CN114819975A - Transaction risk identification method and device - Google Patents
Transaction risk identification method and deviceDownload PDFInfo
- Publication number
- CN114819975A CN114819975ACN202210466159.7ACN202210466159ACN114819975ACN 114819975 ACN114819975 ACN 114819975ACN 202210466159 ACN202210466159 ACN 202210466159ACN 114819975 ACN114819975 ACN 114819975A
- Authority
- CN
- China
- Prior art keywords
- transaction
- risk
- business transaction
- customer
- prediction model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Computer Security & Cryptography (AREA)
- Finance (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Translated fromChinese
Description
Translated fromChinese技术领域technical field
本申请涉及数据处理技术领域，尤其涉及一种交易风险识别方法和装置。The present application relates to the technical field of data processing, and in particular, to a transaction risk identification method and device.
背景技术Background technique
在反洗钱的各工作环节中，有效检测出存在洗钱风险的可疑交易行为是较为重要的一个环节。In each work link of anti-money laundering, the effective detection of suspicious transactions with money laundering risks is a more important link.
基于此，银行等金融系统，为了能够及时发现存在洗钱嫌疑的业务交易行为，就需要识别业务交易是否存在洗钱风险。然而，目前识别业务交易的洗钱风险的识别准确度较低，导致识别结果较高的误报率。Based on this, financial systems such as banks need to identify whether there is a money laundering risk in business transactions in order to timely detect business transactions suspected of money laundering. However, the current identification accuracy for identifying money laundering risks in business transactions is low, resulting in a high false positive rate in identification results.
发明内容SUMMARY OF THE INVENTION
本申请提供了一种交易风险识别方法和装置，以提高识别存在洗钱风险的业务交易的准确性。The present application provides a transaction risk identification method and device to improve the accuracy of identifying business transactions with money laundering risks.
为了实现如上目的，一方面本申请提供了一种交易风险识别方法，包括：In order to achieve the above purpose, on the one hand, the present application provides a transaction risk identification method, including:
获得业务交易的交易特征信息、所述业务交易涉及到的客户的客户信息以及所述客户关联的历史业务交易的交易特征信息；Obtain transaction feature information of the business transaction, customer information of the customer involved in the business transaction, and transaction feature information of the historical business transaction associated with the customer;
基于所述业务交易的交易特征，利用第一预测模型确定所述业务交易存在洗钱风险的第一风险概率，所述第一预测模型为利用标注有洗钱风险标签的多个第一业务交易样本的交易特征信息训练得到的；Based on the transaction characteristics of the business transaction, a first prediction model is used to determine the first risk probability of money laundering risk in the business transaction. It is obtained by training the transaction feature information;
基于业务交易涉及到的客户信息，利用第二预测模型确定所述业务交易存在洗钱风险的第二风险概率，所述第二预测模型为利用标注有洗钱风险标签的多个第二业务交易样本涉及到的客户样本的客户信息训练得到的；Based on the customer information involved in the business transaction, a second prediction model is used to determine the second risk probability of the money laundering risk in the business transaction, and the second prediction model is to use a plurality of second business transaction samples marked with money laundering risk labels. The customer information of the received customer samples is obtained by training;
基于所述历史业务交易的交易特征信息，利用第三预测模型确定所述业务交易存在洗钱风险的第三风险概率，所述第三预测模型为利用多个第三业务交易样本涉及到的客户样本关联的历史业务交易的交易特征信息训练得到的，所述第三业务交易样本标注有洗钱风险标签；Based on the transaction feature information of the historical business transaction, a third prediction model is used to determine a third risk probability of money laundering risk in the business transaction, and the third prediction model is a customer sample involved by using a plurality of third business transaction samples The third business transaction sample is marked with a money laundering risk label, obtained by training the transaction feature information of the associated historical business transaction;
基于所述第一风险概率、第二风险概率和第三风险概率，确定所述业务交易的洗钱风险识别结果。Based on the first risk probability, the second risk probability and the third risk probability, a money laundering risk identification result of the business transaction is determined.
在一种可能的实现方式中，所述基于所述第一风险概率、第二风险概率和第三风险概率，确定所述业务交易的洗钱风险识别结果，包括：In a possible implementation manner, the determining the money laundering risk identification result of the business transaction based on the first risk probability, the second risk probability and the third risk probability includes:
确定所述第一风险概率对应的第一权重、所述第二风险概率对应的第二权重以及所述第三风险概率对应的第三权重；determining a first weight corresponding to the first risk probability, a second weight corresponding to the second risk probability, and a third weight corresponding to the third risk probability;
基于所述第一风险概率的第一权重，第二风险概率的第二权重以及第三风险概率的第三权重，确定所述第一风险概率、第二风险概率和第三风险概率的综合风险概率，将所述综合风险概率确定为所述业务交易存在洗钱风险的概率。Based on the first weight of the first risk probability, the second weight of the second risk probability and the third weight of the third risk probability, determine the comprehensive risk of the first risk probability, the second risk probability and the third risk probability probability, and the comprehensive risk probability is determined as the probability that the business transaction has money laundering risk.
在又一种可能的实现方式中，所述第一权重、第二权重和第三权重为基于所述第一预测模型的第一混淆矩阵、所述第二预测模型的第二混淆矩阵以及第三预测模型的第三混淆矩阵确定的。In yet another possible implementation manner, the first weight, the second weight and the third weight are a first confusion matrix based on the first prediction model, a second confusion matrix based on the second prediction model, and a third The third confusion matrix of the three prediction models was determined.
在又一种可能的实现方式中，所述获得业务交易的交易特征信息、所述业务交易涉及到的客户的客户信息以及所述客户关联的历史业务交易的交易特征信息，包括：In another possible implementation manner, the obtaining transaction feature information of the business transaction, the customer information of the customer involved in the business transaction, and the transaction feature information of the historical business transaction associated with the customer include:
获得业务交易的交易特征信息以及所述业务交易涉及到的客户的客户信息，所述客户的客户信息至少包括：发起所述业务交易的发起客户的客户信息；Obtain the transaction feature information of the business transaction and the client information of the client involved in the business transaction, where the client information at least includes: the client information of the initiating client who initiated the business transaction;
基于所述发起客户的客户信息，获得所述发起客户关联的历史业务交易的交易特征信息。Based on the customer information of the initiating customer, transaction feature information of the historical business transaction associated with the initiating customer is obtained.
在又一种可能的实现方式中，在所述获得业务交易的交易特征信息、所述业务交易涉及到的客户的客户信息以及所述客户关联的历史业务交易的交易特征信息之后，还包括：In another possible implementation manner, after obtaining the transaction feature information of the business transaction, the customer information of the customer involved in the business transaction, and the transaction feature information of the historical business transaction associated with the customer, the method further includes:
确定所述业务交易的交易特征信息对应的第一特征向量；determining the first feature vector corresponding to the transaction feature information of the business transaction;
确定所述业务交易涉及到的客户的客户信息对应的第二特征向量；determining the second feature vector corresponding to the customer information of the customer involved in the business transaction;
确定所述历史业务交易的交易特征信息对应的第三特征向量；determining a third feature vector corresponding to the transaction feature information of the historical business transaction;
所述基于所述业务交易的交易特征，利用第一预测模型确定所述业务交易存在洗钱风险的第一风险概率，包括：The first risk probability that the business transaction has money laundering risk is determined by using the first prediction model based on the transaction characteristics of the business transaction, including:
将所述第一特征向量输入第一预测模型，得到所述第一预测模型输出的所述业务交易存在洗钱风险的第一风险概率；Inputting the first feature vector into a first prediction model to obtain a first risk probability that the business transaction output by the first prediction model has a money laundering risk;
所述基于业务交易涉及到的客户信息，利用第二预测模型确定所述业务交易存在洗钱风险的第二风险概率，包括：The second risk probability that the business transaction has money laundering risk is determined by using the second prediction model based on the customer information involved in the business transaction, including:
将所述第二特征向量输入到所述第二预测模型，得到所述第二预测模型输出的所述业务交易存在洗钱风险的第二风险概率；Inputting the second feature vector into the second prediction model to obtain a second risk probability that the business transaction output by the second prediction model has money laundering risk;
所述基于所述历史业务交易的交易特征信息，利用第三预测模型确定所述业务交易存在洗钱风险的第三风险概率，包括：The third risk probability that the business transaction has money laundering risk is determined by using a third prediction model based on the transaction feature information of the historical business transaction, including:
将所述第三特征向量输入到第三预测模型，得到所述第三预测模型输出的所述业务交易存在洗钱风险的第三风险概率。The third feature vector is input into a third prediction model to obtain a third risk probability of money laundering risk in the business transaction output by the third prediction model.
又一方面，本申请还提供了一种交易风险识别装置，包括：In another aspect, the present application also provides a transaction risk identification device, comprising:
信息获得单元，用于获得业务交易的交易特征信息、所述业务交易涉及到的客户的客户信息以及所述客户关联的历史业务交易的交易特征信息；an information obtaining unit, configured to obtain transaction feature information of a business transaction, customer information of customers involved in the business transaction, and transaction feature information of historical business transactions associated with the customer;
第一预测单元，用于基于所述业务交易的交易特征，利用第一预测模型确定所述业务交易存在洗钱风险的第一风险概率，所述第一预测模型为利用标注有洗钱风险标签的多个第一业务交易样本的交易特征信息训练得到的；The first prediction unit is configured to determine, based on the transaction characteristics of the business transaction, a first risk probability that the business transaction has a money laundering risk by using a first prediction model, where the first prediction model is to use a multi-data tag marked with a money laundering risk label. It is obtained by training the transaction feature information of the first business transaction sample;
第二预测单元，用于基于业务交易涉及到的客户信息，利用第二预测模型确定所述业务交易存在洗钱风险的第二风险概率，所述第二预测模型为利用标注有洗钱风险标签的多个第二业务交易样本涉及到的客户样本的客户信息训练得到的；The second prediction unit is configured to use a second prediction model to determine the second risk probability of money laundering risk in the business transaction based on the customer information involved in the business transaction. The customer information of the customer sample involved in the second business transaction sample is obtained by training;
第三预测单元，用于基于所述历史业务交易的交易特征信息，利用第三预测模型确定所述业务交易存在洗钱风险的第三风险概率，所述第三预测模型为利用多个第三业务交易样本涉及到的客户样本关联的历史业务交易的交易特征信息训练得到的，所述第三业务交易样本标注有洗钱风险标签；A third prediction unit, configured to use a third prediction model to determine a third risk probability of money laundering risk in the business transaction based on the transaction feature information of the historical business transaction, and the third prediction model is to use a plurality of third business transactions The third business transaction sample is marked with a money laundering risk label, obtained by training the transaction feature information of the historical business transaction associated with the customer sample involved in the transaction sample;
风险确定单元，用于基于所述第一风险概率、第二风险概率和第三风险概率，确定所述业务交易的洗钱风险识别结果。A risk determination unit, configured to determine a money laundering risk identification result of the business transaction based on the first risk probability, the second risk probability and the third risk probability.
在又一种可能的实现方式中，所述风险确定单元，包括：In yet another possible implementation, the risk determination unit includes:
权重确定单元，用于确定所述第一风险概率对应的第一权重、所述第二风险概率对应的第二权重以及所述第三风险概率对应的第三权重；a weight determination unit, configured to determine a first weight corresponding to the first risk probability, a second weight corresponding to the second risk probability, and a third weight corresponding to the third risk probability;
风险概率确定单元，用于基于所述第一风险概率的第一权重，第二风险概率的第二权重以及第三风险概率的第三权重，确定所述第一风险概率、第二风险概率和第三风险概率的综合风险概率，将所述综合风险概率确定为所述业务交易存在洗钱风险的概率。a risk probability determination unit, configured to determine the first risk probability, the second risk probability and the The comprehensive risk probability of the third risk probability is determined as the probability that the business transaction has money laundering risk.
在又一种可能的实现方式中，所述权重确定单元中的第一权重、第二权重和第三权重为基于所述第一预测模型的第一混淆矩阵、所述第二预测模型的第二混淆矩阵以及第三预测模型的第三混淆矩阵确定的。In yet another possible implementation manner, the first weight, the second weight and the third weight in the weight determination unit are a first confusion matrix based on the first prediction model, a first confusion matrix based on the second prediction model The second confusion matrix and the third confusion matrix of the third prediction model are determined.
在又一种可能的实现方式中，所述信息获得单元，包括：In yet another possible implementation, the information obtaining unit includes:
第一信息获得单元，用于获得业务交易的交易特征信息以及所述业务交易涉及到的客户的客户信息，所述客户的客户信息至少包括：发起所述业务交易的发起客户的客户信息；a first information obtaining unit, configured to obtain transaction feature information of a business transaction and client information of a client involved in the business transaction, where the client information at least includes: client information of the initiating client who initiated the business transaction;
第二信息获得单元，用于基于所述发起客户的客户信息，获得所述发起客户关联的历史业务交易的交易特征信息。The second information obtaining unit is configured to obtain transaction feature information of historical business transactions associated with the initiating customer based on the customer information of the initiating customer.
在又一种可能的实现方式中，还包括：In yet another possible implementation, it also includes:
第一向量确定单元，用于在所述获得业务交易的交易特征信息、所述业务交易涉及到的客户的客户信息以及所述客户关联的历史业务交易的交易特征信息之后，确定所述业务交易的交易特征信息对应的第一特征向量；A first vector determination unit, configured to determine the business transaction after obtaining the transaction characteristic information of the business transaction, the customer information of the customer involved in the business transaction, and the transaction characteristic information of the historical business transaction associated with the customer The first feature vector corresponding to the transaction feature information of ;
第二向量确定单元，用于确定所述业务交易涉及到的客户的客户信息对应的第二特征向量；A second vector determination unit, configured to determine a second feature vector corresponding to the customer information of the customer involved in the business transaction;
第三向量确定单元，用于确定所述历史业务交易的交易特征信息对应的第三特征向量；A third vector determination unit, configured to determine a third feature vector corresponding to the transaction feature information of the historical business transaction;
所述第一预测单元，包括：The first prediction unit includes:
第一预测子单元，用于将所述第一特征向量输入第一预测模型，得到所述第一预测模型输出的所述业务交易存在洗钱风险的第一风险概率；a first prediction subunit, configured to input the first feature vector into a first prediction model to obtain a first risk probability of money laundering risk in the business transaction output by the first prediction model;
所述第二预测单元，包括：The second prediction unit includes:
第二预测子单元，用于将所述第二特征向量输入到所述第二预测模型，得到所述第二预测模型输出的所述业务交易存在洗钱风险的第二风险概率；a second prediction subunit, configured to input the second feature vector into the second prediction model, to obtain a second risk probability of money laundering risk in the business transaction output by the second prediction model;
所述第三预测单元，包括：The third prediction unit includes:
第三预测子单元，用于将所述第三特征向量输入到第三预测模型，得到所述第三预测模型输出的所述业务交易存在洗钱风险的第三风险概率。The third prediction subunit is configured to input the third feature vector into a third prediction model, and obtain a third risk probability of money laundering risk in the business transaction output by the third prediction model.
由以上可知，在本申请实施例中，分别从业务交易的交易特征信息、涉及到的客户的客户信息以及客户关联的历史业务交易的交易特征信息三个维度分析该业务交易存在洗钱风险的风险情况，并最终结合这三个维度的分析结果综合识别业务交易存在洗钱行为的风险情况，从而可以更为全面分析业务交易可能存在洗钱风险的可能性，可以提升识别存在洗钱风险的业务交易的准确性和可靠性。As can be seen from the above, in the embodiment of the present application, the risk of money laundering risk in the business transaction is analyzed from three dimensions: the transaction feature information of the business transaction, the customer information of the involved customer, and the transaction feature information of the historical business transaction associated with the customer. Finally, combined with the analysis results of these three dimensions, we can comprehensively identify the risk of money laundering in business transactions, so that we can more comprehensively analyze the possibility of money laundering risks in business transactions, and improve the accuracy of identifying business transactions with money laundering risks. sturdiness and reliability.
附图说明Description of drawings
为了更清楚地说明本申请实施例中的技术方案，下面将对实施例描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据提供的附图获得其他的附图。In order to explain the technical solutions in the embodiments of the present application more clearly, the following briefly introduces the drawings that are used in the description of the embodiments. Obviously, the drawings in the following description are only the embodiments of the present application. For those of ordinary skill in the art, other drawings can also be obtained according to the provided drawings without any creative effort.
图1示出了本申请实施例提供的交易风险识别方法一种流程示意图；FIG. 1 shows a schematic flowchart of a transaction risk identification method provided by an embodiment of the present application;
图2示出了本申请实施例提供的交易风险识别方法的又一种流程示意图；FIG. 2 shows another schematic flowchart of the transaction risk identification method provided by the embodiment of the present application;
图3示出了本申请实施例提供的交易风险识别装置的一种组成结构示意图。FIG. 3 shows a schematic structural diagram of a transaction risk identification device provided by an embodiment of the present application.
具体实施方式Detailed ways
本申请实施例的方案可以适用于银行业等涉及到金融业务的场景，通过本申请的方案可以较为准确和可靠的识别出存在洗钱风险的交易业务。The solutions of the embodiments of the present application can be applied to scenarios involving financial services such as the banking industry, and transaction services with money laundering risks can be more accurately and reliably identified through the solutions of the present application.
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本申请一部分实施例，而不是全部的实施例。基于本申请中的实施例，本领域普通技术人员在没有付出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present application.
如图1所示，其示出了本申请实施例提供的交易风险识别方法的一种流程示意图，本实施例的方法可以包括：As shown in FIG. 1, which shows a schematic flowchart of a transaction risk identification method provided by an embodiment of the present application, the method of this embodiment may include:
S101，获得业务交易的交易特征信息、业务交易涉及到的客户的客户信息以及客户关联的历史业务交易的交易特征信息。S101: Obtain transaction feature information of a business transaction, customer information of a customer involved in the business transaction, and transaction feature information of historical business transactions associated with the customer.
其中，业务交易可以是涉及到金融交易行为的任意类型的业务交易。如，业务交易可以为转账、存款以及支付等业务交易。The business transaction may be any type of business transaction involving financial transaction behavior. For example, business transactions may be business transactions such as transfer, deposit, and payment.
业务交易的交易特征信息可以是反映该业务交易的交易内容以及交易方式等特征的信息。业务交易的交易特征信息可以为选择的业务交易的类型以及用户填写的业务交易的内容信息中获得，具体获得交易特征信息的方式不加限制。如，业务交易的交易特征可以包括：业务交易的名称、业务交易的交易渠道(如基于哪些软件或者平台的业务交易)、业务交易的金额以及业务交易的交易类型等特征中的部分或者全部。The transaction feature information of the business transaction may be information reflecting features such as the transaction content and transaction method of the business transaction. The transaction feature information of the business transaction can be obtained from the type of the selected business transaction and the content information of the business transaction filled in by the user, and the specific manner of obtaining the transaction feature information is not limited. For example, the transaction characteristics of the business transaction may include: the name of the business transaction, the transaction channel of the business transaction (such as which software or platform the business transaction is based on), the amount of the business transaction, and the transaction type of the business transaction. Some or all of the characteristics.
业务交易涉及到的客户是指该业务交易所涉及到的多个参与方主体，业务交易涉及到的客户可以有一个或者多个。如，以转账业务为例，转账涉及到的客户包括发起转账的发起客户，以及接受转账的个人或者企业等客户。The customers involved in a business transaction refer to multiple participants involved in the business transaction, and there may be one or more customers involved in the business transaction. For example, taking the transfer business as an example, the customers involved in the transfer include the initiating customer who initiates the transfer, and the individual or enterprise that accepts the transfer.
客户的客户信息可以包括从业务交易的内容中获得的客户的信息，如，可以为客户的姓名、证件号码以及手机号码等，还可以包括从数据库获得的该客户的学历、年龄以及职业等用户画像信息，对此不加限制。The customer's customer information can include the customer's information obtained from the content of the business transaction, such as the customer's name, certificate number and mobile phone number, etc., and can also include the customer's education, age and occupation obtained from the database. Image information, there is no restriction on this.
业务交易涉及到的客户所关联的历史业务交易是指历史上客户发起或者参与的业务交易，可选的，可以是客户曾经发起的业务交易。历史业务交易的交易特征信息具体包括的信息种类可以与上面业务交易的交易特征信息类似，在此不再赘述。The historical business transaction related to the customer involved in the business transaction refers to the business transaction initiated or participated in by the customer in the history, and optionally, the business transaction initiated by the customer. The type of information specifically included in the transaction feature information of the historical business transaction may be similar to the transaction feature information of the above business transaction, and details are not described herein again.
S102，基于业务交易的交易特征，利用第一预测模型确定业务交易存在洗钱风险的第一风险概率。S102. Based on the transaction characteristics of the business transaction, use a first prediction model to determine a first risk probability that the business transaction has a money laundering risk.
其中，第一预测模型为利用标注有洗钱风险标签的多个第一业务交易样本的交易特征信息训练得到的。The first prediction model is obtained by training using transaction feature information of a plurality of first business transaction samples marked with money laundering risk labels.
其中，业务交易样本的洗钱风险标签是用于表示该业务交易样本是否存在洗钱风险的标签。例如，业务交易样本为存在洗钱行为的业务交易，那么该业务交易样本为正样本，该业务交易样本的洗钱风险标签为1；而如果该业务交易样本为不存在洗钱行为的业务交易，那么该业务交易样本为负样本，其洗钱风险标签可以标记为0。The money laundering risk label of the business transaction sample is a label used to indicate whether the business transaction sample has money laundering risk. For example, if the business transaction sample is a business transaction with money laundering behavior, then the business transaction sample is a positive sample, and the money laundering risk label of the business transaction sample is 1; and if the business transaction sample is a business transaction without money laundering behavior, then the business transaction sample is a business transaction without money laundering behavior. The business transaction sample is a negative sample, and its money laundering risk label can be marked as 0.
为了便于区分，将用于训练模型的业务交易称为业务交易样本。类似的，为了区分，训练不同种预测模型所采用的业务交易样本，分别采用第一业务交易样本、第二业务交易样本和第三业务交易样本区分。但是可以理解的是，在实际应用中，训练不同预测模型所采用的多个业务交易样本可以是相同的，也可以是不同的，对此不加限制。For the convenience of distinction, the business transactions used to train the model are called business transaction samples. Similarly, in order to distinguish the business transaction samples used for training different prediction models, the first business transaction sample, the second business transaction sample and the third business transaction sample are respectively used to distinguish. However, it can be understood that, in practical applications, multiple business transaction samples used for training different prediction models may be the same or different, and there is no restriction on this.
其中，该第一预测模型适用于结合业务交易的交易特征预测业务交易是否存在洗钱风险的预测模型。The first prediction model is suitable for a prediction model for predicting whether there is a money laundering risk in a business transaction based on transaction characteristics of the business transaction.
该第一预测模型可以为任意经过训练的神经网络模型或者机器学习模型等，如第一预测模型可以为预测出的分类器等，对此不加限制。The first prediction model may be any trained neural network model or machine learning model, for example, the first prediction model may be a predicted classifier, etc., which is not limited.
可以理解的是，第一业务交易样本的交易特征信息的具体内容可以如前面业务交易的交易特征信息所包含的信息种类相似，在此不再赘述。It can be understood that the specific content of the transaction feature information of the first business transaction sample may be similar to the type of information included in the transaction feature information of the preceding business transaction, and details are not described herein again.
基于多个第一业务交易样本的交易特征信息训练该第一预测模型可以采用有监督训练方式进行训练，具体训练方式可以有多种可能，本申请对此不加限制。The training of the first prediction model based on the transaction feature information of the plurality of first business transaction samples may be performed in a supervised training manner, and the specific training manner may have various possibilities, which are not limited in this application.
S103，基于业务交易涉及到的客户信息，利用第二预测模型确定业务交易存在洗钱风险的第二风险概率。S103 , based on the customer information involved in the business transaction, use a second prediction model to determine a second risk probability that the business transaction has money laundering risk.
其中，该第二预测模型为利用标注有洗钱风险标签的多个第二业务交易样本涉及到的客户样本的客户信息训练得到的。Wherein, the second prediction model is obtained by training using customer information of customer samples involved in a plurality of second business transaction samples marked with money laundering risk labels.
为了便于区分，将第二业务交易样本涉及到的客户称为客户样本。客户样本的客户信息所涉及到的种类可以参见前面客户的客户信息的相关介绍，在此不再赘述。For the convenience of distinction, the customers involved in the second business transaction sample are called customer samples. For the types involved in the customer information of the customer sample, please refer to the relevant introduction of the customer information of the customer above, which will not be repeated here.
该第二预测模型同样可以为利用多个第二业务交易样本训练出的神经网络模型或者机器学习模型等，对此不加限制。The second prediction model may also be a neural network model or a machine learning model, etc. trained by using a plurality of second business transaction samples, which is not limited.
类似的，第二预测模型可以采用有监督训练方式进行训练得到，具体训练方式可以有多种可能，对此不加限制。Similarly, the second prediction model can be obtained by training in a supervised training method, and the specific training method may have multiple possibilities, which is not limited.
S104，基于历史业务交易的交易特征信息，利用第三预测模型确定业务交易存在洗钱风险的第三风险概率。S104 , based on the transaction feature information of the historical business transaction, use a third prediction model to determine a third risk probability that the business transaction has a money laundering risk.
其中，该第三预测模型为利用多个第三业务交易样本涉及到的客户样本关联的历史业务交易的交易特征信息训练得到的，第三业务交易样本同样标注有洗钱风险标签。The third prediction model is obtained by training using transaction feature information of historical business transactions associated with customer samples involved in a plurality of third business transaction samples, and the third business transaction samples are also marked with money laundering risk labels.
第三预测模型可以是经过训练的神经网络模型或者机器学习模型等，对此不加限制。训练该第三预测模型可以是任意有监督的训练方式，具体训练过程不加限制。The third prediction model may be a trained neural network model or a machine learning model, etc., which is not limited. Training the third prediction model may be any supervised training method, and the specific training process is not limited.
在一种可能的实现方式中，考虑到发起业务交易的发起客户的客户信息对于交易业务是否存在洗钱风险的影响较大，因此，本申请中可以业务交易的客户可以是发起该业务交易的发起客户的客户信息。相应的，可以基于该发起客户的客户信息，获得该发起客户关联的历史业务交易的交易特征信息。在此基础上，可以基于该发起客户关联的历史业务交易的交易特征信息，并利用该第三预测模型确定业务交易存在洗钱风险的第三风险概率。In a possible implementation manner, considering that the customer information of the initiating customer who initiates the business transaction has a great influence on whether there is a money laundering risk in the transaction business, the customer who can initiate the business transaction in this application may be the one who initiates the business transaction. Customer information of the customer. Correspondingly, transaction feature information of historical business transactions associated with the initiating customer may be obtained based on the customer information of the initiating customer. On this basis, based on the transaction feature information of the historical business transaction associated with the initiating customer, the third prediction model can be used to determine the third risk probability that the business transaction has money laundering risk.
S105，基于第一风险概率、第二风险概率和第三风险概率，确定业务交易的洗钱风险识别结果。S105: Determine the money laundering risk identification result of the business transaction based on the first risk probability, the second risk probability and the third risk probability.
其中，该洗钱风险识别结果用于表征业务交易存在洗钱风险的可能性，或者是，业务交易是否存在洗钱风险。Wherein, the money laundering risk identification result is used to represent the possibility that the business transaction has money laundering risk, or whether the business transaction has money laundering risk.
在一种可能的实现方式中，本申请可以还可以确定第一预测模型预测出的第一风险概率对应的第一权重，第二预测模型预测出的第二风险概率对应的第二权重，以及该第二预测模型预测出的第三风险概率对应的第三权重。相应的，可以基于该第一风险概率的第一权重，第二风险概率的第二权重以及第三风险概率的第三权重，确定该第一风险概率、第二风险概率和第三风险概率的综合风险概率，将该综合风险概率确定为该业务交易存在洗钱风险的概率。In a possible implementation manner, the present application may further determine the first weight corresponding to the first risk probability predicted by the first prediction model, the second weight corresponding to the second risk probability predicted by the second prediction model, and The third weight corresponding to the third risk probability predicted by the second prediction model. Correspondingly, based on the first weight of the first risk probability, the second weight of the second risk probability, and the third weight of the third risk probability, the first risk probability, the second risk probability and the third risk probability can be determined. The comprehensive risk probability is determined as the probability that the business transaction has money laundering risk.
如，可以计算该第一风险概率、第二风险概率和第三风险概率的加权求和，将加权求和结果作为综合风险概率。For example, the weighted summation of the first risk probability, the second risk probability and the third risk probability may be calculated, and the weighted summation result may be used as the comprehensive risk probability.
其中，第一权重、第二权重和第三权重可以是预先根据需要设定。Wherein, the first weight, the second weight and the third weight may be set in advance as required.
在一种可能的实现方式中，为了能够提高确定出的综合风险概率的准确性和合理性，本申请可以在训练出第一预测模型、第二预测模型和第三预测模型之后，分别确定第一预测模型的第一混淆矩阵，第二预测模型的第二混淆矩阵和第三预测模型的第三混淆矩阵。In a possible implementation manner, in order to improve the accuracy and rationality of the determined comprehensive risk probability, the present application may determine the first prediction model, the second prediction model and the third prediction model respectively after training the first prediction model, second prediction model and third prediction model. A first confusion matrix for the prediction model, a second confusion matrix for the second prediction model and a third confusion matrix for the third prediction model.
在此基础上，基于所述第一预测模型的第一混淆矩阵、所述第二预测模型的第二混淆矩阵以及第三预测模型的第三混淆矩阵，确定出第一权重、第二权重和第三权重。On this basis, based on the first confusion matrix of the first prediction model, the second confusion matrix of the second prediction model, and the third confusion matrix of the third prediction model, the first weight, the second weight and the third weight.
如，对于第一预测模型、第二预测模型和第三预测模型中任意一个预测模型而言，预测模型的混淆矩阵表征预测模型的预测准确度越高，该预测模型对应的权重越高。For example, for any one of the first prediction model, the second prediction model and the third prediction model, the confusion matrix of the prediction model indicates that the higher the prediction accuracy of the prediction model is, the higher the weight corresponding to the prediction model is.
其中，第一权重、第二权重和第三权重之和可以为1。The sum of the first weight, the second weight and the third weight may be 1.
预测模型的混淆矩阵是指混淆矩阵是用于总结分类模型(预测模型)预测结果的情形分析表，它以矩阵形式将训练样本集中各训练样本的真实类别与分类模型预测的预测类别两个标准进行汇总。由此可知，通过预测模型的混淆矩阵可以反映出预测模型预测准确度等信息，基于此，结合预测模型的混淆矩阵可以更为合理的配置不同预测模型的权重，以使得预测准确程度较高的预测模型对应的权重更高。The confusion matrix of the prediction model means that the confusion matrix is a situation analysis table used to summarize the prediction results of the classification model (prediction model). to summarize. It can be seen that the confusion matrix of the prediction model can reflect the prediction accuracy and other information of the prediction model. Based on this, the confusion matrix of the prediction model can be used to configure the weights of different prediction models more reasonably, so that the prediction accuracy is higher. The prediction model corresponds to a higher weight.
在实际应用中，本申请可以结合第一混淆矩阵、第二混淆矩阵和第三混淆矩阵，并基于贝叶斯估计算法分别确定出第一预测模型对应的第一权重，第二预测模型对应的第二权重以及第三预测模型对应的第三权重。在此基础上，每次获得业务交易之后，可以结合三种维度的预测模型分别确定出的风险概率以及每种预测模型对应的权重，可以结合贝叶斯估计算法进行加权融合，从而可以得到融合后的风险概率。In practical applications, the present application can combine the first confusion matrix, the second confusion matrix and the third confusion matrix, and determine the first weight corresponding to the first prediction model based on the Bayesian estimation algorithm, and the corresponding weight of the second prediction model. The second weight and the third weight corresponding to the third prediction model. On this basis, after each business transaction is obtained, the risk probability determined by the three-dimensional prediction models and the corresponding weight of each prediction model can be combined, and the Bayesian estimation algorithm can be used for weighted fusion, so that the fusion can be obtained. subsequent risk probability.
由以上可知，在本申请实施例中，分别从业务交易的交易特征信息、涉及到的客户的客户信息以及客户关联的历史业务交易的交易特征信息三个维度分析该业务交易存在洗钱风险的风险情况，并最终结合这三个维度的分析结果综合识别业务交易存在洗钱行为的风险情况，从而可以更为全面分析业务交易可能存在洗钱风险的可能性，可以提升识别存在洗钱风险的业务交易的准确性和可靠性。It can be seen from the above that in the embodiment of the present application, the risk of money laundering risk in the business transaction is analyzed from three dimensions, namely, the transaction feature information of the business transaction, the customer information of the involved customers, and the transaction feature information of the historical business transactions associated with the customer. Finally, combined with the analysis results of these three dimensions, we can comprehensively identify the risk of money laundering in business transactions, so that we can more comprehensively analyze the possibility of money laundering risks in business transactions, and improve the accuracy of identifying business transactions with money laundering risks. sturdiness and reliability.
可以理解的是，在本申请实施例中，在获得业务交易的三个维度的信息之后，还需要执行去除奇异值等数据预处理操作，对此不加限制。It can be understood that, in this embodiment of the present application, after obtaining the information of the three dimensions of the business transaction, it is also necessary to perform data preprocessing operations such as removing singular values, which is not limited.
在实际应用中，本申请在基于预测模型预测某个维度的信息对应的风险概率之前，还可以先对该维度的信息进行特征提取，下面结合一种可能实现方式进行说明。In practical applications, before predicting the risk probability corresponding to the information of a certain dimension based on the prediction model, the present application may first perform feature extraction on the information of the dimension, which is described below with reference to a possible implementation manner.
如图2所示，其示出了本申请实施例提供的交易风险识别方法的又一种流程示意图，本实施例的方法可以包括：As shown in FIG. 2, which shows another schematic flowchart of the transaction risk identification method provided by the embodiment of the present application, the method of the present embodiment may include:
S201，获得业务交易的交易特征信息以及所述业务交易涉及到的客户的客户信息。S201: Obtain transaction feature information of a business transaction and customer information of customers involved in the business transaction.
基于所述发起客户的客户信息，获得所述发起客户关联的历史业务交易的交易特征信息。Based on the customer information of the initiating customer, transaction feature information of the historical business transaction associated with the initiating customer is obtained.
其中，客户的客户信息至少包括：发起业务交易的客户的客户信息。Wherein, the client information of the client includes at least: client information of the client who initiates the business transaction.
S202，基于发起业务交易的发起客户的客户信息，获得该发起客户关联的历史业务交易的交易特征信息。S202, based on the customer information of the initiating customer who initiates the business transaction, obtain transaction feature information of the historical business transaction associated with the initiating customer.
如，可以将该发起客户历史上发起的各业务交易均确定为该发起客户关联的历史业务交易。For example, each business transaction initiated by the initiating customer in the history may be determined as the historical business transaction associated with the initiating customer.
S203，确定该业务交易的交易特征信息对应的第一特征向量，该业务交易涉及到的客户的客户信息对应的第二特征向量，该发起客户关联的历史业务交易的交易特征信息对应的第三特征向量。S203: Determine a first feature vector corresponding to the transaction feature information of the business transaction, a second feature vector corresponding to the customer information of the customer involved in the business transaction, and a third feature vector corresponding to the transaction feature information of the historical business transaction associated with the initiating customer Feature vector.
其中，从业务交易的交易特征信息中提取特征信息，可以得到第一特征向量，基于交易特征信息构建第一特征向量的方式可以有多种可能，本申请对此不加限制。类似的，第二特征向量和第三特征向量同样可以通过特征提取或者向量转换等方式得到，具体构建方式同样不加限制。The first feature vector can be obtained by extracting the feature information from the transaction feature information of the business transaction, and there are multiple possibilities for constructing the first feature vector based on the transaction feature information, which is not limited in this application. Similarly, the second feature vector and the third feature vector can also be obtained through feature extraction or vector transformation, and the specific construction methods are also not limited.
S204，将第一特征向量输入预先训练出的第一预测模型，得到该业务交易存在洗钱风险的第一风险概率。S204: Input the first feature vector into the pre-trained first prediction model to obtain the first risk probability that the business transaction has money laundering risk.
其中，该第一预测模型为利用标注有洗钱风险标签的多个第一业务交易样本的交易特征信息训练得到的。The first prediction model is obtained by training using transaction feature information of a plurality of first business transaction samples marked with money laundering risk labels.
S205，将第二特征向量输入预先训练出的第二预测模型，得到该业务交易存在洗钱风险的第二风险概率。S205: Input the second feature vector into the pre-trained second prediction model to obtain the second risk probability that the business transaction has money laundering risk.
其中，第二预测模型为利用标注有洗钱风险标签的多个第二业务交易样本涉及到的客户样本的客户信息训练得到的。Wherein, the second prediction model is obtained by training using customer information of customer samples involved in a plurality of second business transaction samples marked with money laundering risk labels.
S206，将该第三特征向量输入到第三预测模型，得到该业务交易存在洗钱风险的第三风险概率。S206: Input the third feature vector into a third prediction model to obtain a third risk probability that the business transaction has money laundering risk.
其中，第三预测模型为利用多个第三业务交易样本涉及到的客户样本关联的历史业务交易的交易特征信息训练得到的，该第三业务交易样本标注有洗钱风险标签。The third prediction model is obtained by training using transaction feature information of historical business transactions associated with customer samples involved in a plurality of third business transaction samples, and the third business transaction samples are marked with a money laundering risk label.
S207，获得配置的第一风险概率对应的第一权重、第二风险概率对应的第二权重以及该第三风险概率对应的第三权重。S207: Obtain a first weight corresponding to the configured first risk probability, a second weight corresponding to the second risk probability, and a third weight corresponding to the third risk probability.
其中，该第一权重、第二权重和第三权重可以为基于第一预测模型的第一混淆矩阵、第二预测模型的第二混淆矩阵以及第三预测模型的第三混淆矩阵，并基于贝叶斯估计算法确定出的。The first weight, the second weight and the third weight may be the first confusion matrix based on the first prediction model, the second confusion matrix based on the second prediction model, and the third confusion matrix based on the third prediction model, and are based on the It is determined by the Yess estimation algorithm.
其中，第一混淆矩阵为在训练出第一预测模型之后，基于多个第一业务交易样本各自标注的洗钱风险标签以及第一预测模型预测出的各第一业务交易样本的风险概率构建出的。The first confusion matrix is constructed based on the money laundering risk labels marked by the multiple first business transaction samples and the risk probability of each first business transaction sample predicted by the first prediction model after the first prediction model is trained. .
类似的，第二混淆矩阵为在训练出第二预测模型之后，基于多个第二业务交易样本各自标注的洗钱风险标签以及第二预测模型预测出的各第二业务交易样本的风险概率构建出的。Similarly, the second confusion matrix is constructed based on the money laundering risk labels marked by the multiple second business transaction samples and the risk probability of each second business transaction sample predicted by the second prediction model after the second prediction model is trained. of.
第三混淆矩阵为在训练出第三预测模型之后，基于多个第三业务交易样本各自标注的洗钱风险标签以及第三预测模型预测出的各第三业务交易样本的风险概率构建出的。The third confusion matrix is constructed based on the money laundering risk labels marked by the multiple third business transaction samples and the risk probability of each third business transaction sample predicted by the third prediction model after the third prediction model is trained.
S208，基于第一风险概率、第一风险概率对应的第一权重、第二风险概率、第二风险概率对应的第二权重、第三风险概率以及第三风险概率对应的第三权重，计算第一风险概率、第二风险概率和第三风险概率的加权和，将加权和确定为该业务交易存在洗钱风险的风险概率。S208, based on the first risk probability, the first weight corresponding to the first risk probability, the second risk probability, the second weight corresponding to the second risk probability, the third risk probability, and the third weight corresponding to the third risk probability, calculate the first risk probability The weighted sum of the first risk probability, the second risk probability and the third risk probability is determined as the risk probability that the business transaction has money laundering risk.
需要说明的，本实施例步骤S207和S209以结合贝叶斯估计算法对三个维度的风险概率进行加权融合来确定该业务交易最终的风险概率为例说明，但是结合第一风险概率、第二风险概率和第三风险概率，通过其他方式来确定业务交易最终存在洗钱风险的风险概率也同样适用于本申请，对此不加限制。It should be noted that steps S207 and S209 in this embodiment are described by taking the Bayesian estimation algorithm for weighted fusion of the risk probability of three dimensions to determine the final risk probability of the business transaction as an example, but combining the first risk probability and the second risk probability The risk probability and the third risk probability, and other methods to determine the risk probability that a business transaction ultimately has money laundering risk are also applicable to this application, and there are no restrictions on this.
本实施例结合了贝叶斯估计算法对三个维度确定的风险概率进行加权融合，以确定业务交易最终存在洗钱风险的概率。This embodiment combines the Bayesian estimation algorithm to perform weighted fusion of the risk probabilities determined by the three dimensions, so as to determine the probability that the business transaction ultimately has money laundering risks.
为了便于理解贝叶斯估计算法的加权融合过程，下面对贝叶斯估计的简单原理进行说明：In order to facilitate the understanding of the weighted fusion process of the Bayesian estimation algorithm, the simple principle of Bayesian estimation is described below:
给定C个不同的类别(如，在本案场景中可以认为是存在洗钱风险和不存在洗钱风险两个类别)，观测值x(如，可以对应为本案中业务交易的交易特征信息、客户的客户信息以及历史交易的交易特征信息)，K个分类器(如本案中K可以取3，分类器为预测模型)，对于每个分类器λi，真实的行为类别标签ω(如，真实的洗钱风险标签)，则预测的行为类别标签为：Given C different categories (for example, in the scenario of this case, it can be considered that there are two categories of money laundering risk and no money laundering risk), the observed value x (for example, can correspond to the transaction feature information of the business transaction in this case, the customer's customer information and transaction feature information of historical transactions), K classifiers (for example, K can be 3 in this case, and the classifier is a prediction model), for each classifier λi, the real behavior category label ω (such as the real money laundering risk label), the predicted behavior category label is:
其中，P(ω|x)为在观测值x的前提下，分类器预测出行为类别标签为真实的行为类别标签ω的概率，即分类器正确输出的概率。P(ω,λi|x)为观测值x的前提下，分类器λi预测出的行为类别标签为真实的行为类别标签ω的概率。Among them, P(ω|x) is the probability that the classifier predicts that the behavior class label is the real behavior class label ω under the premise of the observation value x, that is, the probability that the classifier outputs the correct output. Under the premise that P(ω,λi|x) is the observation value x, the probability that the behavior class label predicted by the classifier λi is the real behavior class label ω.
设单个分类器的输出为随机变量
即，which is,
其中，
其中，
在此基础上，每个分类器的综合决策为：On this basis, the comprehensive decision of each classifier is:
其中，在输入为x、分类器为λi，且带权重的输出概率条件下，分类器正确输出的概率。由于公式四中
其中，
对应本申请的一种交易风险识别方法，本申请还提供了一种交易风险识别装置。Corresponding to a transaction risk identification method of the present application, the present application further provides a transaction risk identification device.
如图3所示，其示出了本申请实施例提供的交易风险识别装置的一种组成结构示意图，本实施例的装置可以包括：As shown in FIG. 3 , which shows a schematic structural diagram of a transaction risk identification device provided in an embodiment of the present application, the device in this embodiment may include:
信息获得单元301，用于获得业务交易的交易特征信息、所述业务交易涉及到的客户的客户信息以及所述客户关联的历史业务交易的交易特征信息；an
第一预测单元302，用于基于所述业务交易的交易特征，利用第一预测模型确定所述业务交易存在洗钱风险的第一风险概率，所述第一预测模型为利用标注有洗钱风险标签的多个第一业务交易样本的交易特征信息训练得到的；The
第二预测单元303，用于基于业务交易涉及到的客户信息，利用第二预测模型确定所述业务交易存在洗钱风险的第二风险概率，所述第二预测模型为利用标注有洗钱风险标签的多个第二业务交易样本涉及到的客户样本的客户信息训练得到的；The
第三预测单元304，用于基于所述历史业务交易的交易特征信息，利用第三预测模型确定所述业务交易存在洗钱风险的第三风险概率，所述第三预测模型为利用多个第三业务交易样本涉及到的客户样本关联的历史业务交易的交易特征信息训练得到的，所述第三业务交易样本标注有洗钱风险标签；The
风险确定单元305，用于基于所述第一风险概率、第二风险概率和第三风险概率，确定所述业务交易的洗钱风险识别结果。A
在一种可能的实现方式中，所述风险确定单元，包括：In a possible implementation manner, the risk determination unit includes:
权重确定单元，用于确定所述第一风险概率对应的第一权重、所述第二风险概率对应的第二权重以及所述第三风险概率对应的第三权重；a weight determination unit, configured to determine a first weight corresponding to the first risk probability, a second weight corresponding to the second risk probability, and a third weight corresponding to the third risk probability;
风险概率确定单元，用于基于所述第一风险概率的第一权重，第二风险概率的第二权重以及第三风险概率的第三权重，确定所述第一风险概率、第二风险概率和第三风险概率的综合风险概率，将所述综合风险概率确定为所述业务交易存在洗钱风险的概率。a risk probability determination unit, configured to determine the first risk probability, the second risk probability and the The comprehensive risk probability of the third risk probability is determined as the probability that the business transaction has money laundering risk.
在又一种可能的实现方式中，所述权重确定单元中的第一权重、第二权重和第三权重为基于所述第一预测模型的第一混淆矩阵、所述第二预测模型的第二混淆矩阵以及第三预测模型的第三混淆矩阵确定的。In yet another possible implementation manner, the first weight, the second weight and the third weight in the weight determination unit are a first confusion matrix based on the first prediction model, a first confusion matrix based on the second prediction model The second confusion matrix and the third confusion matrix of the third prediction model are determined.
在又一种可能的实现方式中，所述信息获得单元，包括：In yet another possible implementation, the information obtaining unit includes:
第一信息获得单元，用于获得业务交易的交易特征信息以及所述业务交易涉及到的客户的客户信息，所述客户的客户信息至少包括：发起所述业务交易的发起客户的客户信息；a first information obtaining unit, configured to obtain transaction feature information of a business transaction and client information of a client involved in the business transaction, where the client information at least includes: client information of the initiating client who initiated the business transaction;
第二信息获得单元，用于基于所述发起客户的客户信息，获得所述发起客户关联的历史业务交易的交易特征信息。The second information obtaining unit is configured to obtain transaction feature information of historical business transactions associated with the initiating customer based on the customer information of the initiating customer.
在又一种可能的实现方式中，还包括：In yet another possible implementation, it also includes:
第一向量确定单元，用于在所述获得业务交易的交易特征信息、所述业务交易涉及到的客户的客户信息以及所述客户关联的历史业务交易的交易特征信息之后，确定所述业务交易的交易特征信息对应的第一特征向量；a first vector determination unit, configured to determine the business transaction after obtaining the transaction feature information of the business transaction, the customer information of the customer involved in the business transaction, and the transaction feature information of the historical business transaction associated with the customer The first feature vector corresponding to the transaction feature information of ;
第二向量确定单元，用于确定所述业务交易涉及到的客户的客户信息对应的第二特征向量；A second vector determination unit, configured to determine a second feature vector corresponding to the customer information of the customer involved in the business transaction;
第三向量确定单元，用于确定所述历史业务交易的交易特征信息对应的第三特征向量；A third vector determination unit, configured to determine a third feature vector corresponding to the transaction feature information of the historical business transaction;
所述第一预测单元，包括：The first prediction unit includes:
第一预测子单元，用于将所述第一特征向量输入第一预测模型，得到所述第一预测模型输出的所述业务交易存在洗钱风险的第一风险概率；a first prediction subunit, configured to input the first feature vector into a first prediction model to obtain a first risk probability of money laundering risk in the business transaction output by the first prediction model;
所述第二预测单元，包括：The second prediction unit includes:
第二预测子单元，用于将所述第二特征向量输入到所述第二预测模型，得到所述第二预测模型输出的所述业务交易存在洗钱风险的第二风险概率；a second prediction subunit, configured to input the second feature vector into the second prediction model, to obtain a second risk probability of money laundering risk in the business transaction output by the second prediction model;
所述第三预测单元，包括：The third prediction unit includes:
第三预测子单元，用于将所述第三特征向量输入到第三预测模型，得到所述第三预测模型输出的所述业务交易存在洗钱风险的第三风险概率。The third prediction subunit is configured to input the third feature vector into a third prediction model, and obtain a third risk probability of money laundering risk in the business transaction output by the third prediction model.
需要说明的是，本说明书中的各个实施例均采用递进的方式描述，每个实施例重点说明的都是与其他实施例的不同之处，各个实施例之间相同相似的部分互相参见即可。同时，本说明书中各实施例中记载的特征可以相互替换或者组合，使本领域专业技术人员能够实现或使用本申请。对于装置类实施例而言，由于其与方法实施例基本相似，所以描述的比较简单，相关之处参见方法实施例的部分说明即可。It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. For the same and similar parts among the various embodiments, refer to each other Can. Meanwhile, the features described in each embodiment in this specification can be replaced or combined with each other, so that those skilled in the art can realize or use the present application. As for the apparatus-type embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and for related parts, please refer to the partial descriptions of the method embodiments.
最后，还需要说明的是，在本文中，诸如第一和第二等之类的关系术语仅仅用来将一个实体或者操作与另一个实体或操作区分开来，而不一定要求或者暗示这些实体或操作之间存在任何这种实际的关系或者顺序。而且，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者设备不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者设备所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括要素的过程、方法、物品或者设备中还存在另外的相同要素。Finally, it should also be noted that in this document, relational terms such as first and second are used only to distinguish one entity or operation from another, and do not necessarily require or imply these entities or that there is any such actual relationship or sequence between operations. Moreover, the terms "comprising", "comprising" or any other variation thereof are intended to encompass a non-exclusive inclusion such that a process, method, article or device comprising a list of elements includes not only those elements, but also includes not explicitly listed or other elements inherent to such a process, method, article or apparatus. Without further limitation, an element qualified by the phrase "comprising a..." does not preclude the presence of additional identical elements in the process, method, article, or device that includes the element.
对所公开的实施例的上述说明，使本领域技术人员能够实现或使用本申请。对这些实施例的多种修改对本领域技术人员来说将是显而易见的，本文中所定义的一般原理可以在不脱离本申请的精神或范围的情况下，在其它实施例中实现。因此，本申请将不会被限制于本文所示的这些实施例，而是要符合与本文所公开的原理和新颖特点相一致的最宽的范围。The above description of the disclosed embodiments enables any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, this application is not intended to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
以上仅是本申请的优选实施方式，应当指出，对于本技术领域的普通技术人员来说，在不脱离本申请原理的前提下，还可以做出若干改进和润饰，这些改进和润饰也应视为本申请的保护范围。The above are only the preferred embodiments of the present application. It should be pointed out that for those skilled in the art, without departing from the principles of the present application, several improvements and modifications can also be made, and these improvements and modifications should also be regarded as The protection scope of this application.
Claims (10)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210466159.7ACN114819975A (en) | 2022-04-29 | 2022-04-29 | Transaction risk identification method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210466159.7ACN114819975A (en) | 2022-04-29 | 2022-04-29 | Transaction risk identification method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114819975Atrue CN114819975A (en) | 2022-07-29 |
Family
ID=82509392
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210466159.7APendingCN114819975A (en) | 2022-04-29 | 2022-04-29 | Transaction risk identification method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114819975A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN119444227A (en)* | 2024-10-23 | 2025-02-14 | 中国平安财产保险股份有限公司 | Suspicious transaction assessment method and device, electronic device and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7805362B1 (en)* | 2006-10-10 | 2010-09-28 | United Services Automobile Association (Usaa) | Methods of and systems for money laundering risk assessment |
| WO2019157946A1 (en)* | 2018-02-13 | 2019-08-22 | 阿里巴巴集团控股有限公司 | Anti-money laundering method, apparatus, and device |
| CN112598311A (en)* | 2020-12-29 | 2021-04-02 | 中国农业银行股份有限公司 | Risk operation identification model construction method and risk operation identification method |
| CN113095927A (en)* | 2021-02-23 | 2021-07-09 | 广发证券股份有限公司 | Method and device for identifying suspicious transactions of anti-money laundering |
| KR20210138297A (en)* | 2020-05-12 | 2021-11-19 | (주)이엘온소프트 | Risk assessment system for anti-money laundering and method thereof |
| CN114187154A (en)* | 2021-12-13 | 2022-03-15 | 平安付科技服务有限公司 | Anti-money laundering identification method and device, computer equipment and storage medium |
- 2022
- 2022-04-29CNCN202210466159.7Apatent/CN114819975A/enactivePending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7805362B1 (en)* | 2006-10-10 | 2010-09-28 | United Services Automobile Association (Usaa) | Methods of and systems for money laundering risk assessment |
| WO2019157946A1 (en)* | 2018-02-13 | 2019-08-22 | 阿里巴巴集团控股有限公司 | Anti-money laundering method, apparatus, and device |
| KR20210138297A (en)* | 2020-05-12 | 2021-11-19 | (주)이엘온소프트 | Risk assessment system for anti-money laundering and method thereof |
| CN112598311A (en)* | 2020-12-29 | 2021-04-02 | 中国农业银行股份有限公司 | Risk operation identification model construction method and risk operation identification method |
| CN113095927A (en)* | 2021-02-23 | 2021-07-09 | 广发证券股份有限公司 | Method and device for identifying suspicious transactions of anti-money laundering |
| CN114187154A (en)* | 2021-12-13 | 2022-03-15 | 平安付科技服务有限公司 | