CN115712627A

CN115712627A - Intelligent financial anti-fraud method and system based on artificial intelligence

Info

Publication number: CN115712627A
Application number: CN202211611835.1A
Authority: CN
Inventors: 郭一迪; 陈辰; 王震; 段美宁
Original assignee: Hangyin Consumer Finance Co ltd
Current assignee: Hangyin Consumer Finance Co ltd
Priority date: 2022-12-15
Filing date: 2022-12-15
Publication date: 2023-02-24

Abstract

According to the intelligent financial anti-fraud method and system based on artificial intelligence, the basic fraud query tables corresponding to the financial fraud collating sets are determined, the confidence factors CM1 of each fraud of each type of fraud of each financial fraud collating set with each fraud description information are obtained preliminarily, then the basic fraud query tables are fused to obtain the accurate selected fraud query table used for indicating that each type of preset fraud has the selected confidence factor of each preset fraud description information, therefore, when the selected fraud of the object to be identified is determined through the selected fraud query table, the accuracy and the efficiency can be effectively improved, meanwhile, the method and system have excellent applicability, the fraud can be intervened in time conveniently, and anti-fraud is achieved.

Description

Intelligent financial anti-fraud method and system based on artificial intelligence

Technical Field

The application relates to the field of data processing, in particular to an intelligent financial anti-fraud method and system based on artificial intelligence.

Background

With the development of the internet and information technology, the number of groups conducting transactions on the internet is increasing, and a large number of lawbreakers conducting financial fraud are bred. In financial fraud, a lawbreaker typically induces stress on a target object through a channel of telephone, network, information, etc., and conducts fraudulent acts such as credit fraud, loan fraud, funding fraud, ticket fraud, etc. Along with the generation of intelligent financial means and tools, the intelligent finance carries out information acquisition and analysis on high-risk financial fraud behaviors through legal compliance means, completes the identification and intervention of financial fraud, is favorable for maintaining normal financial ecological environment, and is a popular subject in the current intelligent financial field. However, in the current financial anti-fraud technology, there is room for improvement in the accurate and efficient identification of fraudulent activities.

Disclosure of Invention

The invention aims to provide an intelligent financial anti-fraud method and system based on artificial intelligence so as to improve the problems.

The technical scheme of the embodiment of the application is realized as follows:

in a first aspect, an embodiment of the present application provides an intelligent financial anti-fraud method based on artificial intelligence, which is applied to an intelligent financial anti-fraud system, and the method includes:

determining not less than one financial fraud consolidation set, wherein each financial fraud consolidation set comprises fraud behaviors corresponding to not less than one object and fraud behavior description information;

for each financial fraud arrangement set, acquiring a basic fraud behavior query table corresponding to the financial fraud arrangement set; the basic fraud behavior look-up table is used for indicating that each type of fraud behavior in the financial fraud settling set has a confidence factor CM1 of each fraud behavior description information in the financial fraud settling set;

determining a selected fraud lookup table based on a plurality of the base fraud lookup tables; the selected fraud query table is used for indicating that each type of preset fraud has a selected confidence factor of each preset fraud description information, each selected confidence factor is determined based on at least one confidence factor CM1 that a preset fraud corresponding to the selected confidence factor in a plurality of basic fraud query tables has a preset fraud description information corresponding to the selected confidence factor, each type of preset fraud is one of fraud corresponding to each financial fraud sorting set, and each preset fraud description information is one of fraud description information corresponding to each financial fraud sorting set;

determining at least one piece of selected fraud behavior description information of an object to be identified, and determining the selected fraud behavior corresponding to the object to be identified through a plurality of pieces of selected fraud behavior description information and the selected fraud behavior lookup table.

As an implementation manner, for each financial fraud arrangement set, the obtaining a basic fraud behavior lookup table corresponding to the financial fraud arrangement set includes:

for each type of fraudulent conduct in the financial fraud arrangement set and each fraudulent conduct description information, obtaining a first number of objects having the fraudulent conduct and a second number of objects having the fraudulent conduct and the fraudulent conduct description information in the financial fraud arrangement set, and determining a confidence factor CM1 that the fraudulent conduct has the fraudulent conduct description information according to the first number and the second number;

and acquiring a basic fraud behavior query table corresponding to the financial fraud arrangement set based on each confidence factor CM1 corresponding to each type of fraud behavior in the financial fraud arrangement set.

As one embodiment, the determining the selected fraud lookup table based on a plurality of the basic fraud lookup tables includes:

for each type of the preset fraudulent conduct and each type of the preset fraudulent conduct description information, determining a plurality of basic fraudulent conduct query tables including a fraudulent conduct query Table1 of a confidence factor CM1 corresponding to the type of the preset fraudulent conduct, and determining that the type of the preset fraudulent conduct has a combined confidence factor of the preset fraudulent conduct description information based on a third number of the plurality of fraudulent conduct query tables Table1 and the confidence factor CM1 of the preset fraudulent conduct with the preset fraudulent conduct description information in each fraudulent conduct query Table 1;

and determining a selected fraud look-up table based on the combined confidence factor that each type of the preset fraud has each preset fraud description information.

As an embodiment, the determining the selected fraud lookup table based on the combined confidence factor that each of the preset fraud behaviors have each of the preset fraud description information includes:

determining a fraud behavior description information relation graph corresponding to each preset fraud behavior description information; a component in the fraud description information relation diagram is used for indicating the preset fraud description information, and the preset fraud description information pointed by the lower component of the component in the fraud description information relation diagram is the lower fraud description information of the preset fraud description information pointed by the component;

and correcting the merged confidence factor of each preset fraud behavior description information of each preset fraud behavior based on the fraud behavior description information relation diagram to obtain the selected confidence factor of each preset fraud behavior description information of each preset fraud behavior.

As an implementation manner, for each type of the preset fraud and each preset fraud description information, modifying the merged confidence factor that the preset fraud of the type has the preset fraud description information based on the fraud description information relationship diagram, so as to obtain the selected confidence factor that the preset fraud of the type has the preset fraud description information, including:

when the class of preset fraudulent conduct has the combined confidence factor of the preset fraudulent conduct description information which is greater than the preset confidence factor and the fraudulent conduct description information relation graph does not contain the lower component of the first component, determining the combined confidence factor of the class of preset fraudulent conduct having the preset fraudulent conduct description information as the selected confidence factor of the class of preset fraudulent conduct having the preset fraudulent conduct description information, wherein the first component is the component which is used for indicating the preset fraudulent conduct description information in the fraudulent conduct description information relation graph;

when the preset fraud behavior of the class is provided with the combined confidence factor of the preset fraud behavior description information and is greater than the preset confidence factor and the fraud behavior description information relation graph comprises the subordinate component of the first component, taking the maximum combined confidence factor in the preset fraud behavior description information combined confidence factor of the class and the selected confidence factor of the preset fraud behavior description information used for indicating by each subordinate component of the preset fraud behavior of the class and the third subordinate component as the selected confidence factor of the preset fraud behavior description information provided by the class;

when the combined confidence factor of the preset fraud behavior description information of the class of preset fraud behaviors is smaller than or equal to the preset confidence factor and a second component in the fraud behavior description information relation graph comprises not less than one third component, determining that the class of preset fraud behaviors has the selected confidence factor of the preset fraud behavior description information based on the selected confidence factor of the preset fraud behavior description information pointed by a plurality of lower components of the first component of the class of preset fraud behaviors; the second elements comprise lower-level elements of the first elements and elements with derivative relations with the first elements, and the preset cheating behaviors of the type are provided with combined confidence factors of preset cheating behavior description information used for indicating by each third element and are larger than the preset confidence factors;

when the preset fraud behavior of the class has the combined confidence factor of the preset fraud behavior description information which is less than or equal to the preset confidence factor and the fraud behavior description information relation graph does not contain the lower component of the first component, or the preset fraud behavior has the combined confidence factor of the preset fraud behavior description information which is less than or equal to the preset confidence factor and the fraud behavior description information relation graph does not contain the third component, determining the preset confidence factor as the selected confidence factor that the preset fraud behavior of the class has the preset fraud behavior description information.

