CN111709844A - Insurance money laundering personnel detection method and device and computer readable storage medium - Google Patents
Insurance money laundering personnel detection method and device and computer readable storage medium Download PDFInfo
- Publication number
- CN111709844A CN111709844A CN202010404117.1A CN202010404117A CN111709844A CN 111709844 A CN111709844 A CN 111709844A CN 202010404117 A CN202010404117 A CN 202010404117A CN 111709844 A CN111709844 A CN 111709844A
- Authority
- CN
- China
- Prior art keywords
- money laundering
- insurance
- preset
- information
- business
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
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
- G06Q40/08—Insurance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Abstract
The embodiment of the application discloses a method and a device for detecting insurance money laundering personnel and a computer readable storage medium, wherein the method comprises the following steps: acquiring an insurance business feature matrix related to one or more insurance business information and a trained insurance money laundering detection model; inputting the insurance business feature matrix into an insurance money laundering detection model; taking the output result of the insurance money laundering detection model as the detection result of whether the insurance business information relates to insurance money laundering; determining the related personnel determined as insurance business information related to the insurance money laundering as suspects of the insurance money laundering; the insurance money laundering detection model is obtained by training a preset machine learning model by using a money laundering business feature matrix containing preset money laundering features as training data. The scheme of the embodiment does not depend on expert rules, effectively screens abnormal money laundering personnel, improves the working efficiency, reduces the difficulty of investigation, quickly processes mass data, and greatly overcomes the defects of the traditional screening mode.
Description
Technical Field
The present disclosure relates to criminal screening technologies, and more particularly, to a method and apparatus for detecting an insurance money laundering employee, and a computer-readable storage medium.
Background
Many lawbreakers' money laundering means have spread to the basic insurance field and begin to erode the financial insurance system, and because the anti-money laundering risk awareness of the basic insurance organization is poor, and many hidden risks exist in the internal control mechanism and the business process, it is necessary to accurately identify the money laundering risk.
The existing insurance field money laundering business is developed mainly by accumulating artificial experiences for years and carrying out level division on insurance customers through a plurality of rules summarized manually, the rules of customers with different levels are slightly different, and the customers are further screened and then manually screened.
The prior art scheme depends heavily on expert rules, the rules are small and limited, the rules are difficult to apply in a big data scene, a screened result set is still huge, and a great deal of manual energy is still needed. Moreover, the expert rules are only applicable to partial regions and specific crowds, the expert opinions in different regions are different, the rules may be different, and the rules are difficult to reuse.
Disclosure of Invention
The embodiment of the application provides a method and a device for detecting insurance money laundering personnel and a computer readable storage medium, which can accurately identify money laundering risks without depending on expert rules, effectively screen abnormal money laundering personnel, improve the working efficiency, reduce the troubleshooting difficulty, quickly process mass data and greatly overcome the defects of the traditional screening mode.
The embodiment of the application provides a method for detecting an insurance money laundering worker, which can comprise the following steps:
acquiring an insurance business feature matrix related to one or more insurance business information and a trained insurance money laundering detection model; the insurance service characteristic matrix is a matrix used for representing corresponding information of preset service characteristics in each piece of insurance service information;
inputting the insurance business feature matrix into the insurance money laundering detection model as input data;
taking the output result of the insurance money laundering detection model as the detection result of whether the one or more insurance business information relates to insurance money laundering;
determining the related personnel determined as insurance business information related to the insurance money laundering as suspects of the insurance money laundering; the relevant persons include any one or more of: policemen, insureds, and beneficiaries;
the insurance money laundering detection model is obtained by training a preset machine learning model by taking a money laundering business feature matrix containing preset money laundering features as training data.
The embodiment of the present application further provides an insurance money laundering staff detection apparatus, which may include a processor and a computer-readable storage medium, where instructions are stored in the computer-readable storage medium, and when the instructions are executed by the processor, the insurance money laundering staff detection method described in any one of the above is implemented.
The embodiment of the application also provides a computer readable storage medium, wherein a computer program is stored on the computer readable storage medium, and when being executed by a processor, the computer program realizes the insurance money laundering staff detection method.
