CN111127185A - Credit fraud identification model construction method and device - Google Patents
Credit fraud identification model construction method and device Download PDFInfo
- Publication number
- CN111127185A CN111127185A CN201911167779.5A CN201911167779A CN111127185A CN 111127185 A CN111127185 A CN 111127185A CN 201911167779 A CN201911167779 A CN 201911167779A CN 111127185 A CN111127185 A CN 111127185A
- Authority
- CN
- China
- Prior art keywords
- model
- behavior
- credit
- variables
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/03—Credit; Loans; Processing thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2465—Query processing support for facilitating data mining operations in structured databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2474—Sequence data queries, e.g. querying versioned data
Abstract
The invention provides a credit fraud identification model construction method and a credit fraud identification model construction device, wherein the method comprises the following steps: constructing a relational network according to the relational data of the credit customer, generating a network characteristic vector and constructing a first depth model; constructing a behavior sequence according to the behavior data of the credit customer, generating a behavior feature vector and constructing a second depth model; performing feature engineering processing to extract attribute feature variables, relationship feature variables, behavior feature variables and rule variables of the credit customer and construct a width model; the depth model and the width model are jointly trained to build a credit fraud recognition model. In the invention, the fraud identification model is constructed by fusing the anti-fraud rule, the relationship network characteristic, the attribute characteristic, the behavior characteristic and the like, so that the method has better adaptability and accuracy for identifying the fraud risk.
Description
Technical Field
The invention relates to the field of machine learning, in particular to a credit fraud identification model construction method and device.
Background
In the credit field, fraud risk is an important content in wind control management, and with the continuous development of popular finance, more and more institutions continue to expand credit business and fraud cases also show a growing trend. The traditional fraud identification method is that a wind control expert identifies fraud by establishing a rule against fraud, but as the industry develops, the ways of fraud and means of fraud are constantly changing. In recent years, with the development of machine learning technology, some financial institutions have started to introduce machine learning into fraud risk identification, and identification of fraud risk is performed by establishing an anti-fraud model based on anti-fraud rules by means of supervised classification.
In the fraud identification mode based on the anti-fraud rules, the wind control expert formulates the anti-fraud rules through the expert experience of the wind control expert, and if the credit application client hits one or more rules, the client is regarded as a fraudulent client. Or a fraud identification model based on the rule characteristics is built, a logistics regression model is built by extracting the rule characteristics of the historical clients and the characteristics of the application information, the new application clients are scored according to the model, and the clients with high scores are regarded as the fraud clients.
However, for the anti-fraud rule-based approach, the specification of the rule relies primarily on historical data and expert experience, and the rule is likely to work when the means and manner of fraud change. The anti-fraud rules are effective in aiming at the occurring fraud cases, but the anti-fraud rules are often ineffective for novel fraud modes.
In a fraud identification mode based on a relationship network, a relationship network between clients is established through the association relationship (such as contacts, relatives, friends, family addresses, company names, company addresses, company telephones and the like) between the clients, and a wind control staff judges whether a new client is fraudulent or not or whether the new client is group fraud or not by checking abnormal point information such as the consistency of association rules (such as the same company telephone and different company names and the like) in a visual mode; an organization constructs a scoring model to score the customers by extracting the first-degree relation characteristics, the second-degree relation characteristics and the network characteristics (such as the degrees of nodes, edge betweenness and the like) of the customers in the relation network, and the customers with high scores are judged as the fraudulent customers; there is also an organization to calculate the fraud risk of new customers based on whether historical customers are fraudulent, using algorithms for tag propagation over the relationship network.
However, for the fraud identification method based on the relational network, no matter the consistency of the association rules, the first-stage and second-stage relational characteristics and the network characteristics are extracted based on expert experience on the basis of the relational network, certain information compression exists, the extraction of the characteristics also depends on the experience of the experts, and the homogeneity and the structural equivalence of the network cannot be well expressed.
Disclosure of Invention
The embodiment of the invention provides a credit fraud identification model construction method and a credit fraud identification model construction device, which are used for at least solving the problem that credit risk identification cannot be effectively carried out in a credit fraud identification mode based on an anti-fraud rule or based on a relationship network in the related art.