As an implementation manner, the determining, by using a plurality of pieces of selected fraud description information and the selected fraud lookup table, a selected fraud corresponding to the object to be identified includes:

determining a selected matching coefficient of the object to be identified and each type of preset fraudulent conduct according to a plurality of selected fraudulent conduct description information and the selected fraudulent conduct lookup table;

determining the selected fraudulent behavior corresponding to the object to be identified based on the selected matching coefficient of the object to be identified and each type of preset fraudulent behavior;

the step of determining the selected matching coefficient between the object to be identified and the preset fraud of each type according to the description information of the selected fraud and the selected fraud lookup table for each type of the preset fraud comprises one or more of the following implementation processes:

determining a prominent distribution factor of each preset fraudulent behavior description information to the fraudulent behavior based on the selected fraudulent behavior query table, and determining a selected matching coefficient of the object to be identified and the class of preset fraudulent behavior according to the prominent distribution factors corresponding to a plurality of preset fraudulent behavior description information and each selected fraudulent behavior description information;

determining a confidence factor CM2 that any object has the class-preset fraud, and determining a selected matching coefficient of the object to be identified and the class-preset fraud according to the confidence factor CM2 and the class-preset fraud which has the selected confidence factor of each piece of preset fraud description information;

inputting the selected fraud description information into a fraud identification network to obtain a selected matching coefficient of the object to be identified and the class of preset fraud; the fraud identification network is obtained by debugging based on the selected fraud lookup table.

As an implementation manner, for each class of the preset fraudulent conduct, the determining, by using the prominent distribution factor corresponding to the plurality of preset fraudulent conduct description information and each of the selected fraudulent conduct description information, a selected matching coefficient between the object to be identified and the class of the preset fraudulent conduct includes:

determining target fraud description information in each selected fraud description information; wherein each of the selected fraud descriptor does not include a subordinate fraud descriptor of the target fraud descriptor;

determining a fraud behavior coincidence description information Cluster1 of each target fraud behavior description information and preset fraud behavior description information Pre-info1 in each preset fraud behavior description information, and determining a selected matching coefficient of the object to be identified and the class of preset fraud behaviors based on a prominent distribution factor corresponding to each fraud behavior description information in the fraud behavior coincidence description information Cluster 1; said class of predetermined fraudulent activities has a selected confidence factor greater than a confidence factor CM4 for each of said predetermined fraudulent activity description information Pre-info 1.

As an implementation manner, for each type of the preset fraudulent conduct, determining a selected matching coefficient between the object to be identified and the preset fraudulent conduct based on a highlight distribution factor corresponding to each fraudulent conduct description information in the fraudulent conduct coincidence description information Cluster1 includes:

determining a matching coefficient MC1 of the object to be identified and the preset fraudulent behaviors based on a highlight distribution factor corresponding to each fraudulent behavior description information in the fraudulent behavior coincidence description information Cluster 1;

determining preset fraud behavior description information Pre-info2 in each preset fraud behavior description information Pre-info 1; each piece of preset fraud description information Pre-info1 does not contain lower-level fraud description information of the preset fraud description information Pre-info2;

determining a fraud behavior coincidence description information Cluster2 of each piece of second fraud behavior description information and each piece of selected fraud behavior description information, and determining a matching coefficient MC2 of the object to be identified and the class of preset fraud behaviors based on a highlight distribution factor corresponding to each piece of fraud behavior description information in the fraud behavior coincidence description information Cluster 2;

and determining the selected matching coefficient of the object to be identified and the preset fraud behaviors based on the matching coefficient MC1 and the matching coefficient MC2.

As an implementation manner, the determining, according to the confidence factor CM2 and the selected confidence factor of each piece of preset fraud description information of the class-preset fraud, the selected matching coefficient of the object to be recognized and the class-preset fraud includes:

determining a confidence factor CM3 that the class of preset fraudulent conduct has the target fraudulent conduct description information at the same time based on the selected confidence factor that the class of preset fraudulent conduct has each preset fraudulent conduct description information;

and determining the selected matching coefficient of the object to be identified and the preset fraud behaviors according to the confidence factor CM2 and the confidence factor CM3.

In a second aspect, an embodiment of the present application provides an intelligent financial anti-fraud system, including a processor and a memory, where the memory stores a computer program for implementing the method described above when executed by the processor.

The embodiment of the application has at least the following beneficial effects:

In the description that follows, additional features will be set forth, in part, in the description. These features will be in part apparent to those of ordinary skill in the art upon examination of the following and the accompanying drawings or may be learned by production or use. The features of the present application may be realized and attained by practice or use of various aspects of the methodologies, instrumentalities and combinations particularly pointed out in the detailed examples that follow.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the disclosure.

Fig. 1 is a schematic view of an application scenario of an intelligent financial anti-fraud method based on artificial intelligence according to an embodiment of the present application.

Fig. 2 is a flowchart of an intelligent financial anti-fraud method based on artificial intelligence according to an embodiment of the present application.

Fig. 3 is a schematic diagram illustrating a functional module architecture of an intelligent financial anti-fraud device according to an embodiment of the present application.

Fig. 4 is a schematic diagram illustrating an intelligent financial anti-fraud system according to an embodiment of the present disclosure.

Detailed Description

In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be described in further detail with reference to the attached drawings, the described embodiments should not be considered as limiting the present application, and all other embodiments obtained by a person of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

In the following description, references to "some embodiments," "as an implementation," "in an implementation," describe a subset of all possible embodiments, but it is understood that "some embodiments," "as an implementation," "in an implementation" may be the same subset or a different subset of all possible embodiments, and may be combined with each other without conflict.

In the following description, references to the terms "first \ second \ third", and the like, merely distinguish between similar objects and do not denote a particular order, but rather denote a particular order, and it is to be understood that "first \ second \ third", where permissible, may be interchanged either in a particular order or in a sequential order so that embodiments of the application described herein may be practiced in other than that shown or described herein. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of the present application only and is not intended to be limiting of the application.