Compared with the related art, the embodiment of the application comprises the following steps: acquiring an insurance business feature matrix related to one or more insurance business information and a trained insurance money laundering detection model; the insurance service characteristic matrix is a matrix used for representing corresponding information of preset service characteristics in each piece of insurance service information; inputting the insurance business feature matrix into the insurance money laundering detection model as input data; taking the output result of the insurance money laundering detection model as the detection result of whether the one or more insurance business information relates to insurance money laundering; determining the related personnel determined as insurance business information related to the insurance money laundering as suspects of the insurance money laundering; the relevant persons include any one or more of: policemen, insureds, and beneficiaries; the insurance money laundering detection model is obtained by training a preset machine learning model by taking a money laundering business feature matrix containing preset money laundering features as training data. Through the scheme of the embodiment, the money laundering risk can be accurately identified based on the trained machine learning model without depending on expert rules, abnormal money laundering personnel are effectively screened, the working efficiency is improved, the troubleshooting difficulty is reduced, mass data are rapidly processed, and the defects of the traditional screening mode are greatly overcome.
Additional features and advantages of the application will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of the application. Other advantages of the present application may be realized and attained by the instrumentalities and combinations particularly pointed out in the specification and the drawings.
Drawings
The accompanying drawings are included to provide an understanding of the present disclosure and are incorporated in and constitute a part of this specification, illustrate embodiments of the disclosure and together with the examples serve to explain the principles of the disclosure and not to limit the disclosure.
FIG. 1 is a flow chart of an insurance money laundering employee detection method according to an embodiment of the present application;
FIG. 2 is a flowchart of a method for obtaining an insurance money laundering detection model according to an embodiment of the present application;
FIG. 3 is a schematic diagram of a method for obtaining an insurance money laundering detection model according to an embodiment of the present application;
FIG. 4 is a block diagram of the detection apparatus for insurance money laundering personnel according to the embodiment of the present application.
Detailed Description
The present application describes embodiments, but the description is illustrative rather than limiting and it will be apparent to those of ordinary skill in the art that many more embodiments and implementations are possible within the scope of the embodiments described herein. Although many possible combinations of features are shown in the drawings and discussed in the detailed description, many other combinations of the disclosed features are possible. Any feature or element of any embodiment may be used in combination with or instead of any other feature or element in any other embodiment, unless expressly limited otherwise.
The present application includes and contemplates combinations of features and elements known to those of ordinary skill in the art. The embodiments, features and elements disclosed in this application may also be combined with any conventional features or elements to form a unique inventive concept as defined by the claims. Any feature or element of any embodiment may also be combined with features or elements from other inventive aspects to form yet another unique inventive aspect, as defined by the claims. Thus, it should be understood that any of the features shown and/or discussed in this application may be implemented alone or in any suitable combination. Accordingly, the embodiments are not limited except as by the appended claims and their equivalents. Furthermore, various modifications and changes may be made within the scope of the appended claims.
Further, in describing representative embodiments, the specification may have presented the method and/or process as a particular sequence of steps. However, to the extent that the method or process does not rely on the particular order of steps set forth herein, the method or process should not be limited to the particular sequence of steps described. Other orders of steps are possible as will be understood by those of ordinary skill in the art. Therefore, the particular order of the steps set forth in the specification should not be construed as limitations on the claims. Further, the claims directed to the method and/or process should not be limited to the performance of their steps in the order written, and one skilled in the art can readily appreciate that the sequences may be varied and still remain within the spirit and scope of the embodiments of the present application.
The embodiment of the application provides a method for detecting an insurance money laundering worker, and as shown in fig. 1, the method may include steps S101 to S104:
s101, acquiring an insurance business feature matrix related to one or more insurance business information and a trained insurance money laundering detection model; the insurance business characteristic matrix is a matrix used for representing corresponding information of preset business characteristics in each piece of insurance business information; the insurance money laundering detection model is obtained by training a preset machine learning model by taking a money laundering business feature matrix containing preset money laundering features as training data.
In the exemplary embodiment of the application, aiming at the anti-money laundering business in the insurance field, a method for identifying abnormal people based on a machine learning model is provided, an unsupervised mode (such as an isolated forest algorithm) in machine learning is adopted, a large number of artificial rules are not needed, some specific incidence relations can be learned from data, and people with high abnormal money laundering risks can be excavated.
In an exemplary embodiment of the present application, the embodiment scheme may input a large amount of insurance business information into a trained insurance money laundering detection model based on a large amount of insurance business daily processed by an insurance enterprise, so as to screen money laundering business.
In an exemplary embodiment of the present application, to facilitate identification of the insurance money laundering detection model, the volume of insurance business information may be converted into an insurance business feature matrix relating to one or more pieces of insurance business information as input data to the insurance money laundering detection model.