According to an embodiment of the invention, there is provided a credit fraud identification model construction method including: establishing a relational network according to the relational data of the credit customer, generating a network characteristic vector, and establishing a first depth model according to the network characteristic vector; constructing a behavior sequence according to the behavior data of the credit customer, generating a behavior feature vector, and constructing a second depth model according to the behavior feature vector; performing feature engineering processing on the attribute data, the relationship data and the behavior data to extract attribute feature variables, relationship feature variables, behavior feature variables and rule variables of the credit customer, and constructing a width model according to the attribute feature variables, the relationship feature variables, the behavior feature variables and the rule variables; the first depth model, the second depth model, and the width model are jointly trained to construct a credit fraud recognition model.
Optionally, before the building the relationship network according to the relationship data of the credit customer, the method further includes: and acquiring attribute data, relationship data and behavior data of the credit customer.
Optionally, the relationship network and the behavior sequence are respectively converted into the network feature vector and the behavior feature vector by using an embedding algorithm.
Optionally, after the jointly training the first depth model, the second depth model and the width model to construct the credit fraud recognition model, the method further includes: and using a weighted sum of the predictions based on the first depth model, the second depth model and the width model as the credit customer fraud risk prediction.
According to another embodiment of the present invention, there is provided a credit fraud identification model building apparatus including: the first depth model building module is used for building a relational network according to the relational data of the credit customers, generating network characteristic vectors and building a first depth model according to the network characteristic vectors; the second depth model building module is used for building a behavior sequence according to the behavior data of the credit customer, generating a behavior feature vector and building a second depth model according to the behavior feature vector; the width model building module is used for performing feature engineering processing on the attribute data, the relationship data and the behavior data to extract attribute characteristic variables, relationship characteristic variables, behavior characteristic variables and rule variables of the credit customer and building a width model according to the attribute characteristic variables, the relationship characteristic variables, the behavior characteristic variables and the rule variables; and the joint training module is used for carrying out joint training on the first depth model, the second depth model and the width model so as to construct a credit fraud recognition model.
Optionally, the apparatus further comprises: and the acquisition module is used for acquiring the attribute data, the relationship data and the behavior data of the credit customer.
Optionally, the relationship network and the behavior sequence are respectively converted into the network feature vector and the behavior feature vector by using an embedding algorithm.
Optionally, the apparatus further comprises: a prediction module to use a weighted sum of the predictions based on the first depth model, the second depth model, and the width model as the credit customer fraud risk prediction.
According to a further embodiment of the present invention, there is also provided a storage medium having a computer program stored therein, wherein the computer program is arranged to perform the steps of any of the above method embodiments when executed.
According to yet another embodiment of the present invention, there is also provided an electronic device, including a memory in which a computer program is stored and a processor configured to execute the computer program to perform the steps in any of the above method embodiments.
In the embodiment of the invention, the fraud identification model is constructed by fusing the anti-fraud rule, the relationship network characteristic, the attribute characteristic, the behavior characteristic and the like, so that the method has better adaptability and accuracy for identifying the fraud risk.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the invention without limiting the invention. In the drawings:
FIG. 1 is a flow diagram of a credit fraud identification model construction method according to an embodiment of the invention;
FIG. 2 is a flow diagram of a credit fraud identification model construction method according to an alternative embodiment of the invention;
FIG. 3 is a data processing flow diagram according to an embodiment of the invention;
FIG. 4 is a schematic diagram of the structure of a credit fraud identification model building apparatus according to an embodiment of the present invention;
FIG. 5 is a block diagram of a credit fraud identification model building apparatus according to an alternative embodiment of the invention.
Detailed Description
The invention will be described in detail hereinafter with reference to the accompanying drawings in conjunction with embodiments. It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict.
It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order.
In the embodiment, a credit fraud identification model construction method is provided, fig. 1 is a flow chart of the method according to the embodiment of the invention, and as shown in fig. 1, the flow chart comprises the following steps:
step S102, a relational network is built according to the relational data of the credit customer, a network feature vector is generated, and a first depth model is built according to the network feature vector;
step S104, constructing a behavior sequence according to the behavior data of the credit customer, generating a behavior feature vector, and constructing a second depth model according to the behavior feature vector;
step S106, performing feature engineering processing on the attribute data, the relationship data and the behavior data to extract attribute feature variables, relationship feature variables, behavior feature variables and rule variables of the credit customer and construct a width model;
and step S108, performing joint training on the first depth model, the second depth model and the width model to construct a credit fraud recognition model.