The intelligent financial anti-fraud method based on artificial intelligence provided by the embodiment of the application can be executed by electronic devices such as an intelligent financial anti-fraud system, wherein the intelligent financial anti-fraud system can be various types of terminals such as a notebook computer, a tablet computer, a desktop computer, a mobile device (e.g., a mobile phone, a portable music player, a personal digital assistant, a dedicated messaging device, a portable game device) and the like, and can also be implemented as a server. The server may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing basic cloud computing services such as a cloud service, a cloud database, cloud computing, a cloud function, cloud storage, a Network service, cloud communication, middleware service, a domain name service, a security service, a Content Delivery Network (CDN), a big data and artificial intelligence platform, and the like. For example, the following description will clearly and completely describe the technical solution in the embodiments of the present application with reference to the drawings in the embodiments of the present application in an exemplary application in which the intelligent financial anti-fraud system is implemented as a server.

Fig. 1 is a schematic view illustrating an application scenario of an intelligent financial anti-fraud method based on artificial intelligence according to an embodiment of the present application. Wherein, a plurality of terminal devices 100 and the intelligent financial anti-fraud system 300 are connected in communication through the network 200. The intelligent financial anti-fraud system 300 is used for executing the method provided by the embodiment of the application. Specifically, the present application provides an artificial intelligence based intelligent financial anti-fraud method, which is applied to an intelligent financial anti-fraud system 300, as shown in fig. 2, and the method includes:

step S10: no less than one financial fraud collage is determined.

In one embodiment, each financial fraud arrangement set includes fraud behaviors corresponding to at least one object and fraud behavior description information, the fraud behaviors are credit fraud, bill fraud, funding fraud, loan fraud and the like, the fraud behavior description information is tagging information describing the fraud behaviors, such as specific contained text information, such as "national deposit", "no mortgage deposit", "ultra-low interest rate", "guarantee deposit", "commission charge" and the like, and it is understood that the specific content of the fraud behavior description information is not limited thereto and can be configured as required. The object in the financial fraud arrangement set can be a cheater or a cheated person, the fraud behavior and the fraud behavior description information are obtained by analyzing the session information of the cheater and the cheated person obtained in the legal compliance range, when the object is the cheater, the fraud behavior is the behavior of the cheater, and when the object is the cheated person, the fraud behavior is the behavior of the cheated person. Different objects in the same financial fraud finishing set may have the same fraud or different fraud, and in addition, different objects may have different fraud descriptors while having the same fraud.

Step S20: and for each financial fraud arrangement set, acquiring a basic fraud behavior query table corresponding to the financial fraud arrangement set.

In one embodiment, the base fraud-related look-up table for each financial fraud arrangement is used to indicate that each type of fraud in the financial fraud arrangement has a confidence factor (confidence factor CM 1) for each fraud-related descriptive information in the financial fraud arrangement, i.e. a confidence factor CM1 for indicating that an object has each fraud-related descriptive information when having any of the fraud in the financial fraud arrangement, the confidence factor characterizing the likelihood that each type of fraud in the financial fraud arrangement has each fraud-related descriptive information in the financial fraud arrangement. For each type of fraud and each fraud description in each financial fraud collation, a first number of objects in the financial fraud collation having the type of fraud and a second number of objects in the financial fraud collation having the type of fraud and having the fraud description may be obtained. For example, one financial fraud consolidation set comprising fraud "credit fraud" and fraud description information "cash-out," a first number of objects having credit fraud in the financial fraud consolidation set and a second number of objects having and having cash-out fraud description information may be obtained. For each type of fraud and each fraud description in each financial fraud finishing set, determining a confidence factor CM1 that the type of fraud has the fraud description in proportion to the second number of objects that have the type of fraud and have the fraud description and the first number of objects that have the type of fraud. And for each financial fraud arrangement set, a fraud behavior determinant and a fraud behavior description information determinant corresponding to the financial fraud arrangement set can be obtained, and a confidence factor CM1 of any fraud behavior description information of any type is determined through the fraud behavior determinant and the fraud behavior description information determinant. For each financial fraud collation set, the fraud description information determinant corresponding to the financial fraud collation set can be determined by the fraud description information of each object in the financial fraud collation set.

The fraud descriptor determinant is an M × N determinant, and an element in the determinant is a value corresponding to an object in a corresponding row (or column), and the value is used for indicating whether the object has fraud descriptor in the corresponding column (or row). And M is the number of objects in the financial fraud arrangement set, and N is the number of classes of the fraud behavior description information in the financial fraud arrangement set. The expression form of the value may be any feasible form, for example, discrete vectors defined as 0 and 1, if the value is 1, it indicates that the correspondence has corresponding fraud description information, and if the value is 0, it indicates that the correspondence does not have corresponding fraud description information, which is not limited in this embodiment of the application. In addition, for each financial fraud collation, the fraud determinant corresponding to the financial fraud collation can be determined by the fraud corresponding to each object in the financial fraud collation. The fraud determinant is an M × H determinant, and elements in the determinant are values corresponding to objects in corresponding rows (or columns), and the values are used for indicating whether the objects have fraud in corresponding columns (or rows). Wherein M is the number of objects in the financial fraud arrangement set, and H is the number of classes of fraud in the financial fraud arrangement set. Based on the principle that the determinant of the fraud description information is consistent, the representation form of the value in the determinant of the fraud may be any feasible form, for example, discrete vectors defined as 0 and 1, if the value is 1, it indicates that the correspondence has corresponding fraud, and if the value is 0, it indicates that the correspondence does not have corresponding fraud, which is not limited in the embodiment of the present application.

The confidence factor CM1 can be calculated with reference to the following formula:

CM1=Q2/Q1；

the element value of the determinant of the fraud behavior description information is 1, which represents that the corresponding object has corresponding fraud behavior description information, the element value of 0 represents that the corresponding object does not have corresponding fraud behavior description information and the element value of the determinant of the fraud behavior is 1, which represents that the corresponding object has corresponding fraud behavior, when the element value of 0 represents that the corresponding object does not have corresponding fraud behavior, Q1 is a first number of objects having fraud behavior, and Q2 is a second number of objects having fraud behavior and having fraud behavior description information. Further, for each financial fraud settling set, after a confidence factor CM1 of each type of fraud in the financial fraud settling set, which has the description information of each fraud in the financial fraud settling set, is obtained, a basic fraud query table corresponding to the fraud is generated based on a plurality of confidence factors CM1. The element values in the fraud look-up table are confidence factors that the corresponding fraud has the description information of the object fraud.

Step S30: a selected fraud lookup table is determined based on a plurality of underlying fraud lookup tables.

In one embodiment, since the fraud behaviors corresponding to the objects in the different financial fraud settling sets and the fraud behavior description information may not be adapted, each confidence factor in the corresponding basic fraud behavior query table has an error with the actual confidence factor, and in order to improve the accuracy of the confidence factor CM1 of any fraud behavior having any fraud behavior description information, the basic fraud behavior query tables corresponding to the financial fraud settling sets are fused to obtain the selected fraud behavior query table. The selected fraud behavior lookup table is used for indicating that each type of preset fraud behavior has a selected confidence factor of each preset fraud behavior description information, each type of preset fraud behavior is one type of fraud behavior in all fraud behaviors corresponding to each financial fraud arrangement set, and each preset fraud behavior description information is one piece of fraud behavior description information in all fraud behavior description information corresponding to each financial fraud arrangement set. For example, the preset fraud may be fraud that is simultaneously present in each basic fraud lookup table, and the preset fraud description information is fraud description information that is simultaneously present in each basic fraud lookup table; or the preset fraud behavior is the fraud behavior with the frequency meeting the preset times in each basic fraud behavior query table, and the preset fraud behavior description information is all corresponding fraud behavior description information of each preset fraud behavior in each basic fraud behavior query table; or the preset fraudulent conduct may be all fraudulent conduct corresponding to each basic fraudulent conduct query table, and the preset fraudulent conduct description information is all fraudulent conduct description information corresponding to each basic fraudulent conduct query table.