In an exemplary embodiment of the present application, the obtaining of the insurance service feature matrix regarding one or more pieces of insurance service information may include:
acquiring one or more insurance service information, and extracting one or more preset characteristics from the one or more insurance service information;
listing information of one or more preset characteristics corresponding to each insurance service information in a vector form according to a preset sequence to obtain an insurance service characteristic vector corresponding to each insurance service information;
and forming the insurance service characteristic matrix by all insurance service characteristic vectors corresponding to all insurance service information.
In an exemplary embodiment of the present application, the insurance service information may include, but is not limited to, any one or more of the following: information on the policy information table, information on the applicant information table, information on the insured person information table, information on the beneficiary information table, and information on the policy transaction record table.
In an exemplary embodiment of the present application, useful information for money laundering risk identification, i.e., the above-described preset features, may be extracted from the above-described various information tables.
In exemplary embodiments of the present application, the preset features may include, but are not limited to, any one or more of the following:
the policy is selected from the group consisting of a form of failure, a check-out transaction time, a check-out loss accumulated amount, a non-related person check-out accumulated acquisition amount of insurance service information, a consistency of a check-out bank card number and a transaction bank card number, a policy amount of cash payment, a total number of policy applicant changes, a policy applicant change time difference, a policy service change number of whether a policy applicant change time period is before and after holidays and within a preset time period (for example, within 3 months), an applicant age, applicant nationality, applicant occupation, and a consistency of an applicant and a check-out transaction person.
In an exemplary embodiment of the present application, after acquiring the preset features of one or more pieces of insurance business information, the preset features may be preprocessed to acquire an insurance business feature vector as an input of an insurance money laundering detection model, so as to identify whether one or more pieces of insurance business information are business information related to insurance money laundering, thereby facilitating to list related personnel of the insurance business information as money laundering criminal suspects, and thus, a target may be locked, and further detection and detection may be facilitated.
In an exemplary embodiment of the present application, the preprocessing of the preset features may include: feature vectorization and normalization operations.
In an exemplary embodiment of the present application, the feature vectorizing the preset features may include: listing information of one or more preset characteristics corresponding to each insurance service information in a vector form according to a preset sequence to obtain an insurance service characteristic vector corresponding to each insurance service information; and forming an insurance service characteristic matrix by using all insurance service characteristic vectors corresponding to all insurance service information.
In an exemplary embodiment of the present application, normalizing the preset features may include:
carrying out non-dimensionalization operation on the insurance service characteristic matrix, and removing the maximum value max and the minimum value min in each column in the insurance service characteristic matrix to obtain the characteristic value y of each columni：
In an exemplary embodiment of the present application, for information of one or more preset features corresponding to each piece of insurance service information, a corresponding insurance service feature vector may be obtained, multiple pieces of insurance service information correspond to multiple insurance service feature vectors, and the multiple insurance service feature vectors may be combined into an insurance service feature matrix.
In the exemplary embodiment of the present application, since the measurement criteria of the feature dimensions are different, and the value ranges of the feature values of different dimensions are also different, a de-dimensioning operation is performed. And the characteristic value can be normalized by a maximum and minimum normalization method. That is, for each column of features of the insurance service feature matrix, the maximum value max in the feature column and the minimum value min in the feature column are removed, and the feature value y of each column is obtainedi：
In an exemplary embodiment of the present application, obtaining a trained insurance money laundering detection model may include:
calling a pre-created and trained insurance money laundering detection model; alternatively, the first and second electrodes may be,
and establishing a machine learning model, training the machine learning model, and acquiring the insurance money laundering detection model.
In the exemplary embodiment of the present application, in the process of money laundering risk screening, a trained insurance money laundering detection model may be directly called, or a machine learning model may be temporarily established and trained to obtain one insurance money laundering detection model.
In an exemplary embodiment of the present application, as shown in fig. 2 and 3, the creating a machine learning model and training the machine learning model, and the obtaining the insurance money laundering detection model may include steps S201 to S203:
s201, obtaining a money laundering insurance business information set which is determined to be related to an insurance money laundering business, obtaining one or more preset money laundering characteristics from the money laundering insurance business information set, and forming a money laundering business characteristic matrix by the one or more preset money laundering characteristics as training data.
In an exemplary embodiment of the present application, training data (or training data sets) may be constructed by extracting one or more preset money laundering characteristics based on a pre-saved money laundering insurance business information set that has been determined to be involved in an insurance money laundering business.
In an exemplary embodiment of the present application, the money laundering characteristic may be an information characteristic having money laundering behavior direction that does not conform to normal insurance logic in insurance service information corresponding to a copy of money laundering insurance service.