Before step S102 in this embodiment, the method further includes: and acquiring attribute data, relationship data and behavior data of the credit customer.
In steps S102 and S104 of the present embodiment, an embedding algorithm may be adopted to convert the relationship network and the behavior sequence into the network feature vector and the behavior feature vector, respectively.
In step S108 of the present embodiment, a weighted sum of the prediction results based on the first depth model, the second depth model, and the width model may be used as the fraud risk prediction result of the credit customer.
In order to facilitate understanding of the technical solutions provided by the present invention, the following detailed description is made with reference to specific scenario embodiments.
In the credit fraud identification model construction method provided by the embodiment, the anti-fraud rule, the relationship network characteristics, the attribute characteristics, the behavior characteristics and the like can be fused to construct the fraud identification model, so that the method has better adaptability and accuracy rate for identifying fraud risks.
The embodiment mainly describes a method for constructing a fraud identification model by fusing attribute characteristics, relationship characteristics, behavior characteristics and rule characteristics of a customer in a credit scene.
As shown in fig. 2, the main steps are as follows:
step S201, customer information acquisition. Attribute data of the credit customer such as sex, age, address, company and the like, relationship data of spouse, parent, friend, colleague, mobile phone online person and the like, and behavior data of using the app in the credit application are obtained.
Step S202, a relational network is constructed and a network feature vector is generated. And constructing a relationship network according to relationship data of the client, wherein the relationship data comprises strong association relationships such as spouses, parents, friends, colleagues, contacts and the like, and weak association relationships such as companies, call records, addresses and the like. In the relation network, each client takes an identity card number as a unique identifier as a client node, each contact comprises a communication contact and takes a mobile phone number as a unique identifier as a contact node, and in addition, the communication contact also comprises an address node and a company node. Based on the constructed relationship network, a network embedding algorithm, such as algorithms of Deepwalk, Node2Vec, LINE, TADW, CENE and the like, is adopted to generate vectorization expression of the client relationship network.
Step S203, constructing a behavior sequence and generating a behavior feature vector. And extracting browsed pages and operated controls according to browsing and operating logs of a client in the APP, and generating a behavior sequence according to the time sequence. For example: page A, page B, input box C and selection box D. And then according to a skip-gram algorithm, embedding the behavior sequence of the user to obtain the vectorized expression of the browsing behavior of the client.
And step S204, processing the characteristic engineering. According to the experience summarized by the wind control experts, feature engineering processing is carried out on the customer data according to the attribute data, the relationship network, the APP behavior data and the like, and the construction attribute feature variable, the relationship feature variable, the behavior feature variable and the like of the customer are extracted. Such as the age, gender, number of overdue days of credit card, the same unit name of the associated historical customer unit, and the dwell time of page a.
In step S205, the vectorization of the customer relationship network in step S202 and the vectorization of the customer behavior in step S203 are used to construct a depth (deep) model, and the feature variable user in step S204 is used to construct a width (wide) model.
The width model is suitable for discrete features, a large number of nonlinear discrete features are constructed by a series of discrete artificial features through cross product (cross product), the correlation among the features is learned, and for continuous features, the continuous features are often converted into discrete features through a binning method, so that the model has good interpretability and memory, but the generalization capability is weak. The depth model is suitable for continuous features, and the feature correlation is captured by learning the continuous features or performing embedding processing on discrete features into the continuous features. According to the method, the width model and the depth model are combined in a joint training mode, discrete features and continuous features are processed respectively, the memorability and the interpretability of the width model and the generalization ability of the depth model can be combined, known fraud risks can be identified, and more concealed and unknown fraud risks can be learned through better utilization.
And step S206, performing combined training on the depth model and the width model. And taking the weighted sum of the depth model result and the width model result as a prediction result. And simultaneously optimizing parameters of the depth model and the width model according to the historical data, and finally obtaining a weighted sum of the results of the two models as a prediction result. Wherein the distribution of each model weight can be learned through joint training. In the joint training, not only the parameters of the respective models but also the weighted weights of the two model results can be learned.
