pieces of user information corresponding to each piece of user information comprises the following steps: constructing an intermediate portrait according to the user historical behaviors and the categories of the user historical behaviors; constructing the user representation from the intermediate representation; the constructing an intermediate portrait according to the intermediate portrait and the type of the intermediate portrait specifically includes: calculating the average value of the vector of the intermediate portrait, and expressing the semantic meaning of the intermediate portrait through the average value of the vector; calculating the vector average value of the intermediate image of the same type according to the average value of the vectors of the intermediate image, and taking the vector average value as the intermediate image; when the intermediate representation includes one or more of the types, one or more of the intermediate representations are constructed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011149392.XA CN112270596A (en) | 2020-10-23 | 2020-10-23 | Risk control system and method based on user portrait construction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011149392.XA CN112270596A (en) | 2020-10-23 | 2020-10-23 | Risk control system and method based on user portrait construction |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112270596A true CN112270596A (en) | 2021-01-26 |
Family
ID=74341863
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011149392.XA Withdrawn CN112270596A (en) | 2020-10-23 | 2020-10-23 | Risk control system and method based on user portrait construction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112270596A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112862582A (en) * | 2021-02-18 | 2021-05-28 | 深圳无域科技技术有限公司 | User portrait generation system and method based on financial wind control |
| CN112991077A (en) * | 2021-02-18 | 2021-06-18 | 深圳无域科技技术有限公司 | Financial risk control system and method |
| CN113571157A (en) * | 2021-04-20 | 2021-10-29 | 杭州袋虎信息技术有限公司 | Intelligent risk person psychological image recognition system based on FMT characteristics |
| CN115146155A (en) * | 2022-06-28 | 2022-10-04 | 广东圣火传媒科技股份有限公司 | Dynamic user portrait management system |
| CN117033729A (en) * | 2023-08-08 | 2023-11-10 | 北京健康在线技术开发有限公司 | Doctor imaging method, apparatus, device and storage medium for evaluating doctor ability |
-
2020
- 2020-10-23 CN CN202011149392.XA patent/CN112270596A/en not_active Withdrawn
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112862582A (en) * | 2021-02-18 | 2021-05-28 | 深圳无域科技技术有限公司 | User portrait generation system and method based on financial wind control |
| CN112991077A (en) * | 2021-02-18 | 2021-06-18 | 深圳无域科技技术有限公司 | Financial risk control system and method |
| CN112991077B (en) * | 2021-02-18 | 2024-03-19 | 深圳无域科技技术有限公司 | Financial risk control system and method |
| CN112862582B (en) * | 2021-02-18 | 2024-03-22 | 深圳无域科技技术有限公司 | User portrait generation system and method based on financial wind control |
| CN113571157A (en) * | 2021-04-20 | 2021-10-29 | 杭州袋虎信息技术有限公司 | Intelligent risk person psychological image recognition system based on FMT characteristics |
| CN115146155A (en) * | 2022-06-28 | 2022-10-04 | 广东圣火传媒科技股份有限公司 | Dynamic user portrait management system |
| CN115146155B (en) * | 2022-06-28 | 2023-08-25 | 广东圣火传媒科技股份有限公司 | Dynamic user portrayal management system |
| CN117033729A (en) * | 2023-08-08 | 2023-11-10 | 北京健康在线技术开发有限公司 | Doctor imaging method, apparatus, device and storage medium for evaluating doctor ability |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112270596A (en) | Risk control system and method based on user portrait construction | |
| Keramati et al. | A proposed classification of data mining techniques in credit scoring | |
| CN109739844B (en) | Data classification method based on attenuation weight | |
| CN111861697B (en) | Loan multi-head data-based user portrait generation method and system | |
| CN112700324A (en) | User loan default prediction method based on combination of Catboost and restricted Boltzmann machine | |
| CN111340086A (en) | Method, system, medium and terminal for processing label-free data | |
| Fan et al. | Improved ML-based technique for credit card scoring in internet financial risk control | |
| CN111783829A (en) | Financial anomaly detection method and device based on multi-label learning | |
| El Seddawy et al. | Enhanced K-mean algorithm to improve decision support system under uncertain situations | |
| CN113762151A (en) | Fault data processing method and system and fault prediction method | |
| Zhang et al. | A precision marketing strategy of e-commerce platform based on consumer behavior analysis in the era of big data | |
| CN115797044B (en) | Credit wind control early warning method and system based on cluster analysis | |
| CN112506930B (en) | Data insight system based on machine learning technology | |
| CN115293641A (en) | Enterprise risk intelligent identification method based on financial big data | |
| CN114066173A (en) | Capital flow behavior analysis method and storage medium | |
| WO1992017853A2 (en) | Direct data base analysis, forecasting and diagnosis method | |
| Quan | Stock prediction by searching similar candlestick charts | |
| Kulothungan | Loan Forecast by Using Machine Learning | |
| Şişci et al. | Machine Learning as a Tool for Achieving Digital Transformation | |
| CN112348220A (en) | Credit risk assessment prediction method and system based on enterprise behavior pattern | |
| CN115953166B (en) | Customer information management method and system based on big data intelligent matching | |
| CN112926816B (en) | Vendor evaluation method, device, computer device and storage medium | |
| CN115187387B (en) | Identification method and equipment for risk merchant | |
| Naidovich et al. | Survival analysis in credit scoring | |
| Zeng et al. | Classification of Stock Information Disclosure Level Based on SVM and CNN |
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 | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20210126 |