In exemplary embodiments of the present application, the preset money laundering characteristics may include any one or more of:
the current form is a failure form;
the time length difference between the release transaction time and the application transaction time is smaller than a preset time length threshold value;
the accumulated refund loss sum is greater than or equal to a preset first sum threshold;
the accumulated acquired amount of the refund of the irrelevant personnel of the insurance service information is greater than or equal to a preset second amount threshold;
the number of the refund bank card is inconsistent with the number of the transacted bank card;
the insurance policy amount paid by cash is greater than or equal to a preset third amount threshold value;
the total change times of the policy applicant is greater than or equal to a preset first time threshold value;
the policy service change times in the preset time length are greater than or equal to a preset second time threshold;
the policy applicant changes the time difference to be less than or equal to the preset time difference threshold value;
the policy applicant changes the time period before and after the holiday;
the age of the applicant is greater than a preset age threshold;
the state of the applicant is a foreign state;
the employment enterprises of the applicant belong to the preset money laundering high-risk enterprise range; and the number of the first and second groups,
the policyholder is inconsistent with the retirement transactor.
In the exemplary embodiment of the present application, the preset money laundering characteristics can be set and combined according to different requirements and different application scenarios, and no detailed limitation is made on specific contents.
In an exemplary embodiment of the present application, the forming of the money laundering transaction feature matrix from one or more preset money laundering features may include:
acquiring a money laundering business feature vector with the money laundering feature corresponding to each piece of money laundering insurance business information, and forming a money laundering business feature matrix by all money laundering business feature vectors corresponding to all money laundering insurance business information;
the money laundering service feature vector is information for listing one or more money laundering features corresponding to each money laundering insurance service information in a vector form according to a preset sequence.
The method further comprises the following steps: preprocessing the money laundering business feature matrix before the money laundering business feature matrix is used as the training data; the pretreatment comprises the following steps: missing value supplementation and normalization operations;
supplementing the missing value of the money laundering business feature matrix comprises:
filling missing value parts of each money laundering characteristic with missing value in the money laundering business characteristic matrix by adopting any one or more of the following modes: mean filling, mode filling, median filling and 0-filling;
the normalization operation of the money laundering business feature matrix comprises the following steps:
carrying out dimensionless operation on the money laundering service characteristic matrix, and removing the maximum value MAX and the minimum value MIN in each column in the money laundering service characteristic matrix to obtain the characteristic value x of each columni：
In an exemplary embodiment of the present application, for one or more preset money laundering characteristics corresponding to each piece of money laundering insurance business information, a corresponding money laundering business characteristic vector may be obtained, multiple pieces of money laundering insurance business information correspond to multiple money laundering business characteristic vectors, and the multiple money laundering business characteristic vectors may be combined into a money laundering business characteristic matrix.
In an exemplary embodiment of the present application, the constructed money laundering business feature matrix may be the following matrix a, which may be represented as:
wherein f isaiAnd fmiMoney laundering features representing different dimensions (i.e., different types of money laundering features), faiA value (which may be a specific numerical value, such as a representative numerical value of answers of days, "yes" or "no", etc., and may also be a specific answer description) of an nth sample (nth money laundering insurance business information) in a dimension a (category a money laundering characteristic); i is a natural number from 1 to n, n is dataTotal number of samples collected (i.e., total number of money laundering insurance business information).
In an exemplary embodiment of the present application, for each obtained money laundering policy transaction record (corresponding to a piece of money laundering insurance business information, respectively), a corresponding money laundering business feature vector can be obtained through money laundering feature extraction. However, the characteristic value of the money laundering characteristic corresponding to each money laundering policy transaction record may be missing, and the missing value part needs to be complemented. The missing value can be completed by any one or more of the following methods:
1. mean value filling: calculating an average value through the part of the non-missing values in a money laundering characteristic column, and completing the average value to the part of the missing values;
2. mode filling: calculating a mode through a part of the money laundering characteristic column which is not the missing value, and complementing the mode to the missing value part;
3. filling the median: calculating a median from the part of the money laundering characteristic column which is not the missing value, and complementing the median to the missing value part;
4. filling with 0: 0 is filled directly into all missing value parts.
In the exemplary embodiment of the present application, which filling manner is specifically adopted may be selected according to the missing value padding effect of the last experiment.