Fig. 3 shows a specific data processing flow of the present embodiment. As shown in fig. 3, in the model construction method provided in this embodiment, a relationship network and a behavior sequence of a client are converted into a vector through embedding, and the vector is used for training a depth model. And the attribute characteristics, the rule characteristics, the relation characteristics and the behavior characteristics generated by the characteristic engineering are used for training the linear model. The depth model and the linear model are then trained simultaneously, and the weighted sum of the results of the two models is used as the final predicted fraud recognition result. The method ensures that the fraud identification is a colleague of accuracy, improves the generalization capability of the model, finds out the fraud characteristics which can not be summarized by the expert experience, and can identify a novel fraud mode more timely and improve the adaptability of the model.
Through the above description of the embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but the former is a better implementation mode in many cases. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which is stored in a storage medium (e.g., ROM/RAM, magnetic disk, optical disk) and includes instructions for enabling a terminal device (e.g., a mobile phone, a computer, a server, or a network device) to execute the method according to the embodiments of the present invention.
In this embodiment, a credit fraud identification model construction apparatus is further provided, and the apparatus is used to implement the foregoing embodiments and preferred embodiments, which have already been described and are not described again. As used below, the term "module" may be a combination of software and/or hardware that implements a predetermined function. Although the means described in the embodiments below are preferably implemented in software, an implementation in hardware, or a combination of software and hardware is also possible and contemplated.
Fig. 4 is a block diagram showing the construction of the credit fraud recognition model building apparatus according to the embodiment of the present invention, which includes a first depth model building module 10, a second depth model building module 20, a width model building module 30, and a joint training module 40, as shown in fig. 4.
The first depth model building module 10 is used for building a relational network according to the relational data of the credit customers, generating network characteristic vectors and building a first depth model according to the network characteristic vectors; .
The second depth model building module 20 is used for building a behavior sequence according to the behavior data of the credit customer, generating a behavior feature vector and building a second depth model according to the behavior feature vector;
the width model building module 30 is configured to perform feature engineering processing on the attribute data, the relationship data and the behavior data to extract attribute feature variables, relationship feature variables, behavior feature variables and rule variables of the credit customer, and build a width model according to the attribute feature variables, the relationship feature variables, the behavior feature variables and the rule variables;
and the joint training module 40 is used for performing joint training on the first depth model, the second depth model and the width model to construct a credit fraud recognition model.
Fig. 5 is a block diagram of the construction of the credit fraud identification model building apparatus according to an alternative embodiment of the present invention, which includes an acquisition module 50 and a prediction module 60, as shown in fig. 5, in addition to all of the modules shown in fig. 4.
An obtaining module 50 for obtaining attribute data, relationship data and behavior data of the credit customer.
A prediction module 60 for using a weighted sum of the predictions based on the first depth model, the second depth model and the width model as the credit customer fraud risk prediction.
It should be noted that, the above modules may be implemented by software or hardware, and for the latter, the following may be implemented, but not limited to: the modules are all positioned in the same processor; alternatively, the modules are respectively located in different processors in any combination.
Embodiments of the present invention also provide a storage medium having a computer program stored therein, wherein the computer program is arranged to perform the steps of any of the above method embodiments when executed.
Optionally, in this embodiment, the storage medium may include, but is not limited to: various media capable of storing computer programs, such as a usb disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic disk, or an optical disk.
Embodiments of the present invention also provide an electronic device comprising a memory having a computer program stored therein and a processor arranged to run the computer program to perform the steps of any of the above method embodiments.
Optionally, the electronic apparatus may further include a transmission device and an input/output device, wherein the transmission device is connected to the processor, and the input/output device is connected to the processor.