For each selected confidence factor in the selected fraud query table, the selected confidence factor is obtained by obtaining each confidence factor CM1 of the preset fraud corresponding to the selected confidence factor in all basic fraud query tables, wherein the preset fraud corresponding to the selected confidence factor has the preset fraud description information corresponding to the selected confidence factor. For example, for each type of preset fraud and each preset fraud description information, it may be determined in all basic fraud lookup tables that the type of preset fraud has each confidence factor CM1 of the preset fraud description information, and then the confidence factors of each determined confidence factor CM1 are averaged, and the average value is used as the selected confidence factor that the type of preset fraud has the preset fraud description information; or determining an allocation factor (for representing the importance or the reliability of the corresponding basic fraud query table, which can be embodied by a weight) of each basic fraud query table, and also for indicating the importance or the reliability of the corresponding financial fraud arrangement set, and then multiplying each confidence factor CM1 of the preset fraud with the preset fraud description information by the corresponding allocation factor, and then summing the result to obtain a selected confidence factor of the preset fraud with the preset fraud description information.

For each type of preset fraud and each type of preset fraud description information, a fraud lookup Table (regarded as a fraud lookup Table 1) including the confidence factor CM1 of the type of preset fraud in each basic fraud lookup Table may be determined, and the fraud lookup Table1 including the involvement information of the type of preset fraud is determined in each basic fraud lookup Table. And then, determining the third number of each fraudulent behavior query Table1 and the confidence factor CM1 that the preset fraudulent behavior in each fraudulent behavior query Table1 has the preset fraudulent behavior description information, and determining that the preset fraudulent behavior in each class has the combined confidence factor of the preset fraudulent behavior description information. For example, the result of summing the confidence factors of the confidence factors CM1 of the preset fraud behaviors having the preset fraud behavior description information in each fraud behavior query Table1 is used, and then the result of the ratio of the confidence factors to the third number of the fraud behavior query Table1 is used as the combined confidence factor of the preset fraud behavior description information of the preset fraud behaviors; or determining the distribution factor of each fraud behavior query Table1, then performing multiplication operation on the confidence factor CM1 of the preset fraud behavior with the preset fraud behavior description information in each fraud behavior query Table1 and the distribution factor of the corresponding fraud behavior query Table1 to obtain a product, summing the products corresponding to each fraud behavior query Table1 to obtain a sum value, performing proportional operation on the obtained sum value and the third number of the fraud behavior query Table1 to obtain a proportional value, and taking the proportional value as the combined confidence factor of the preset fraud behavior description information of the preset fraud behavior. Based on the above, the merged confidence factor of each preset fraud behavior with each preset fraud behavior description information is obtained, and then the selected fraud behavior query table is determined based on the merged confidence factors.

In one embodiment, when the selected fraud query table is determined by the merged confidence factor that each type of preset fraud has each preset fraud description information, for each type of preset fraud and each preset fraud description information, the merged confidence factor that the preset fraud has the preset fraud description information of the type may be determined as the selected confidence factor that the preset fraud has the preset fraud description information of the type, and then the selected fraud query table is determined by the selected confidence factor that each type of preset fraud has each preset fraud description information of the type.

In one embodiment, for each piece of predetermined fraud description information, there is a case that: one piece of preset fraud description information is lower fraud description information of another piece of preset fraud description information, for example, the preset fraud description information "fictitious transaction" is lower fraud description information of "cash-over". For each type of preset fraud and each type of preset fraud description information, the confidence factor that the class of preset fraud has the class of preset fraud description information is greater than the confidence factor that the class of preset fraud has the lower level fraud description information of the class of preset fraud description information. On the basis, in order to improve the accuracy of the selected fraud behavior lookup table, a fraud behavior description information relation graph corresponding to each preset fraud behavior description information is determined, the fraud behavior description information relation graph is used for indicating the upper-level and lower-level relations among the preset fraud behavior description information, wherein one component in the fraud behavior description information relation graph is used for indicating one preset fraud behavior description information, and in addition, the preset fraud behavior description information indicated by the lower-level component of one component in the fraud behavior description information relation graph is the lower-level fraud behavior description information of the preset fraud behavior description information indicated by the component.

After the fraud behavior description information relation graph corresponding to each preset fraud behavior description information is determined, the merged confidence factor of each preset fraud behavior description information of each type of preset fraud behavior can be corrected through the fraud behavior description information relation graph, and the selected confidence factor of each preset fraud behavior description information of each type of preset fraud behavior is obtained. The selected confidence factor with the preset fraud behavior description information as the preset fraud behavior can be updated and optimized by the preset updating function on the combined confidence factor with the preset fraud behavior description information as the preset fraud behavior. For example, for each type of preset fraud and each preset fraud description information, when the type of preset fraud has a lower component whose combined confidence factor of the preset fraud description information is greater than a preset confidence factor and whose fraud description information relationship diagram does not include a first component (a component in the fraud description information relationship diagram for indicating the preset fraud description information), the combined confidence factor that the type of preset fraud has the preset fraud description information is determined as the selected confidence factor that the type of preset fraud has the preset fraud description information. For each type of preset fraud and each preset fraud description information, when the type of preset fraud has a combined confidence factor of the preset fraud description information which is greater than a preset confidence factor and a lower component of the fraud description information relation diagram does not contain a first component, determining that the type of preset fraud has a selected confidence factor of the preset fraud description information and a selected confidence factor of the preset fraud description information (namely the lower fraud description information of the preset fraud description information) for indicating that each lower component of the type of preset fraud has the first component, and determining that the type of preset fraud has the selected confidence factor of the preset fraud description information. Specifically, the maximum value of the combination confidence factor that the preset fraudulent behavior of the category has the preset fraudulent behavior description information and the selected confidence factor that the preset fraudulent behavior of the category has each type of lower-level fraudulent behavior description information of the preset fraudulent behavior description information may be determined as the selected confidence factor that the preset fraudulent behavior of the category has the preset fraudulent behavior description information.

For each type of preset fraudulent conduct and each preset fraudulent conduct description information, when the merged confidence factor of the preset fraudulent conduct description information of the type of preset fraudulent conduct is smaller than or equal to the preset confidence factor and the second component in the relation graph of the fraudulent conduct description information comprises not less than one third component, the selected confidence factor of the preset fraudulent conduct description information of the type of preset fraudulent conduct with the first component is determined through the selected confidence factor of the preset fraudulent conduct description information pointed by each lower component of the first component of the type of preset fraudulent conduct (namely, each lower fraudulent conduct description information of the preset fraudulent conduct description information), wherein the second component comprises each lower component of the first component and the component having a derivative relation with the first component, such as all lower components of the first component, and the merged confidence factor of the preset fraudulent conduct description information pointed by each third component of the type of preset fraudulent conduct is larger than the preset confidence factor.