In an exemplary embodiment of the present application, a feature normalization operation may be performed next to the money laundering transaction feature matrix populated with missing values. Because the measurement standards of the feature dimensions are different, and the value ranges of the feature values of different dimensions are also different, the de-dimensionalization operation is performed. And the characteristic value can be normalized by a maximum and minimum normalization method. That is, for each column of features of the money laundering transaction feature matrix, the maximum value MAX and the minimum value MIN in the feature column are removed, and the feature value x of each column is obtainedi：
In an exemplary embodiment of the present application, the method may further include:
after normalization operation is carried out on the money laundering service characteristic matrix, calculating the correlation between any two characteristic dimensions in the money laundering service characteristic matrix through a preset correlation calculation formula; and removing any one of the two characteristic dimensions of which the correlation calculation result is greater than a preset correlation threshold value, so as to realize the screening of the money laundering service characteristic matrix.
In the exemplary embodiment of the present application, if the machine learning model is constructed, the efficiency of calculation may be affected and even the prediction accuracy may be affected due to the large feature dimension in the money laundering business feature matrix. Some features in the money laundering business feature matrix may contain the same or repeated information, so that feature dimensions are large, and therefore, after the machine learning model is obtained, a part of features with large correlation can be discarded through feature correlation analysis and selection.
In an exemplary embodiment of the present application, the preset correlation calculation formula (i.e., the feature correlation analysis method) may include: the following calculation of the Pearson correlation coefficient corr (i.e. by calculating the characteristic dimension fiAnd a characteristic dimension fjCorrelation corr) of (a):
wherein, cov (f)i，fj) Representing a characteristic dimension fiAnd a characteristic dimension fjThe covariance of (a) of (b),
representing a characteristic dimension fiThe standard deviation of (a) is determined,
representing a characteristic dimension fjE (X) represents the mean value of X.
In an exemplary embodiment of the present application, for the pearson correlation coefficient corr between two feature dimensions in the money laundering service feature matrix, if the pearson correlation coefficient corr of any two feature dimensions is calculated to be greater than a preset correlation threshold, for example, 0.85, only one feature dimension of the two feature dimensions may be retained.
In an exemplary embodiment of the present application, after the money laundering business feature matrix performs the above-mentioned correlation calculation, the data in the money laundering business feature matrix obtained finally may be used as training data for training the insurance money laundering detection model.
S202, a machine learning model is constructed by adopting a preset machine learning algorithm.
In an exemplary embodiment of the present application, the machine learning algorithm may be an unsupervised isolated forest algorithm in machine learning.
In the exemplary embodiment of the present application, step S201 and step S202 may be performed simultaneously without being separated from each other, or any step may be performed first, and the order of the two steps is not limited.
S203, training the machine learning model by adopting the training data to obtain the insurance money laundering detection model.
In an exemplary embodiment of the application, the constructed money laundering business feature matrix is input into an insurance money laundering detection model based on an isolated forest algorithm, the isolated forest algorithm can train and learn information data in the money laundering business feature matrix by constructing an isolated tree (iTree), calculate abnormal points under different feature dimensions and different density distributions, score abnormal values of each row of transaction records input into the money laundering business feature matrix, and give abnormal value scores of each transaction record (the numerical range can be between 0 and 1, and the higher the score is, the more abnormal the score is). Finally, the first few results (for example, 100) with the highest abnormal scores are output as a prediction list (a money-washing related insurance policy list) for further verification by public security authorities.
And S102, inputting the insurance business feature matrix into the insurance money laundering detection model as input data.
In an exemplary embodiment of the present application, after acquiring an insurance business feature matrix (including one or more preset features) and an insurance money laundering detection model regarding one or more insurance business information through step S101, the insurance business feature matrix may be directly input into the insurance money laundering detection model as an input.
S103, taking the output result of the insurance money laundering detection model as the detection result of whether the one or more insurance business information relates to insurance money laundering.
In an exemplary embodiment of the application, through an isolated forest algorithm included in the insurance money laundering detection model, whether corresponding multiple insurance policy business information has suspicion of money laundering can be directly calculated according to information of preset features corresponding to each insurance business feature vector in an input insurance business feature matrix, so that detection results of the multiple insurance policy business information are obtained.
S104, determining the related personnel determined as insurance business information related to the insurance money laundering as suspects of the insurance money laundering; the relevant persons include any one or more of: policyholder, insured person, and beneficiary.
In an exemplary embodiment of the present application, if one or more policy service information relates to a money laundering service, the applicant, insured person, beneficiary, etc. associated with such policy service information may all be determined to be a suspect of money laundering, thereby facilitating the targeting of public security authorities for further investigation of money laundering crimes.