It will be apparent to those skilled in the art that the modules or steps of the present invention described above may be implemented by a general purpose computing device, they may be centralized on a single computing device or distributed across a network of multiple computing devices, and alternatively, they may be implemented by program code executable by a computing device, such that they may be stored in a storage device and executed by a computing device, and in some cases, the steps shown or described may be performed in an order different than that described herein, or they may be separately fabricated into individual integrated circuit modules, or multiple ones of them may be fabricated into a single integrated circuit module. Thus, the present invention is not limited to any specific combination of hardware and software.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. A credit fraud identification model construction method, comprising:
establishing a relational network according to the relational data of the credit customer, generating a network characteristic vector, and establishing a first depth model according to the network characteristic vector;
constructing a behavior sequence according to the behavior data of the credit customer, generating a behavior feature vector, and constructing a second depth model according to the behavior feature vector;
performing feature engineering processing on the attribute data, the relationship data and the behavior data to extract attribute feature variables, relationship feature variables, behavior feature variables and rule variables of the credit customer, and constructing a width model according to the attribute feature variables, the relationship feature variables, the behavior feature variables and the rule variables;
the first depth model, the second depth model, and the width model are jointly trained to construct a credit fraud recognition model.
2. The method according to claim 1, before building a relationship network from the credit customer's relationship data, further comprising:
and acquiring attribute data, relationship data and behavior data of the credit customer.
3. The method of claim 1, wherein the relational network and the behavior sequence are converted into the network feature vector and the behavior feature vector respectively by using an embedding algorithm.
4. The method of claim 1, wherein after jointly training the first depth model, the second depth model, and the width model to build a credit fraud recognition model, further comprising:
and using a weighted sum of the predictions based on the first depth model, the second depth model and the width model as the credit customer fraud risk prediction.
5. A credit fraud identification model building apparatus, comprising:
the first depth model building module is used for building a relational network according to the relational data of the credit customers, generating network characteristic vectors and building a first depth model according to the network characteristic vectors;
the second depth model building module is used for building a behavior sequence according to the behavior data of the credit customer, generating a behavior feature vector and building a second depth model according to the behavior feature vector;
the width model building module is used for performing feature engineering processing on the attribute data, the relationship data and the behavior data to extract attribute characteristic variables, relationship characteristic variables, behavior characteristic variables and rule variables of the credit customer and building a width model according to the attribute characteristic variables, the relationship characteristic variables, the behavior characteristic variables and the rule variables;
and the joint training module is used for carrying out joint training on the first depth model, the second depth model and the width model so as to construct a credit fraud recognition model.
6. The apparatus of claim 5, further comprising:
and the acquisition module is used for acquiring the attribute data, the relationship data and the behavior data of the credit customer.
7. The apparatus of claim 5, wherein the relationship network and the behavior sequence are respectively converted into the network feature vector and the behavior feature vector by using an embedding algorithm.
8. The apparatus of claim 5, further comprising:
a prediction module to use a weighted sum of the predictions based on the first depth model, the second depth model, and the width model as the credit customer fraud risk prediction.
9. A computer-readable storage medium, in which a computer program is stored, wherein the computer program is arranged to perform the method of any of claims 1 to 4 when executed.
10. An electronic device comprising a memory and a processor, wherein the memory has stored therein a computer program, and wherein the processor is arranged to execute the computer program to perform the method of any of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911167779.5A CN111127185A (en) | 2019-11-25 | 2019-11-25 | Credit fraud identification model construction method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911167779.5A CN111127185A (en) | 2019-11-25 | 2019-11-25 | Credit fraud identification model construction method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111127185A true CN111127185A (en) | 2020-05-08 |
Family
ID=70496542
