CN116757835A - Method and device for monitoring transaction risk in credit card customer credit - Google Patents
Method and device for monitoring transaction risk in credit card customer credit Download PDFInfo
- Publication number
- CN116757835A CN116757835A CN202310805182.9A CN202310805182A CN116757835A CN 116757835 A CN116757835 A CN 116757835A CN 202310805182 A CN202310805182 A CN 202310805182A CN 116757835 A CN116757835 A CN 116757835A
- Authority
- CN
- China
- Prior art keywords
- risk
- transaction
- vector
- information
- feature vector
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 67
- 238000012544 monitoring process Methods 0.000 title claims abstract description 26
- 239000013598 vector Substances 0.000 claims abstract description 179
- 238000001514 detection method Methods 0.000 claims abstract description 76
- 238000012545 processing Methods 0.000 claims abstract description 25
- 238000000605 extraction Methods 0.000 claims abstract description 18
- 238000012954 risk control Methods 0.000 claims abstract description 14
- 238000004590 computer program Methods 0.000 claims description 23
- 230000008451 emotion Effects 0.000 claims description 14
- 238000012549 training Methods 0.000 claims description 11
- 239000002131 composite material Substances 0.000 claims 1
- 238000013473 artificial intelligence Methods 0.000 abstract description 4
- 230000003993 interaction Effects 0.000 description 12
- 238000010586 diagram Methods 0.000 description 9
- 238000005516 engineering process Methods 0.000 description 8
- 230000008569 process Effects 0.000 description 8
- 230000006870 function Effects 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 238000012795 verification Methods 0.000 description 5
- 239000000284 extract Substances 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000012806 monitoring device Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 238000013136 deep learning model Methods 0.000 description 2
- 230000008909 emotion recognition Effects 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000007935 neutral effect Effects 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000000750 progressive effect Effects 0.000 description 2
- 239000002699 waste material Substances 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 239000013604 expression vector Substances 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
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
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/096—Transfer learning
-
- 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
- G06Q20/00—Payment architectures, schemes or protocols
- G06Q20/38—Payment protocols; Details thereof
- G06Q20/40—Authorisation, e.g. identification of payer or payee, verification of customer or shop credentials; Review and approval of payers, e.g. check credit lines or negative lists
- G06Q20/401—Transaction verification
- G06Q20/4016—Transaction verification involving fraud or risk level assessment in transaction processing
-
- 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
- G06Q40/04—Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Accounting & Taxation (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Finance (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Technology Law (AREA)
- Computing Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Software Systems (AREA)
- Marketing (AREA)
- Mathematical Physics (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computer Security & Cryptography (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The application discloses a method and a device for monitoring transaction risk in credit card client credit, which can be used in the technical field of artificial intelligence, and the method comprises the following steps: extracting risk trend vectors of a plurality of risks corresponding to the client information, carrying out feature extraction on the transaction information through a multi-task learning network to obtain a comprehensive feature vector, and obtaining a low-order cross feature vector according to the client information and the transaction information; forming an input vector based on the risk trend vector, the comprehensive feature vector and the low-order cross feature vector, and inputting the input vector into a preset risk detection model to obtain a risk detection result; and executing a preset risk control strategy based on the risk detection result. The application can improve the accuracy of credit card transaction risk detection and the processing efficiency of credit card transaction, and improve the experience of credit card clients.
Description
Technical Field
The application relates to the technical field of transaction monitoring, in particular to the technical field of artificial intelligence, and particularly relates to a method and a device for monitoring transaction risk in credit card client loan.
Background
This section is intended to provide a background or context to the embodiments of the application that are recited in the claims. The description herein is not admitted to be prior art by inclusion in this section.
With the continuous development of technology, the transaction amount of credit card transactions is gradually increased, and due to the characteristics of credit card transactions in advance, credit card transactions may have risk actions such as false transactions, etc., which result in property loss of customers and banks. The main measures adopted in the present credit risk monitoring are to deploy expert experience rules and artificial intelligent models through an intelligent decision engine, and automatically control or manually check tasks of risk clients or transactions. However, due to the increase in credit card transactions and the complexity of the transaction, the tasks of manual outbound verification and risk disposal are also increased. Part of companies are used for reducing the outbound workload, introducing an intelligent outbound mode for risk task verification, when the intelligent outbound problem is solved, different models are usually trained for different tasks to identify various different types of risks, the current models are used for directly identifying and predicting client information and transaction information to obtain risks, the detection accuracy of the transaction risks is low, the generalization capability is weak, a large number of tasks for invalid manual outbound verification or risk treatment can be possibly caused, and further transaction efficiency is reduced and manpower resources are wasted.
Disclosure of Invention
An object of the present application is to provide a method for monitoring transaction risk in credit card client credit, which improves accuracy of credit card transaction risk detection and processing efficiency of credit card transaction, and improves experience of credit card client. Another object of the present application is to provide a distributed combined application running environment anomaly detection device. It is a further object of the application to provide a distributed system. It is a further object of the application to provide a computer device. It is a further object of the application to provide a readable medium.