In the exemplary embodiment of the application, an unsupervised mode in machine learning is adopted, a large number of manual rules are not needed to be relied on, some specific association relations can be learned from data, and people with high abnormal money laundering risks can be mined. Specifically, an unsupervised isolated forest algorithm in machine learning can be utilized, an insurance business feature matrix related to one or more insurance business information is detected by constructing an insurance money laundering detection model, a policy list corresponding to the insurance business information with abnormal suspicion of money laundering is effectively given, and the traditional money laundering risk screening mode which depends on expert rules, is low in working efficiency, high in investigation difficulty and low in mass data processing speed is greatly improved.
The embodiment of the present application further provides an insurance money laundering staff detecting device 1, as shown in fig. 4, which may include a processor 11 and a computer-readable storage medium 12, where the computer-readable storage medium 12 stores instructions, and when the instructions are executed by the processor 11, the insurance money laundering staff detecting method described in any one of the above items is implemented.
The embodiment of the application also provides a computer readable storage medium, wherein a computer program is stored on the computer readable storage medium, and when being executed by a processor, the computer program realizes the insurance money laundering staff detection method.
In the exemplary embodiments of the present application, any of the foregoing method embodiments are applicable to the apparatus embodiment and the computer-readable storage medium embodiment, and are not described in detail herein.
It will be understood by those of ordinary skill in the art that all or some of the steps of the methods, systems, functional modules/units in the devices disclosed above may be implemented as software, firmware, hardware, and suitable combinations thereof. In a hardware implementation, the division between functional modules/units mentioned in the above description does not necessarily correspond to the division of physical components; for example, one physical component may have multiple functions, or one function or step may be performed by several physical components in cooperation. Some or all of the components may be implemented as software executed by a processor, such as a digital signal processor or microprocessor, or as hardware, or as an integrated circuit, such as an application specific integrated circuit. Such software may be distributed on computer readable media, which may include computer storage media (or non-transitory media) and communication media (or transitory media). The term computer storage media includes volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data, as is well known to those of ordinary skill in the art. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, Digital Versatile Disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store the desired information and which can accessed by a computer. In addition, communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media as known to those skilled in the art.
Claims (10)
1. An insurance money laundering employee detection method, the method comprising:
acquiring an insurance business feature matrix related to one or more insurance business information and a trained insurance money laundering detection model; the insurance service characteristic matrix is a matrix used for representing corresponding information of preset service characteristics in each piece of insurance service information;
inputting the insurance business feature matrix into the insurance money laundering detection model as input data;
taking the output result of the insurance money laundering detection model as the detection result of whether the one or more insurance business information relates to insurance money laundering;
determining the related personnel determined as insurance business information related to the insurance money laundering as suspects of the insurance money laundering; the relevant persons include any one or more of: policemen, insureds, and beneficiaries;
the insurance money laundering detection model is obtained by training a preset machine learning model by taking a money laundering business feature matrix containing preset money laundering features as training data.
2. The insurance money laundering employee detection method according to claim 1, wherein the obtaining of the insurance service characteristic matrix regarding the one or more insurance services information comprises:
acquiring one or more insurance service information, and extracting one or more preset characteristics from the one or more insurance service information;
listing information of one or more preset characteristics corresponding to each insurance service information in a vector form according to a preset sequence to obtain an insurance service characteristic vector corresponding to each insurance service information;
and forming the insurance service characteristic matrix by all insurance service characteristic vectors corresponding to all insurance service information.
3. The insurance money laundering personnel detection method according to claim 2, wherein the insurance service information comprises any one or more of: information of the policy information table, information of the applicant information table, information of the insured person information table, information of the beneficiary information table and information of the policy transaction record table;
the preset features include any one or more of:
the method comprises the steps of determining whether a form is invalid, the time for handling the refund, the accumulated amount of the refund loss, the accumulated amount of the refund obtained by the irrelevant personnel of insurance service information, the consistency of the card number of the refund bank and the card number of the bank for handling the refund, the amount of the policy paid by cash, the total number of changes of the policy applicant, the difference of the change time of the policy applicant, whether the change time period of the policy applicant is around holidays, the number of changes of the policy service within a preset time length, the age of the applicant, the national nationality of the applicant, the occupation of the applicant and the consistency of the applicant and the refund handler.
4. The insurance money laundering employee detection method of claim 1, wherein obtaining the trained insurance money laundering detection model comprises:
calling a pre-created and trained insurance money laundering detection model; alternatively, the first and second electrodes may be,
and establishing a machine learning model, training the machine learning model, and acquiring the insurance money laundering detection model.