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911167779.5A Withdrawn CN111127185A (en) | 2019-11-25 | 2019-11-25 | Credit fraud identification model construction method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111127185A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112116356A (en) * | 2020-09-28 | 2020-12-22 | 中国银行股份有限公司 | Asset characteristic information processing method and device |
| CN113408207A (en) * | 2021-06-24 | 2021-09-17 | 上海硕恩网络科技股份有限公司 | Data mining method based on social network analysis technology |
| CN116205664A (en) * | 2023-04-28 | 2023-06-02 | 成都新希望金融信息有限公司 | Intermediary fraud identification method and device, electronic equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108537587A (en) * | 2018-04-03 | 2018-09-14 | 广州优视网络科技有限公司 | It is lost in user's method for early warning, device, computer readable storage medium and server |
| CN108564460A (en) * | 2018-01-12 | 2018-09-21 | 阳光财产保险股份有限公司 | Real-time fraud detection method under internet credit scene and device |
| US20180307986A1 (en) * | 2017-04-20 | 2018-10-25 | Sas Institute Inc. | Two-phase distributed neural network training system |
| CN109685647A (en) * | 2018-12-27 | 2019-04-26 | 阳光财产保险股份有限公司 | The training method of credit fraud detection method and its model, device and server |
| CN110222816A (en) * | 2019-04-29 | 2019-09-10 | 北京迈格威科技有限公司 | Method for building up, image processing method and the device of deep learning model |
| CN110298663A (en) * | 2018-03-22 | 2019-10-01 | 中国银联股份有限公司 | Based on the wide fraudulent trading detection method learnt deeply of sequence |
-
2019
- 2019-11-25 CN CN201911167779.5A patent/CN111127185A/en not_active Withdrawn
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180307986A1 (en) * | 2017-04-20 | 2018-10-25 | Sas Institute Inc. | Two-phase distributed neural network training system |
| CN108564460A (en) * | 2018-01-12 | 2018-09-21 | 阳光财产保险股份有限公司 | Real-time fraud detection method under internet credit scene and device |
| CN110298663A (en) * | 2018-03-22 | 2019-10-01 | 中国银联股份有限公司 | Based on the wide fraudulent trading detection method learnt deeply of sequence |
| CN108537587A (en) * | 2018-04-03 | 2018-09-14 | 广州优视网络科技有限公司 | It is lost in user's method for early warning, device, computer readable storage medium and server |
| CN109685647A (en) * | 2018-12-27 | 2019-04-26 | 阳光财产保险股份有限公司 | The training method of credit fraud detection method and its model, device and server |
| CN110222816A (en) * | 2019-04-29 | 2019-09-10 | 北京迈格威科技有限公司 | Method for building up, image processing method and the device of deep learning model |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112116356A (en) * | 2020-09-28 | 2020-12-22 | 中国银行股份有限公司 | Asset characteristic information processing method and device |
| CN112116356B (en) * | 2020-09-28 | 2023-11-21 | 中国银行股份有限公司 | Asset characteristic information processing method and device |
| CN113408207A (en) * | 2021-06-24 | 2021-09-17 | 上海硕恩网络科技股份有限公司 | Data mining method based on social network analysis technology |
| CN116205664A (en) * | 2023-04-28 | 2023-06-02 | 成都新希望金融信息有限公司 | Intermediary fraud identification method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107872436A (en) | A kind of account recognition methods, apparatus and system | |
| CN111127185A (en) | Credit fraud identification model construction method and device | |
| CN112308173B (en) | Multi-target object evaluation method based on multi-evaluation factor fusion and related equipment thereof | |
| CN111090807A (en) | Knowledge graph-based user identification method and device | |
| CN107368499B (en) | Client label modeling and recommending method and device | |
| CN114693192A (en) | Wind control decision method and device, computer equipment and storage medium | |
| CN113011884A (en) | Account feature extraction method, device and equipment and readable storage medium | |
| CN113641797A (en) | Data processing method, device, equipment, storage medium and computer program product | |
| CN112328657A (en) | Feature derivation method, feature derivation device, computer equipment and medium | |
| CN110197426A (en) | A kind of method for building up of credit scoring model, device and readable storage medium storing program for executing | |
| CN113011966A (en) | Credit scoring method and device based on deep learning | |
| CN112016850A (en) | Service evaluation method and device | |
| CN113254788B (en) | Big data based recommendation method and system and readable storage medium | |
| CN111465021B (en) | Graph-based crank call identification model construction method | |
| CN115099875A (en) | Data classification method based on decision tree model and related equipment | |
| CN115099988A (en) | Model training method, data processing method, device and computer medium | |
| CN110163761B (en) | Suspicious item member identification method and device based on image processing | |
| CN109474703B (en) | Personalized product combination pushing method, device and system | |
| CN113641767A (en) | Entity relationship extraction method, device, equipment and storage medium | |
| CN112561538A (en) | Risk model creation method and device, computer equipment and readable storage medium | |
| CN113011875A (en) | Text processing method and device, computer equipment and storage medium | |
| CN113076450B (en) | Determination method and device for target recommendation list | |
| CN117151727B (en) | Customer service intelligent switching method based on user behavior analysis | |
| CN115879826B (en) | Fine chemical process quality inspection method, system and medium based on big data | |
| CN116934092A (en) | Operation risk determining method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20200508 |