In order to achieve the above object, one aspect of the present application discloses a method for monitoring risk of transaction in credit of a credit card customer, comprising:
extracting risk trend vectors of a plurality of risks corresponding to the client information, carrying out feature extraction on the transaction information through a multi-task learning network to obtain a comprehensive feature vector, and obtaining a low-order cross feature vector according to the client information and the transaction information;
forming an input vector based on the risk trend vector, the comprehensive feature vector and the low-order cross feature vector, and inputting the input vector into a preset risk detection model to obtain a risk detection result;
and executing a preset risk control strategy based on the risk detection result.
Optionally, the method further comprises:
acquiring customer basic data and transaction original data;
and respectively characterizing and extracting the customer basic data and the transaction original data to obtain customer information and transaction information.
Optionally, the extracting features of the transaction information through the multi-task learning network to obtain the comprehensive feature vector includes:
building training tasks corresponding to the risks respectively;
extracting exclusive feature vectors and shared feature vectors through an exclusive network and a shared network respectively based on the transaction information;
the integrated feature vector is formed based on the dedicated feature vector and the shared feature vector.
Optionally, the obtaining the low-order cross feature vector according to the customer information and the transaction information includes:
vector cross-multiplying the customer information and the transaction information to obtain an outer product;
and inputting the outer product into a preset linear model to obtain the low-order cross feature vector.
Optionally, the forming the input vector based on the risk tendency vector, the integrated feature vector, and the low-order cross feature vector includes:
and splicing the risk tendency vector, the comprehensive feature vector and the low-order cross feature vector to obtain an input vector.
Optionally, inputting the input vector into a preset risk detection model to obtain a risk detection result includes:
inputting the input vector into a preset risk detection model to obtain a detection result vector comprising prediction probabilities of a plurality of risks;
obtaining a weighted prediction probability corresponding to each risk according to the prediction probability of each risk and a preset weight corresponding to the risk;
and adding the weighted prediction probabilities corresponding to all risks to obtain comprehensive prediction probability serving as the risk detection result.
Optionally, the risk control policy includes performing one or more of transaction processing, risk pre-warning, transaction rejection and account stop payment, outbound call, and gray listing operations.
Optionally, the executing a preset risk control policy based on the risk detection result includes:
if the risk detection result is greater than the first preset probability, executing transaction rejection and account stop payment operation;
if the risk detection result is smaller than the first preset probability and larger than the second preset probability, executing transaction rejection operation;
if the risk detection result is smaller than the second preset probability and larger than the third preset probability, executing outbound operation;
if the risk detection result is smaller than the third preset probability and larger than the fourth preset probability, executing the operation of adding the gray list;
and if the risk detection result is smaller than the fourth preset probability, executing transaction processing operation.
Optionally, performing the outbound operation includes:
according to the client information and the transaction information, outbound is carried out, and reply information of the client based on the outbound is obtained;
inputting the reply information into a preset natural language model to obtain a customer representation vector;
and inputting the client representation vector into a preset emotion classifier to obtain the client representation, and determining whether to conduct transaction processing according to the client representation.
The application also discloses a credit card customer in-credit transaction risk monitoring device, which comprises:
the information extraction module is used for extracting risk trend vectors of a plurality of risks corresponding to the client information, carrying out feature extraction on the transaction information through the multi-task learning network to obtain a comprehensive feature vector, and obtaining a low-order cross feature vector according to the client information and the transaction information;
the risk detection module is used for forming an input vector based on the risk trend vector, the comprehensive feature vector and the low-order cross feature vector, and inputting the input vector into a preset risk detection model to obtain a risk detection result;
and the strategy executing module is used for executing a preset risk control strategy based on the risk detection result.
The application also discloses a computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the method as described above when executing the computer program.
The application also discloses a computer readable storage medium storing a computer program which when executed by a processor implements a method as described above.
The application relates to a method for monitoring transaction risk in credit card client credit, which is characterized by extracting risk trend vectors of client information corresponding to a plurality of risks, carrying out feature extraction on transaction information through a multi-task learning network to obtain a comprehensive feature vector, and obtaining a low-order cross feature vector according to the client information and the transaction information; forming an input vector based on the risk trend vector, the comprehensive feature vector and the low-order cross feature vector, and inputting the input vector into a preset risk detection model to obtain a risk detection result; and executing a preset risk control strategy based on the risk detection result. Therefore, the risk trend vector is extracted from the client information according to different risks, and the vector which can reflect the corresponding risk can be obtained, so that the accuracy of predicting each risk type is improved. In addition, the application carries out feature extraction on the transaction information through the multi-task learning network, extracts the feature vector which is unique to each risk and shared by all risks in the transaction information through the multi-task learning network to obtain the comprehensive feature vector, improves the universality of the vector extracted by the transaction information, and further improves the accuracy of risk type prediction. In addition, the application also adopts the client information and the transaction information to obtain the low-order cross feature vector, and combines the client risk and the transaction characteristic to further improve the accuracy of risk prediction. Therefore, the application can improve the accuracy of transaction risk prediction, further improve the accuracy of risk checking and controlling and the processing efficiency of transactions, improve the transaction experience of credit card clients, reduce invalid outbound or manual checking tasks and reduce the waste of manpower resources and cost.
Drawings
In order to more clearly illustrate the embodiments of the application or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art. In the drawings:
FIG. 1 is a flow chart of a method for monitoring risk of an in-flight transaction of a credit card customer according to an embodiment of the application;
FIG. 2 is a flow chart of a method S000 for monitoring risk of an in-flight transaction of a credit card customer according to an embodiment of the application;
FIG. 3 is a flowchart of a method S100 for monitoring risk of a credit card customer in-process transaction to obtain a comprehensive feature vector according to an embodiment of the present application;
FIG. 4 is a flow chart of the method S100 for monitoring risk of transaction in credit card customer credit according to an embodiment of the application to obtain low-order cross feature vectors;
FIG. 5 is a flow chart of the method S200 for monitoring risk of transaction in credit card customer credit according to an embodiment of the application;
FIG. 6 is a flowchart illustrating a method S200 for monitoring risk of an in-credit transaction of a credit card customer according to an embodiment of the application;
FIG. 7 is a flow chart of a method S300 for monitoring risk of an in-flight transaction of a credit card customer according to an embodiment of the application;
FIG. 8 is a flowchart of a method S330 for monitoring risk of an in-flight transaction of a credit card customer according to an embodiment of the application;
fig. 9 is a schematic diagram of a credit card customer in-process transaction risk monitoring device according to an embodiment of the application.
Fig. 10 shows a schematic structural diagram of a computer device suitable for use in implementing embodiments of the present application.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present application more apparent, the embodiments of the present application will be described in further detail with reference to the accompanying drawings. The exemplary embodiments of the present application and their descriptions herein are for the purpose of explaining the present application, but are not to be construed as limiting the application.
It should be noted that the method and the device for monitoring transaction risk in credit card client credit disclosed by the application can be used in the technical field of artificial intelligence and can also be used in any field except the technical field of artificial intelligence, and the application field of the method and the device for monitoring transaction risk in credit card client credit disclosed by the application is not limited.
According to one aspect of the present application, a method for monitoring risk of an in-flight transaction by a credit card customer is disclosed. As shown in fig. 1, in this embodiment, the method includes:
s100: and extracting risk trend vectors of the client information corresponding to a plurality of risks, carrying out feature extraction on the transaction information through a multi-task learning network to obtain a comprehensive feature vector, and obtaining a low-order cross feature vector according to the client information and the transaction information.
After the client information is obtained, risk trend vectors corresponding to the multiple risks respectively can be trained in advance, and the client information is extracted according to the preset risk trend vectors to obtain risk trend vectors corresponding to the client and the multiple risks respectively. Of course, in other embodiments, the risk tendency vectors corresponding to the plurality of risks may be obtained by processing the client information in other manners.
S200: and forming an input vector based on the risk trend vector, the comprehensive feature vector and the low-order cross feature vector, and inputting the input vector into a preset risk detection model to obtain a risk detection result.
S300: and executing a preset risk control strategy based on the risk detection result.
According to the risk prediction method and the risk prediction device, the risk trend vectors are respectively extracted from the client information aiming at different multiple risks, so that the vectors which can reflect corresponding risks can be obtained, and the accuracy of prediction of various risk types is improved. In addition, the application carries out feature extraction on the transaction information through the multi-task learning network, extracts the feature vector which is unique to each risk and shared by all risks in the transaction information through the multi-task learning network to obtain the comprehensive feature vector, improves the universality of the vector extracted by the transaction information, and further improves the accuracy of risk type prediction. In addition, the application also adopts the client information and the transaction information to obtain the low-order cross feature vector, and combines the client risk and the transaction characteristic to further improve the accuracy of risk prediction. Therefore, the application can improve the accuracy of transaction risk prediction, further improve the accuracy of risk checking and controlling and the processing efficiency of transactions, improve the transaction experience of credit card clients, reduce invalid outbound or manual checking tasks and reduce the waste of manpower resources and cost.
In an alternative embodiment, as shown in fig. 2, the method further comprises S000:
s010: customer base data and transaction original data are obtained.
S020: and respectively characterizing and extracting the customer basic data and the transaction original data to obtain customer information and transaction information.
In particular, it will be appreciated that customer base data of customers may be pre-stored, typically by registration or the like, and relevant data during the transaction may also be stored in the transaction system. Therefore, pre-stored customer basic data and transaction original data can be obtained, the customer basic data and the transaction original data are respectively characterized and extracted to obtain customer information and transaction information, so that the characterized and extracted customer information and transaction information can be better used for subsequent risk prediction, and the risk prediction efficiency is improved.
In a specific example, the customer base data includes customer name, gender, age, occupation, usage equipment, and income. The transaction raw data may include a card number, transaction type, transaction amount, transaction currency, transaction country, transaction city, merchant name, merchant type, credit line, and card expiration date, among other transaction details. Whether credit risk, fraud risk and compliance risk occur in the history of the client can also be obtained; and whether historical behavior information such as excess derate, stop payment, early warning and the like is adopted is used for model training.
The risk trend vector, the comprehensive feature vector and the low-order cross feature vector can be realized through a model, so that an input vector is obtained, and a risk detection result is output. The model may include, among other things, an input layer, an embedded layer, a client-transaction interaction layer, and an output layer. The input layer is used for inputting the client basic data and the transaction original data to the embedded layer. The embedded layer performs preliminary processing on the input data, so that subsequent module learning is facilitated.
The preliminary processing of the data by the embedding layer may include characterizing and extracting the customer base data and the transaction raw data, respectively, to obtain customer information and transaction information, which after passing through the embedding layer represent each attribute information of the customer and the transaction. The features of the characterizing extraction of the customer basic data and the transaction original data may be preset, and those skilled in the art may set the features of the characterizing extraction and the specific manner of the characterizing extraction according to actual requirements, which is not limited in the present application.
In an alternative embodiment, as shown in fig. 3, the step S100 of extracting features of the transaction information through the multi-task learning network to obtain the integrated feature vector includes:
s111: and building training tasks corresponding to the risks respectively.
S112: and extracting the exclusive feature vector and the shared feature vector through the exclusive network and the shared network respectively based on the transaction information.
S113: the integrated feature vector is formed based on the dedicated feature vector and the shared feature vector.
In particular, transaction integrity properties associated with multiple risk tendencies in transaction information may be sought through a multiple learning (PLE) network. First, K risks may be defined, and the PLE network corresponding thereto contains K training tasks. The first stage of each task is aimed at extracting a comprehensive characteristic vector for risk trends, which is mainly accomplished by a differentiated expert network. The expert network is mainly used for training out comprehensive characteristic vectors, receiving transaction characteristic vectors and training the comprehensive characteristic vectors through a plurality of full-connection layers.
According to the purposes, the expert network is divided into exclusive and shared, common transaction characteristics are necessarily existed in different risk types, and different personalized task networks are required to be served by training results of the shared network. Let t be s A shared network for specific task k Has t k And a proprietary network. Thus, task is being performed k During training, characteristic information related to risk tendency of clients is provided by the exclusive network and the shared network, and the exclusive characteristic vector and the shared characteristic vector are extracted through the exclusive network and the shared network to respectively form a comprehensive characteristic vector p corresponding to each risk k . Finally, the output result of the corresponding expert network is synthesized through the gate control network to obtain the comprehensive characteristic vector G of the transaction k 。
In an alternative embodiment, as shown in fig. 4, the step of S100 obtaining a low-order cross feature vector according to the customer information and the transaction information includes:
s121: and carrying out vector cross multiplication on the client information and the transaction information to obtain an outer product.
S122: and inputting the outer product into a preset linear model to obtain the low-order cross feature vector.
Specifically, it can be understood that in this alternative embodiment, the client information and the transaction information acquired by the embedding layer are associated, and are directly regarded as respective representing vectors of the client and the transaction, then vector cross multiplication is performed to obtain an outer product, and the obtained result is input into the linear model to obtain a model containing low-order interaction information of the client-transaction, so that the model can not only memorize the low-order interaction of the client-transaction, but also learn the interaction condition of the low-order cross feature, risk tendency and transaction information of the client-transaction, and further improve the prediction effect of the model.
It should be noted that the linear model may be implemented by conventional technical means in the art, and will not be described herein.
In an alternative embodiment, as shown in fig. 5, the forming an input vector by S200 based on the risk tendency vector, the integrated feature vector, and the low-order cross feature vector includes:
s210: and splicing the risk tendency vector, the comprehensive feature vector and the low-order cross feature vector to obtain an input vector.
Specifically, in this optional embodiment, the risk trend vector, the comprehensive feature vector and the low-order cross feature vector may be obtained through a client-transaction interaction layer, and the obtained risk trend vector, the comprehensive feature vector and the low-order cross feature vector are spliced to obtain an input vector, so as to expand the range of the input vector, enable the input vector to have information required for accurately predicting each risk, and improve the accuracy of a risk detection result obtained by inputting the input vector into a preset risk detection model. Therefore, the input vector comprises a high-order interaction part and low-order interaction part information, the high-order interaction part mainly learns risk related information such as client information, transaction information and the like, extracts the characteristics related to risk identification and risk disposal, and mainly trains a user and a transaction embedded vector by using different deep learning models. The low-level interaction part is a collaborative filtering module for memorizing interactions of low-level cross-characteristics of the client-transaction. Conventional mainstream model algorithms often only focus on first-order features of customers and transactions, and require complex manual feature engineering if cross features are considered. Therefore, the method introduces a deep learning algorithm model to mine highly complex feature interaction, improves the accuracy of model prediction, but the deep learning model is easy to cause the problems of over fitting and the like. Therefore, the application adopts the traditional wind control model algorithm to learn the low-order interaction between simple features and then works together with the neural network to assist the deep learning network to be insufficient, thereby achieving better risk identification and treatment effects.
It should be noted that, the risk detection model may be implemented by using technologies such as machine learning or neural network, and a large amount of historical customer information, transaction information and customer behavior information may be trained in advance based on the machine learning or neural network technology to obtain the risk detection model, which is common knowledge in the art and will not be described herein.
In an alternative embodiment, as shown in fig. 6, the step S200 of inputting the input vector into a preset risk detection model to obtain a risk detection result includes:
s221: and inputting the input vector into a preset risk detection model to obtain a detection result vector comprising prediction probabilities of a plurality of risks.
S222: and obtaining the weighted prediction probability corresponding to each risk according to the prediction probability of each risk and the preset weight corresponding to the risk.
S223: and adding the weighted prediction probabilities corresponding to all risks to obtain comprehensive prediction probability serving as the risk detection result.
Specifically, the risk detection model outputs a detection result vector formed by a plurality of prediction probabilities of risks respectively, and the prediction probabilities can be added to obtain a risk detection result of the current transaction of the client. However, the importance of different risks is different, the method further sets corresponding weights for different risks, weights the prediction probability of the risks through the preset weights corresponding to each risk to obtain weighted prediction probability, and then adds the weighted prediction probabilities corresponding to all the risks to obtain comprehensive prediction probability as a risk detection result.
In alternative embodiments, the risk control policy includes performing one or more of transaction processing, risk early warning, transaction rejection and account stop payment, outbound call, and gray listing operations.
Specifically, it can be understood that the present application can adopt different processing strategies according to the probability value of the obtained risk detection result, for example, the transaction with lower risk probability is directly judged to pass, and transaction processing is performed; for transactions with slightly higher risks, risk early warning is carried out on management staff, so that the management staff can find potential risks in time and process the potential risks; for transactions with higher risk, operations such as directly executing transaction rejection, transaction rejection and account stop payment, outbound call, gray list adding and the like can be sequentially considered. According to the application, different operations are executed on transactions with different risks through a hierarchical progressive strategy, so that the safety and the transaction efficiency are considered, and the user experience is improved.
In an alternative embodiment, as shown in fig. 7, the step S300 of executing a preset risk control policy based on the risk detection result includes:
s310: and if the risk detection result is greater than the first preset probability, executing transaction rejection and account stop payment operation.
S320: and if the risk detection result is smaller than the first preset probability and larger than the second preset probability, executing transaction rejection operation.
S330: and if the risk detection result is smaller than the second preset probability and larger than the third preset probability, executing outbound operation.
S340: and if the risk detection result is smaller than the third preset probability and larger than the fourth preset probability, executing the operation of adding the gray list.
S350: and if the risk detection result is smaller than the fourth preset probability, executing transaction processing operation.
Specifically, it can be understood that the first preset probability, the second preset probability, the third preset probability and the fourth preset probability can be preset, so that the first preset probability, the second preset probability, the third preset probability and the fourth preset probability are sequentially reduced, and corresponding strategy operations are executed for different risk detection results. In a specific example, the first preset probability, the second preset probability, the third preset probability and the fourth preset probability may be selected to be 0.98, 0.97, 0.96 and 0.95, respectively.
The specific probability values of the first preset probability, the second preset probability, the third preset probability and the fourth preset probability can be set by a person skilled in the art according to actual situations, and the application is not limited to this.
In an alternative embodiment, as shown in fig. 8, S330 performs the outbound operation including:
s331: and carrying out outbound according to the client information and the transaction information, and acquiring reply information of the client based on the outbound.
S332: and inputting the reply information into a preset natural language model to obtain a customer representation vector.
S333: and inputting the client representation vector into a preset emotion classifier to obtain the client representation, and determining whether to conduct transaction processing according to the client representation.
Specifically, during the task of intelligent outbound, the client, transaction and risk information can be used as parameters to be sent to an intelligent outbound module for outbound verification, and the outbound module can be obtained through conventional technology in the field and is not described herein. Meanwhile, the application adopts emotion recognition technology in intelligent outbound, so that whether fraud risk or credit risk exists can be judged according to emotion tendency answered by a client. Whereas conventional NLP techniques typically only focus on semantic information of text, ignoring the impact of affective information.
In a specific example, the outbound procedure may be implemented by a BERT model. The BERT model is based on a transducer model, belonging to the encoder therein. Based on the existing corpus model of the large-scale outbound scenario for solving the problem of daily customer consultation, a new BERT model is initialized instead of using random parameters for initialization. In practical application, the parameters of the new model are then fine-tuned according to the task of the credit card wind-controlled manual outbound team. The way of generating the new task mainly comprises 5 steps,
1) The text data for manual outbound verification is prepared and can be set according to actual requirements.
2) Converting the original text into a BERT compatible input format;
3) Adding a new Layer above BERT to generate a downstream task model;
4) Training a downstream task model;
5) And deducing a new sample.
In the practical application process, preprocessing is carried out according to text information answered by a client, including word segmentation, stop word removal, part-of-speech tagging and the like, and the processed text is input into a BERT pre-training model to obtain a representation vector of the text.
And inputting the expression vector into an emotion classifier for emotion classification.
Wherein y is the output of the emotion classifier, h is the representation vector of the BERT model, W out And b out Is a parameter of the emotion classifier. Post-processing is carried out on the output of the emotion classifier, including normalization, normalization and other operations on the output so as to obtain a final emotion classification result. And classifying the customer answers into three types of positive, negative and neutral according to the emotion classification result, and outputting an emotion analysis result. And according to the emotion analysis result, carrying out risk judgment and decision by combining the business scene.
Wherein, the positive class: the customers clearly indicate that the consumption belongs to the principal consumption, and the transaction is released if the risk is not found; negative classes: the customer indicates non-self transaction, and then account locking is carried out; neutral class: the customer can not confirm the consumption condition or the transaction condition, or hang up the phone, etc., then send a short message, generate a manual task, and manually check by the business.
Therefore, the emotion recognition technology of the intelligent outbound can judge the true intention and emotion tendency of the client more accurately, thereby improving the accuracy and effect of wind control. The risk conclusion verified by the outbound can be fed back to the risk early warning module, and compared and analyzed with the early warning risk type, so that the follow-up early warning network can be optimized.
The application provides an automatic transaction risk checking and disposing scheme in credit card credit based on multi-task learning, which adopts a multi-task learning method to realize the recognition and disposing of different risk types, thereby reducing the number of models and the consumption of calculation resources and improving the precision and generalization capability of the models by sharing model parameters. In the input stage, the model provides a plurality of risk vectors for risk transaction, simultaneously extracts a plurality of risk characteristic vectors for clients, and uses one risk characteristic to correspond to one client risk trend, thereby overcoming the disadvantage that simple client information and transaction information are spliced as an input set and improving the effectiveness of input data. Meanwhile, creatively proposes to use a multi-task learning model to combine the characteristics of the transaction and the risk tendency of the clients and calculate the risk probability of the transaction. The application can solve the problems of poor multitasking capability, weak risk fine recognition capability, lag risk disposal, relatively low accuracy and other service pain points of the credit card wind control intelligent model. And the risk checking and controlling capability in credit card lending is comprehensively improved, the workload of risk monitoring personnel is reduced, and the working efficiency is improved.
Based on the same principle, the application also discloses a credit card customer in-credit transaction risk monitoring device. As shown in fig. 9, in this embodiment, the apparatus includes an information extraction module 11, a risk detection module 12, and a policy enforcement module 13.
The information extraction module 11 is configured to extract risk trend vectors corresponding to multiple risks of the client information, perform feature extraction on the transaction information through the multi-task learning network to obtain a comprehensive feature vector, and obtain a low-order cross feature vector according to the client information and the transaction information.
The risk detection module 12 is configured to form an input vector based on the risk trend vector, the comprehensive feature vector and the low-order cross feature vector, and input the input vector into a preset risk detection model to obtain a risk detection result.
The policy execution module 13 is configured to execute a preset risk control policy based on the risk detection result.
Since the principle of the device for solving the problem is similar to that of the above method, the implementation of the device can be referred to the implementation of the method, and will not be described herein.
The embodiment of the application also provides computer equipment, which comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor realizes the method when executing the computer program.
Embodiments of the present application also provide a computer-readable storage medium storing a computer program which, when executed by a processor, implements the above-described method.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program for producing a system, apparatus, module, or unit as set forth in the above embodiments, and may be embodied in a computer chip or entity, or in an article of manufacture having some function. A typical implementation device is a computer device, which may be, for example, a personal computer, a laptop computer, a cellular telephone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.
In a typical example, the computer apparatus includes a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the program to implement a method performed by a client as described above, or where the processor executes the program to implement a method performed by a server as described above.
Referring now to FIG. 10, there is illustrated a schematic diagram of a computer device 600 suitable for use in implementing embodiments of the present application.
As shown in fig. 10, the computer apparatus 600 includes a Central Processing Unit (CPU) 601, which can perform various appropriate works and processes according to a program stored in a Read Only Memory (ROM) 602 or a program loaded from a storage section 608 into a Random Access Memory (RAM) 603. In the RAM603, various programs and data required for the operation of the computer device 600 are also stored. The CPU601, ROM602, and RAM603 are connected to each other through a bus 604. An input/output (I/O) interface 605 is also connected to bus 604.
The following components are connected to the I/O interface 605: an input portion 606 including a keyboard, mouse, etc.; an output portion 607 including a Cathode Ray Tube (CRT), a liquid crystal feedback device (LCD), and the like, and a speaker, and the like; a storage section 608 including a hard disk and the like; and a communication section 609 including a network interface card such as a LAN card, a modem, or the like. The communication section 609 performs communication processing via a network such as the internet. The drive 610 is also connected to the I/O interface 605 as needed. Removable media 611 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is mounted on drive 610 as needed, so that a computer program read therefrom is mounted as needed as storage section 608.
In particular, according to embodiments of the present application, the processes described above with reference to flowcharts may be implemented as computer software programs. For example, embodiments of the present application include a computer program product comprising a computer program tangibly embodied on a machine-readable medium, the computer program comprising program code for performing the method shown in the flowchart. In such an embodiment, the computer program may be downloaded and installed from a network through the communication portion 609, and/or installed from the removable medium 611.
Computer readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, does not include transitory computer-readable media (transmission media), such as modulated data signals and carrier waves.
For convenience of description, the above devices are described as being functionally divided into various units, respectively. Of course, the functions of each element may be implemented in the same piece or pieces of software and/or hardware when implementing the present application.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It should also be noted that 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 phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises the element.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments. In particular, for system embodiments, since they are substantially similar to method embodiments, the description is relatively simple, as relevant to see a section of the description of method embodiments.
The foregoing is merely exemplary of the present application and is not intended to limit the present application. Various modifications and variations of the present application will be apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. which come within the spirit and principles of the application are to be included in the scope of the claims of the present application.
Claims (12)
1. A method for monitoring risk of transactions in credit card customers, comprising:
extracting risk trend vectors of a plurality of risks corresponding to the client information, carrying out feature extraction on the transaction information through a multi-task learning network to obtain a comprehensive feature vector, and obtaining a low-order cross feature vector according to the client information and the transaction information;
forming an input vector based on the risk trend vector, the comprehensive feature vector and the low-order cross feature vector, and inputting the input vector into a preset risk detection model to obtain a risk detection result;
and executing a preset risk control strategy based on the risk detection result.
2. The method for monitoring risk of an on-credit transaction for a credit card customer according to claim 1, further comprising:
acquiring customer basic data and transaction original data;
and respectively characterizing and extracting the customer basic data and the transaction original data to obtain customer information and transaction information.
3. The method for monitoring risk of in-credit transaction of a credit card customer according to claim 1, wherein the feature extraction of transaction information through a multi-task learning network to obtain a comprehensive feature vector comprises:
building training tasks corresponding to the risks respectively;
extracting exclusive feature vectors and shared feature vectors through an exclusive network and a shared network respectively based on the transaction information;
the integrated feature vector is formed based on the dedicated feature vector and the shared feature vector.
4. The method of claim 1, wherein said deriving a low-level cross feature vector from said customer information and said transaction information comprises:
vector cross-multiplying the customer information and the transaction information to obtain an outer product;
and inputting the outer product into a preset linear model to obtain the low-order cross feature vector.
5. The method of claim 1, wherein the forming an input vector based on the risk propensity vector, the composite feature vector, and the low order cross feature vector comprises:
and splicing the risk tendency vector, the comprehensive feature vector and the low-order cross feature vector to obtain an input vector.
6. The method for monitoring risk of an on-credit transaction of a credit card customer according to claim 1, wherein inputting the input vector into a predetermined risk detection model to obtain a risk detection result comprises:
inputting the input vector into a preset risk detection model to obtain a detection result vector comprising prediction probabilities of a plurality of risks;
obtaining a weighted prediction probability corresponding to each risk according to the prediction probability of each risk and a preset weight corresponding to the risk;
and adding the weighted prediction probabilities corresponding to all risks to obtain comprehensive prediction probability serving as the risk detection result.
7. The method of claim 1, wherein the risk control policy includes performing one or more of transaction processing, risk pre-warning, transaction rejection and account stop payment, outbound and gray listing.
8. The method of claim 1, wherein the performing a predetermined risk control strategy based on the risk detection result comprises:
if the risk detection result is greater than the first preset probability, executing transaction rejection and account stop payment operation;
if the risk detection result is smaller than the first preset probability and larger than the second preset probability, executing transaction rejection operation;
if the risk detection result is smaller than the second preset probability and larger than the third preset probability, executing outbound operation;
if the risk detection result is smaller than the third preset probability and larger than the fourth preset probability, executing the operation of adding the gray list;
and if the risk detection result is smaller than the fourth preset probability, executing transaction processing operation.
9. The method of claim 1, wherein performing outbound operations comprises:
according to the client information and the transaction information, outbound is carried out, and reply information of the client based on the outbound is obtained;
inputting the reply information into a preset natural language model to obtain a customer representation vector;
and inputting the client representation vector into a preset emotion classifier to obtain the client representation, and determining whether to conduct transaction processing according to the client representation.
10. A credit card customer in-credit transaction risk monitoring apparatus comprising:
the information extraction module is used for extracting risk trend vectors of a plurality of risks corresponding to the client information, carrying out feature extraction on the transaction information through the multi-task learning network to obtain a comprehensive feature vector, and obtaining a low-order cross feature vector according to the client information and the transaction information;
the risk detection module is used for forming an input vector based on the risk trend vector, the comprehensive feature vector and the low-order cross feature vector, and inputting the input vector into a preset risk detection model to obtain a risk detection result;
and the strategy executing module is used for executing a preset risk control strategy based on the risk detection result.
11. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the method of any of claims 1 to 9 when executing the computer program.
12. A computer readable storage medium, characterized in that the computer readable storage medium stores a computer program which, when executed by a processor, implements the method of any of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310805182.9A CN116757835A (en) | 2023-06-30 | 2023-06-30 | Method and device for monitoring transaction risk in credit card customer credit |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310805182.9A CN116757835A (en) | 2023-06-30 | 2023-06-30 | Method and device for monitoring transaction risk in credit card customer credit |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116757835A true CN116757835A (en) | 2023-09-15 |
Family
ID=87956884
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310805182.9A Pending CN116757835A (en) | 2023-06-30 | 2023-06-30 | Method and device for monitoring transaction risk in credit card customer credit |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116757835A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117035434A (en) * | 2023-10-10 | 2023-11-10 | 中国建设银行股份有限公司 | Suspicious transaction monitoring method and suspicious transaction monitoring device |
-
2023
- 2023-06-30 CN CN202310805182.9A patent/CN116757835A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117035434A (en) * | 2023-10-10 | 2023-11-10 | 中国建设银行股份有限公司 | Suspicious transaction monitoring method and suspicious transaction monitoring device |
| CN117035434B (en) * | 2023-10-10 | 2023-12-29 | 中国建设银行股份有限公司 | Suspicious transaction monitoring method and suspicious transaction monitoring device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Chen et al. | Stock prediction using convolutional neural network | |
| CN110070391B (en) | Data processing method and device, computer readable medium and electronic equipment | |
| CN110751557B (en) | Abnormal fund transaction behavior analysis method and system based on sequence model | |
| TW201944305A (en) | Method and apparatus for determining risk probability of service request event | |
| CN111340616B (en) | Method, device, equipment and medium for approving online loan | |
| CN112862298B (en) | Credit evaluation method for user portrait | |
| CN111932020B (en) | Customer churn prediction method and device | |
| CN110032551B (en) | Model generation system, method and prediction system | |
| CN116757835A (en) | Method and device for monitoring transaction risk in credit card customer credit | |
| Huynh et al. | Addressing regulatory requirements on explanations for automated decisions with provenance—A case study | |
| CN108647200A (en) | Talk with intent classifier method and device, equipment and storage medium | |
| Lee | Predicting land prices and measuring uncertainty by combining supervised and unsupervised learning | |
| Laylo | The Impact of AI and Information Technologies on Islamic Charity (Zakat): Modern Solutions for Efficient Distribution | |
| CN113159419A (en) | Group feature portrait analysis method, device and equipment and readable storage medium | |
| CN111353728A (en) | Risk analysis method and system | |
| CN117437020A (en) | Merchant risk judging method and device, electronic equipment and medium | |
| CN117171403A (en) | Data processing method, device, computer equipment and storage medium | |
| CN117076672A (en) | Training method of text classification model, text classification method and device | |
| Bai et al. | Commercial Bank Credit Grading Model Using Genetic Optimization Neural Network and Cluster Analysis | |
| CN113256404A (en) | Data processing method and device | |
| Bakurova et al. | Development of a Productive Credit Decision-Making System Based on the Ontology Model. | |
| CN117556264B (en) | Training method and device for evaluation model and electronic equipment | |
| Ning et al. | A Cost-Sensitive Ensemble Model for e-Commerce Customer Behavior Prediction with Weighted SVM | |
| CN110059905B (en) | Risk quantification method, risk identification system and storage medium | |
| Schmitt | Explainable Automated Machine Learning for Credit Decisions: Enhancing Human Artificial Intelligence Collaboration in Financial Engineering |
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 |