5. The insurance money laundering personnel detection method according to claim 4, wherein the creating and training of the machine learning model, and the obtaining of the insurance money laundering detection model comprises:
obtaining a money laundering insurance business information set that has been determined to be involved in an insurance money laundering business, obtaining one or more preset money laundering characteristics from the money laundering insurance business information set, the money laundering business characteristic matrix being formed of the one or more preset money laundering characteristics as training data;
constructing a machine learning model by adopting a preset machine learning algorithm;
and training the machine learning model by adopting the training data to obtain the insurance money laundering detection model.
6. The insurance money laundering personnel detection method according to claim 5, wherein the preset machine learning algorithm comprises: an isolated forest algorithm;
the preset money laundering characteristics comprise any one or more of:
the current form is a failure form;
the time length difference between the release transaction time and the application transaction time is smaller than a preset time length threshold value;
the accumulated refund loss sum is greater than or equal to a preset first sum threshold;
the accumulated acquired amount of the refund of the irrelevant personnel of the insurance service information is greater than or equal to a preset second amount threshold;
the number of the refund bank card is inconsistent with the number of the transacted bank card;
the insurance policy amount paid by cash is greater than or equal to a preset third amount threshold value;
the total change times of the policy applicant is greater than or equal to a preset first time threshold value;
the policy service change times in the preset time length are greater than or equal to a preset second time threshold;
the policy applicant changes the time difference to be less than or equal to the preset time difference threshold value;
the policy applicant changes the time period before and after the holiday;
the age of the applicant is greater than a preset age threshold;
the state of the applicant is a foreign state;
the employment enterprises of the applicant belong to the preset money laundering high-risk enterprise range; and the number of the first and second groups,
the policyholder is inconsistent with the retirement transactor.
7. The insurance money laundering personnel detection method of claim 5, wherein the forming of the money laundering business feature matrix from one or more preset money laundering features comprises:
acquiring a money laundering business feature vector with the money laundering feature corresponding to each piece of money laundering insurance business information, and forming a money laundering business feature matrix by all money laundering business feature vectors corresponding to all money laundering insurance business information;
the money laundering service characteristic vector is information of one or more money laundering characteristics corresponding to each money laundering insurance service information listed in a vector form according to a preset sequence;
the method further comprises the following steps: preprocessing the money laundering business feature matrix before the money laundering business feature matrix is used as the training data; the pretreatment comprises the following steps: missing value supplementation and normalization operations;
supplementing the missing value of the money laundering business feature matrix comprises:
filling missing value parts of each money laundering characteristic with missing value in the money laundering business characteristic matrix by adopting any one or more of the following modes: mean filling, mode filling, median filling and 0-filling;
the normalization operation of the money laundering business feature matrix comprises the following steps:
carrying out dimensionless operation on the money laundering service characteristic matrix, and removing the maximum value MAX and the minimum value MIN in each column in the money laundering service characteristic matrix to obtain the characteristic value x of each columni：
8. The insurance money laundering personnel detection method according to claim 7, further comprising:
after normalization operation is carried out on the money laundering service characteristic matrix, calculating the correlation between any two characteristic dimensions in the money laundering service characteristic matrix through a preset correlation calculation formula; and removing any one of the two characteristic dimensions of which the correlation calculation result is greater than a preset correlation threshold value, so as to realize the screening of the money laundering service characteristic matrix.
9. An insurance money laundering personnel detection apparatus comprising a processor and a computer readable storage medium having instructions stored therein, wherein the instructions, when executed by the processor, implement the insurance money laundering personnel detection method according to any of claims 1-8.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the insurance money laundering person detection method according to any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010404117.1A CN111709844A (en) | 2020-05-13 | 2020-05-13 | Insurance money laundering personnel detection method and device and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010404117.1A CN111709844A (en) | 2020-05-13 | 2020-05-13 | Insurance money laundering personnel detection method and device and computer readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111709844A true CN111709844A (en) | 2020-09-25 |
Family
ID=72537808
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010404117.1A Pending CN111709844A (en) | 2020-05-13 | 2020-05-13 | Insurance money laundering personnel detection method and device and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111709844A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113568947A (en) * | 2021-07-21 | 2021-10-29 | 众安在线财产保险股份有限公司 | Data processing method, system and computer storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106952162A (en) * | 2016-01-07 | 2017-07-14 | 平安科技（深圳）有限公司 | Money laundering risks rating calculation method and system |
| CN108053087A (en) * | 2017-10-20 | 2018-05-18 | 深圳前海微众银行股份有限公司 | Anti money washing monitoring method, equipment and computer readable storage medium |
| CN110046993A (en) * | 2018-12-15 | 2019-07-23 | 深圳壹账通智能科技有限公司 | Illicit gain legalizes behavior monitoring method, system, computer installation and medium |
| CN110276618A (en) * | 2019-06-28 | 2019-09-24 | 第四范式（北京）技术有限公司 | The method and system for generating money laundering ancestor prediction model, predicting money laundering ancestor |
| CN111127200A (en) * | 2019-11-25 | 2020-05-08 | 中国建设银行股份有限公司 | Method and device for monitoring suspicious transactions of anti-money laundering |
-
2020
- 2020-05-13 CN CN202010404117.1A patent/CN111709844A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106952162A (en) * | 2016-01-07 | 2017-07-14 | 平安科技（深圳）有限公司 | Money laundering risks rating calculation method and system |
| CN108053087A (en) * | 2017-10-20 | 2018-05-18 | 深圳前海微众银行股份有限公司 | Anti money washing monitoring method, equipment and computer readable storage medium |
| CN110046993A (en) * | 2018-12-15 | 2019-07-23 | 深圳壹账通智能科技有限公司 | Illicit gain legalizes behavior monitoring method, system, computer installation and medium |
| CN110276618A (en) * | 2019-06-28 | 2019-09-24 | 第四范式（北京）技术有限公司 | The method and system for generating money laundering ancestor prediction model, predicting money laundering ancestor |
| CN111127200A (en) * | 2019-11-25 | 2020-05-08 | 中国建设银行股份有限公司 | Method and device for monitoring suspicious transactions of anti-money laundering |
Non-Patent Citations (3)
| Title |
|---|
| 向玲等: "保险业反洗钱智能监测模型研究", 《金融电子化》 * |
| 罗素文等: ""机器学习技术在商业银行反洗钱领域的应用"", 《金融电子化》 * |
| 赵东明等: ""寿险业客户洗钱风险等级分类管理工作研究——基于分类模型和系统架构视角"", 《福建金融》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113568947A (en) * | 2021-07-21 | 2021-10-29 | 众安在线财产保险股份有限公司 | Data processing method, system and computer storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111291816B (en) | Method and device for carrying out feature processing aiming at user classification model | |
| CN110188198B (en) | Anti-fraud method and device based on knowledge graph | |
| US8121920B2 (en) | System and method of detecting mortgage related fraud | |
| US8160982B2 (en) | Method for detecting people of interest from information sources | |
| US20140143134A1 (en) | System and method of detecting and assessing multiple types of risks related to mortgage lending | |
| WO2005101265A2 (en) | Systems and methods for investigation of financial reporting information | |
| US11263644B2 (en) | Systems and methods for detecting unauthorized or suspicious financial activity | |
| US20150363875A1 (en) | System and Method for Filtering and Analyzing Transaction Information | |
| US8145585B2 (en) | Automated methods and systems for the detection and identification of money service business transactions | |
| CN110728301A (en) | Credit scoring method, device, terminal and storage medium for individual user | |
| CN112381640A (en) | Service data monitoring method, device, equipment and storage medium | |
| CN115545886A (en) | Overdue risk identification method, overdue risk identification device, overdue risk identification equipment and storage medium | |
| CN111709844A (en) | Insurance money laundering personnel detection method and device and computer readable storage medium | |
| CN113177733B (en) | Middle and small micro enterprise data modeling method and system based on convolutional neural network | |
| CN110570301B (en) | Risk identification method, device, equipment and medium | |
| CN112823502B (en) | Real-time feedback service for resource access rule configuration | |
| CN113822751A (en) | Online loan risk prediction method | |
| CN111861715A (en) | Credit data processing method, device, equipment and storage medium | |
| CN113723522B (en) | Abnormal user identification method and device, electronic equipment and storage medium | |
| CN115345736B (en) | Abnormal behavior detection method for financial transaction | |
| US8078517B1 (en) | Systems and methods for monitoring remittances for reporting requirements | |
| CN108520042B (en) | System and method for realizing suspect case-involved role calibration and role evaluation in detection work | |
| Suhanjoyo et al. | Fraud Detection in Sales of Distribution Companies Using Machine Learning | |
| CN114066643A (en) | Method and device for rapidly mining suspected associated bill based on sliding window and storage medium | |
| Shaikh et al. | Credit Risk Assessment |
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 |