For each type of preset fraud behaviors and each preset fraud behavior description information, when the type of preset fraud behaviors has a combined confidence factor of the preset fraud behavior description information which is less than or equal to a preset confidence factor and a lower component of a fraud behavior description information relation diagram does not contain a first component (namely the lower fraud behavior description information which does not contain the preset fraud behavior description information), or the type of preset fraud behaviors has a combined confidence factor of the preset fraud behavior description information which is less than or equal to the preset confidence factor and a second component does not contain a third component (in the preset fraud behavior description information which is used for indicating that all derivative components of the preset fraud behavior description information point to and do not have preset fraud behavior description information of which the corresponding combined confidence factor is greater than the preset confidence factor), determining the preset confidence factor as the selected confidence factor that the type of preset fraud behaviors have the preset fraud behavior description information.

For any fraud description information (e.g. cash register), when the object has lower level fraud description information (e.g. fictitious transaction) of the fraud description information, the object must have the fraud description information (cash register), after obtaining a fraud description information relation diagram corresponding to each preset fraud description information, the object completes part of upper level fraud description information corresponding to the lower level fraud description information, and simultaneously, for each type of preset fraud, assigns a corresponding value to a combined confidence factor that the preset fraud of the type has the upper level fraud description information, the value is not less than the combined confidence factor that the fraud of the type has any lower level fraud description information of the upper level fraud description information, the upper level fraud description information can be determined as the preset fraud description information, so as to correct the combined confidence factor that the preset fraud of each type has each preset fraud description information through the fraud description information relation diagram after completion, and obtain a selected confidence factor that each type of preset fraud description information has each preset fraud description information.

Step S40: and determining at least one selected fraudulent behavior description information of the object to be identified, and determining the selected fraudulent behavior corresponding to the object to be identified through a plurality of selected fraudulent behavior description information and the selected fraudulent behavior look-up table.

In one embodiment, when determining the selected fraudulent conduct corresponding to the object to be identified, the selected matching coefficient of the object to be identified and each type of preset fraudulent conduct is determined through at least one piece of selected fraudulent conduct description information and the selected fraudulent conduct lookup table of the object to be identified. For example, after the selected matching coefficient of the object to be identified and each type of preset fraudulent conduct is obtained, the preset fraudulent conduct corresponding to at least one maximum selected matching coefficient in the selected matching coefficient is determined as the selected fraudulent conduct of the object to be identified; or for each type of preset fraudulent conduct, determining a score corresponding to the type of preset fraudulent conduct, for example, a selected matching coefficient of the object to be identified and the type of preset fraudulent conduct and a sequence score in all preset fraudulent conduct of the type of preset fraudulent conduct. For example, the preset fraud i corresponds to a score value: ji = (W-Fi, di); di is a selected matching coefficient of the object to be recognized and the preset fraud behaviors in the class; and the value of Fi is 1 to W, and is the score sequence of the preset fraudulent behaviors i in all the preset fraudulent behaviors. And when the W-ri is larger than the preset score sequence and the Di is larger than the preset matching coefficient, determining that the object to be recognized has a preset fraud behavior i.

In one embodiment, when determining the selected matching coefficient between the object to be identified and each preset fraudulent behavior, for each type of preset fraudulent behavior, determining the prominent distribution factor of each preset fraudulent behavior description information to the fraudulent behavior based on the selected fraudulent behavior look-up table, and if the prominent distribution factor of any preset fraudulent behavior description information to the fraudulent behavior is larger, indicating that the degree of the prominent distribution of the preset fraudulent behavior description information to the fraudulent behavior is higher, the prominent distribution factor can be represented by a weight. For each preset fraud description information, when the selected confidence factor that the preset fraud description information of the class of preset fraud is greater than a preset confidence factor threshold value, determining that the preset fraud includes the preset fraud description information. And determining the prominent distribution factor of the preset fraud description information to the fraud through the total number of the preset fraud and the total number of the preset fraud including the preset fraud description information. The fraud activity highlight allocation factor of the preset fraud activity description information may be a value obtained by performing logarithmic solution on a value obtained by dividing the number of preset fraud activities by a value including the total number of preset fraud activities of the preset fraud activity description information.

For each type of preset fraudulent conduct, after the prominent distribution factor of each preset fraudulent conduct description information to the fraudulent conduct is determined through the selected fraudulent conduct query table, the chosen matching coefficient of the object to be identified and the preset fraudulent conduct is determined through the chosen fraudulent conduct description information of the prominent distribution factors corresponding to the multiple preset fraudulent conduct description information. For example, target fraud descriptors in each of the selected fraud descriptors are determined, where each of the selected fraud descriptors does not include any of the target fraud descriptors' subordinate fraud descriptors. That is, the lowest level of selected fraud description information in each selected fraud description information is determined as the target fraud description information, and no such condition exists in each target fraud description information: one target fraud descriptor is a subordinate fraud descriptor of another target fraud descriptor.

Optionally, the preset fraud behavior description information Pre-info1 in each preset fraud behavior description information is determined, the class of preset fraud behaviors has a selected confidence factor of each preset fraud behavior description information Pre-info1 larger than a confidence factor CM4, wherein the confidence factor CM4 is a confidence factor threshold used in obtaining the total number of preset fraud behaviors of the preset fraud behavior description information. Further, a fraudulent behavior coincidence description information Cluster1 of each target fraudulent behavior description information and each preset fraudulent behavior description information Pre-info1 is determined, and a selected matching coefficient of the object to be identified and the preset fraudulent behaviors is determined based on a highlight distribution factor corresponding to each fraudulent behavior description information in the fraudulent behavior coincidence description information Cluster 1.

In addition, in order to increase the matching between the object to be identified and each type of estimated fraudulent conduct, for each type of estimated fraudulent conduct, the matching coefficient MC1 of the object to be identified and each type of estimated fraudulent conduct is determined through the salient distribution factor corresponding to each fraudulent conduct description information in the fraudulent conduct coincidence description information Cluster 1. For a preset fraudulent behavior, the selected matching coefficient may be determined as a matching coefficient MC1 between the object to be identified and the predicted fraudulent behavior, and preset fraudulent behavior description information Pre-info2 in each preset fraudulent behavior description information Pre-info1 is determined, where each preset fraudulent behavior description information Pre-info1 does not include any lower level fraudulent behavior description information of the preset fraudulent behavior description information Pre-info 1. That is, the lowest level of the preset fraud description information Pre-info1 in each preset fraud description information Pre-info1 is determined as the preset fraud description information Pre-info2, and the following conditions do not exist in each preset fraud description information Pre-info 2: one preset fraud descriptor Pre-info2 is a subordinate fraud descriptor of another preset fraud descriptor Pre-info 2. In addition, each preset fraud behavior description information Pre-info2 and the fraud behavior coincidence description information Cluster Cluster2 of each selected fraud behavior description information are determined, and the matching coefficient MC2 of the object to be identified and the class of preset fraud behaviors is determined based on the highlighted distribution factor corresponding to each fraud behavior description information in the fraud behavior coincidence description information Cluster 2.

And for each type of preset fraudulent conduct, after determining the selected matching coefficient of the object to be identified and the type of preset fraudulent conduct, determining the selected matching coefficient of the object to be identified and the type of preset fraudulent conduct through the corresponding matching coefficient MC1 and the matching coefficient MC2. For example, the matching coefficient MC1 and the matching coefficient MC2 are averaged to be the selected matching coefficient. In one embodiment, when the selected matching coefficient of the object to be recognized and each preset fraudulent behavior is determined, for each type of preset fraudulent behavior, the confidence factor CM2 that any object has the preset fraudulent behavior of the type is determined, and then the selected matching coefficient of the object to be recognized and the preset fraudulent behavior of the type is determined according to the selected confidence factor that the preset fraudulent behavior of the type has the description information of each preset fraudulent behavior; and the confidence factor CM2 of the fraud behavior preset by any object in the class is a confidence factor determined in advance by statistics.

For example, target fraud descriptors in each selected fraud descriptor are determined, wherein each selected fraud descriptor does not include any subordinate fraud descriptors of the target fraud descriptor. Namely, the lowest level selected fraud description information in each selected fraud description information is determined as the target fraud description information, and the following conditions do not exist in each target fraud description information: one target fraud descriptor is a lower level fraud descriptor of another target fraud descriptor.

Optionally, when each preset fraud description information includes each target fraud description information, determining that the class of preset fraud has the confidence factor CM3 of each target fraud description information at the same time through the selected confidence factor of each preset fraud description information of the class of preset fraud. Namely, the selected fraud behavior look-up table is selected to determine that the preset fraud behavior of the class has the selected confidence factor of each target fraud behavior description information, then the selected confidence factor that the preset fraud behavior of the class has each target fraud behavior description information is used to determine the confidence factor CM3 that the preset fraud behavior of the class has each target fraud behavior description information, and the confidence factor CM3 is determined as the selected matching coefficient of the object to be recognized and the preset fraud behavior of the class.

For the preset fraud, under the condition that the preset fraud has the selected confidence factors of each target fraud description information, which are not related to each other, the confidence factor CM3 that the preset fraud has all the target fraud description information can be determined by taking continuous product of the selected confidence factors that the preset fraud has the target fraud description information.

In one embodiment, when the selected matching coefficient of the object to be identified and each type of preset fraudulent conduct is determined, the description information of each selected fraudulent conduct of the object to be identified is input into a fraudulent conduct identification network, and the selected matching coefficient of the object to be identified and each type of preset fraudulent conduct is obtained. The fraud identification network is debugged based on the selected fraud look-up table.

In the intelligent financial anti-fraud method and system based on artificial intelligence provided by the embodiment of the application, the confidence factors CM1 of each fraud behavior description information of each financial fraud arrangement set are preliminarily obtained by determining the basic fraud behavior query tables corresponding to each financial fraud arrangement set, and then the basic fraud behavior query tables are fused to obtain the accurate selected fraud behavior query table for indicating that each preset fraud behavior has the selected confidence factor of each preset fraud behavior description information, so that the accuracy and efficiency can be effectively improved when the selected fraud behavior of the object to be identified is determined by the selected fraud behavior query table, and meanwhile, the method and system have excellent applicability, are convenient for timely intervening fraud behaviors, and anti-fraud is realized

In the process of debugging the fraud behavior recognition network, a training template Sample1 set is generated based on a plurality of financial fraud finishing sets, the training template Sample1 set comprises a plurality of training templates Sample1, each training template Sample1 comprises fraud behavior description information when an object has a type of fraud behavior, each training template Sample1 comprises a template supervision mark, and the template supervision mark of each training template Sample1 is used for indicating correct fraud behavior corresponding to the corresponding object. In addition, a fraud query table is selected to generate a training template Sample2 corresponding to each training template Sample1, each training template Sample2 comprises fraud description information that an object has a class of fraud history, each training template Sample2 also comprises a template supervision mark, the template supervision mark of each training template Sample2 is used for indicating correct fraud corresponding to the corresponding object, and each set of training templates Sample1 and training template Sample2 corresponds to the same object and fraud and corresponds to different fraud description information. For each training template, sample1, it is determined that the fraud corresponding to the training template, sample1, has the selected confidence factor of each fraud description information in the training template, sample1, through the selected fraud lookup table. In addition, the fraudulent behavior description information of which the confidence factor is smaller than the generation threshold value is selected from the training templates Sample1 and determined as the fraudulent behavior description information in the corresponding training template Sample2, and the object and the fraudulent behavior corresponding to the training template Sample1 are determined as the object and the fraudulent behavior corresponding to the training template Sample2, so that the corresponding training template Sample2 is generated by changing the fraudulent behavior description information in the training template Sample1 on the premise of ensuring that the object and the fraudulent behavior corresponding to the training template Sample1 are not changed.

For each training template Sample1, because the fraud corresponding to the training template Sample1 has the selected confidence factors of each fraud description information in the training template Sample1 irrelevant to each other, an individual generation threshold value can be determined for each fraud description information, when the selected confidence factor corresponding to the fraud description information is smaller than the corresponding generation threshold value, the fraud description information is retained, if not smaller, the fraud description information is deleted, thus each fraud description information in the corresponding training template Sample2 is obtained, each fraud description information in the training template Sample2 is made to be the fraud description information with the smaller confidence factor corresponding to the fraud, the network capability obtained by debugging is stronger, and the generation threshold values corresponding to each fraud description information can be the same or different and are determined according to the fact. After obtaining each training template Sample2, when the training template Sample2 belongs to the fraud description information of the lower fraud description information of the other upper fraud description information, taking the upper fraud description information corresponding to the fraud description information as the fraud description information in the training template Sample2 at the same time. In addition, after generating each training template Sample2, inputting each fraud description information in each training template Sample1 and each training template Sample2 into an original fraud recognition network, obtaining an estimated matching coefficient of each object corresponding to each training template Sample1 and each training template Sample2 and each preset fraud, determining the estimated fraud of the object corresponding to the training template Sample1 according to the estimated matching coefficient of the object corresponding to each training template Sample1 and each preset fraud, determining the estimated fraud of the object corresponding to the training template Sample2 according to the object corresponding to each training template Sample2 and each preset fraud estimated matching coefficient, determining a convergence evaluation value based on the correct fraud pointed by the template supervision marks of the plurality of training templates Sample1 and each training template Sample2 and the estimated matching coefficients corresponding to each training template Sample1 and each training template Sample2, and obtaining a result of debugging the fraud recognition network when the original fraud recognition network stops. When the training template is in a debugging state, correcting the template supervision mark of the training template Sample2 according to a preset interval and based on the estimated fraudulent behavior of the training template Sample2 at the corresponding moment for each training template Sample2. Wherein the convergence evaluation value (which can be understood as a cost) can be determined by a cross-entropy cost function. Through the network debugging process, semi-supervised learning is carried out when the number of debugging samples is small, the learning capacity and the network identification performance of the fraud identification network can be improved, in addition, for the fraud with low frequency, a large number of debugging templates can be generated for debugging to obtain the fraud identification network, and accurate identification of the fraud with low frequency is completed.

Based on the above embodiments, the present application provides an intelligent financial anti-fraud device, and fig. 3 is an intelligent financial anti-fraud device provided in the present application, as shown in fig. 3, the device 340 includes:

a fraud arrangement determining module 341, configured to determine at least one financial fraud arrangement, where each financial fraud arrangement includes at least one object corresponding fraud and description information of the fraud;

a basic look-up table obtaining module 342, configured to, for each financial fraud arrangement set, obtain a basic fraud behavior look-up table corresponding to the financial fraud arrangement set; the basic fraud behavior look-up table is used for indicating that each type of fraud behavior in the financial fraud settling set has a confidence factor CM1 of each fraud behavior description information in the financial fraud settling set;

a selected look-up table obtaining module 343, configured to determine a selected fraud look-up table based on a plurality of the basic fraud look-up tables; the selected fraud action query table is used for indicating that each type of preset fraud action has a selected confidence factor of each preset fraud action description information, each selected confidence factor is determined based on that the preset fraud action corresponding to the selected confidence factor in a plurality of basic fraud action query tables has not less than one confidence factor CM1 of the preset fraud action description information corresponding to the selected confidence factor, each type of preset fraud action is one type of fraud action in fraud actions corresponding to each financial fraud arrangement set, and each preset fraud action description information is one fraud action description information in fraud action description information corresponding to each financial fraud arrangement set;

the to-be-identified object determining module 344 is configured to determine at least one piece of selected fraud behavior description information of an object to be identified, and determine, through a plurality of pieces of selected fraud behavior description information and the selected fraud behavior lookup table, a selected fraud behavior corresponding to the object to be identified.

The above description of the apparatus embodiments, similar to the above description of the method embodiments, has similar beneficial effects as the method embodiments. For technical details not disclosed in the embodiments of the apparatus of the present application, reference is made to the description of the embodiments of the method of the present application for understanding.

If the technical scheme of the application relates to personal or private information, a product applying the technical scheme of the application clearly informs personal information processing rules before processing the personal information, and obtains personal autonomous consent. If the technical scheme of the application relates to sensitive personal information, before the sensitive personal information is processed, a product applying the technical scheme of the application obtains individual consent, meets the requirement of 'express consent' and is collected in the scope of laws and regulations. For example, at a personal information collection device such as a camera, a clear and significant identifier is set to inform that the personal information collection range is entered, the personal information is collected, and if the person voluntarily enters the collection range, the person is considered as agreeing to collect the personal information; or on the device for processing the personal information, under the condition of informing the personal information processing rule by using obvious identification/information, obtaining personal authorization by modes of popping window information or asking a person to upload personal information of the person by himself, and the like; the personal information processing rule may include information such as a personal information processor, a personal information processing purpose, a processing method, and a type of personal information to be processed.

It should be noted that, in the embodiment of the present application, if the alarm processing method is implemented in the form of a software functional module and is sold or used as an independent product, the alarm processing method may also be stored in a computer-readable storage medium. Based on such understanding, the technical solutions of the embodiments of the present application may be essentially implemented or portions thereof contributing to the related art may be embodied in the form of a software product stored in a storage medium, and including several instructions for enabling an electronic device (which may be a personal computer, a server, or a network device) to execute all or part of the methods described in the embodiments of the present application. And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read Only Memory (ROM), a magnetic disk, or an optical disk. Thus, embodiments of the present application are not limited to any specific combination of hardware and software.

An embodiment of the present application provides an electronic device, which includes a memory and a processor, where the memory stores a computer program that can be run on the processor, and the processor implements the above alarm processing method when executing the computer program.

An embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored, and the computer program, when executed by a processor, implements the above-mentioned alarm processing method. The computer readable storage medium may be transitory or non-transitory.

Embodiments of the present application provide a computer program product, which includes a non-transitory computer readable storage medium storing a computer program, and when the computer program is read and executed by a computer, the computer program implements some or all of the steps of the above method. The computer program product may be embodied in hardware, software or a combination thereof. In an alternative embodiment, the computer program product is embodied in a computer storage medium, and in another alternative embodiment, the computer program product is embodied in a Software product, such as a Software Development Kit (SDK), or the like.

It should be noted that fig. 4 is a schematic diagram of hardware entities of an intelligent financial anti-fraud system 300 according to an embodiment of the present application, and as shown in fig. 4, the hardware entities of the intelligent financial anti-fraud system 300 include: a processor 310, a communication interface 320, and a memory 330, wherein: the processor 310 generally controls the overall operation of the intelligent financial anti-fraud system 300. The communication interface 320 may enable the electronic device to communicate with other terminals or servers via a network. The Memory 330 is configured to store instructions and applications executable by the processor 310, and may also buffer data (e.g., image data, audio data, voice communication data, and video communication data) to be processed or already processed by the processor 310 and modules in the intelligent financial anti-fraud system 300, and may be implemented by a FLASH Memory (FLASH) or a Random Access Memory (RAM). Data may be transferred between processor 310, communication interface 320, and memory 330 via bus 340. Here, it should be noted that: the above description of the storage medium and device embodiments is similar to the description of the method embodiments above, with similar advantageous effects as the method embodiments. For technical details not disclosed in the embodiments of the storage medium and apparatus of the present application, reference is made to the description of the embodiments of the method of the present application for understanding.

It should be appreciated that reference throughout this specification to "one embodiment" or "an embodiment" means that a particular feature, structure or characteristic described in connection with the embodiment is included in at least one embodiment of the present application. Thus, the appearances of the phrases "in one embodiment" or "in an embodiment" in various places throughout this specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. It should be understood that, in the various embodiments of the present application, the sequence numbers of the above-mentioned processes do not mean the execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present application. The above-mentioned serial numbers of the embodiments of the present application are merely for description and do not represent the merits of the embodiments.

It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or 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 apparatus. Without further limitation, an element defined by the phrases "comprising a component of' 8230; \8230;" does not exclude the presence of another like element in a process, method, article, or apparatus that comprises the element.

In the several embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other manners. The above-described device embodiments are merely illustrative, for example, the division of the unit is only a logical functional division, and there may be other division ways in actual implementation, such as: multiple units or components may be combined, or may be integrated into another system, or some features may be omitted, or not implemented. In addition, the coupling, direct coupling or communication connection between the components shown or discussed may be through some interfaces, and the indirect coupling or communication connection between the devices or units may be electrical, mechanical or other forms.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units; can be located in one place or distributed on a plurality of network units; some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, all the functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may be separately regarded as one unit, or two or more units may be integrated into one unit; the integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional unit.

Those of ordinary skill in the art will understand that: all or part of the steps of implementing the method embodiments may be implemented by hardware related to program instructions, and the program may be stored in a computer-readable storage medium, and when executed, executes the steps including the method embodiments; and the aforementioned storage medium includes: various media that can store program codes, such as a removable Memory device, a Read Only Memory (ROM), a magnetic disk, or an optical disk.

Alternatively, the integrated units described above in the present application may be stored in a computer-readable storage medium if they are implemented in the form of software functional modules and sold or used as independent products. Based on such understanding, the technical solutions of the present application may be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for causing an electronic device (which may be a personal computer, a server, or a network device) to execute all or part of the methods described in the embodiments of the present application. And the aforementioned storage medium includes: various media that can store program code, such as removable storage devices, ROMs, magnetic or optical disks, etc.

The above description is only an embodiment of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily think of the changes or substitutions within the technical scope of the present application, and shall cover the scope of the present application.

Claims

1. An intelligent financial anti-fraud method based on artificial intelligence is applied to an intelligent financial anti-fraud system, and comprises the following steps:

for each financial fraud arrangement set, acquiring a basic fraud behavior query table corresponding to the financial fraud arrangement set; the basic fraud behavior lookup table is used for indicating that each type of fraud behavior in the financial fraud arrangement set has a confidence factor CM1 of each fraud behavior description information in the financial fraud arrangement set;

and determining at least one piece of selected fraudulent conduct description information of the object to be identified, and determining the selected fraudulent conduct corresponding to the object to be identified through a plurality of pieces of selected fraudulent conduct description information and the selected fraudulent conduct query table.

2. The method of claim 1, wherein for each financial fraud collation, said obtaining a base fraud behavior look-up table corresponding to the financial fraud collation comprises:

for each type of fraud behaviors in the financial fraud arrangement set and each fraud behavior description information, acquiring a first number of objects having the type of fraud behaviors in the financial fraud arrangement set and a second number of objects having the type of fraud behaviors and having the fraud behavior description information, and determining a confidence factor CM1 that the type of fraud behaviors have the fraud behavior description information according to the first number and the second number;

3. The method of claim 1, wherein determining the selected fraud lookup table based on a plurality of the base fraud lookup tables comprises:

4. The method according to claim 3, wherein determining the selected fraud lookup table based on the combined confidence factor that each of the predetermined fraud profiles of each of the predetermined fraud profiles includes:

determining a fraud behavior description information relation graph corresponding to each preset fraud behavior description information; one component in the fraud description information relation diagram is used for indicating one piece of preset fraud description information, and the preset fraud description information pointed by the lower component of one component in the fraud description information relation diagram is the lower fraud description information of the preset fraud description information pointed by the component;

and correcting the merged confidence factor of each preset fraud behavior description information of each preset fraud behavior based on the fraud behavior description information relation graph to obtain the selected confidence factor of each preset fraud behavior description information of each preset fraud behavior.

5. The method according to claim 4, wherein for each type of the preset fraudulent conduct and each of the preset fraudulent conduct description information, modifying the merged confidence factor that the preset fraudulent conduct of the type has the preset fraudulent conduct description information based on the fraudulent conduct description information relation diagram to obtain the selected confidence factor that the preset fraudulent conduct of the type has the preset fraudulent conduct description information, includes:

when the combined confidence factor of the preset fraud behavior description information of the class of preset fraud behaviors is smaller than or equal to the preset confidence factor and a second component in the fraud behavior description information relation graph comprises not less than one third component, determining that the class of preset fraud behaviors has the selected confidence factor of the preset fraud behavior description information based on the selected confidence factor of the preset fraud behavior description information pointed by a plurality of lower components of the first component of the class of preset fraud behaviors; the second elements comprise all lower-level elements of the first elements and elements with derivative relations with the first elements, and the preset fraud behaviors are similar to each other, and the combined confidence factor of the preset fraud behavior description information used for indicating by each third element is larger than the preset confidence factor;

6. The method according to any one of claims 1 to 5, wherein the determining the selected fraudulent behavior of the object to be identified by the plurality of selected fraudulent behavior description information and the selected fraudulent behavior look-up table comprises:

determining the selected fraudulent conduct corresponding to the object to be identified based on the selected matching coefficient of the object to be identified and each type of preset fraudulent conduct;

determining a prominent distribution factor of each preset fraud description information to fraud based on the selected fraud lookup table, and determining a selected matching coefficient of the object to be identified and the class of preset fraud through the prominent distribution factors corresponding to a plurality of preset fraud description information and each selected fraud description information;

inputting each piece of selected fraudulent conduct description information into a fraudulent conduct identification network to obtain a selected matching coefficient of the object to be identified and the class of preset fraudulent conduct; the fraud identification network is obtained by debugging based on the selected fraud lookup table.

7. The method according to claim 6, wherein for each class of the preset fraudulent activity, the determining the selected matching coefficient between the object to be identified and the class of the preset fraudulent activity by using the prominent distribution factor corresponding to the plurality of preset fraudulent activity description information and each of the selected fraudulent activity description information comprises:

determining target fraud description information in each of the selected fraud description information; wherein each of the selected fraud description information does not include subordinate fraud description information of the target fraud description information;

determining a fraud behavior coincidence description information Cluster1 of each target fraud behavior description information and preset fraud behavior description information Pre-info1 in each preset fraud behavior description information, and determining a selected matching coefficient of the object to be identified and the class of preset fraud behaviors based on a prominent distribution factor corresponding to each fraud behavior description information in the fraud behavior coincidence description information Cluster 1; the class of preset fraud has a selected confidence factor for each of the preset fraud description information Pre-info1 that is greater than a confidence factor CM4.

8. The method according to claim 7, wherein for each type of the preset fraudulent conduct, the determining the selected matching coefficient of the object to be identified and the class of the preset fraudulent conduct based on the highlighted distribution factor corresponding to each fraudulent conduct description information in the fraudulent conduct coincidence description information Cluster1 includes:

determining preset fraud description information Pre-info2 in each preset fraud description information Pre-info 1; each piece of preset fraud description information Pre-info1 does not contain lower-level fraud description information of the preset fraud description information Pre-info2;

determining a fraudulent behavior coincidence description information Cluster2 of each second fraudulent behavior description information and each selected fraudulent behavior description information, and determining a matching coefficient MC2 of the object to be identified and the class of preset fraudulent behaviors based on a highlight distribution factor corresponding to each fraudulent behavior description information in the fraudulent behavior coincidence description information Cluster 2;

9. The method according to claim 6, wherein the determining the selected matching coefficient between the object to be recognized and the predetermined fraud is determined according to the confidence factor CM2 and the predetermined fraud is the selected confidence factor of each of the predetermined fraud description information, and includes:

determining target fraud description information in each of the selected fraud description information; wherein each of the selected fraud descriptor does not include a subordinate fraud descriptor of the target fraud descriptor;

determining a confidence factor CM3 that the class of preset fraudulent conduct has the target fraudulent conduct description information at the same time based on the fact that the class of preset fraudulent conduct has the selected confidence factor of each preset fraudulent conduct description information;

and determining a selected matching coefficient of the object to be recognized and the preset fraud behaviors according to the confidence factor CM2 and the confidence factor CM3.

10. An intelligent financial anti-fraud system comprising a processor and a memory, the memory storing a computer program for implementing the method of any one of claims 1 to 9 when executed by the processor.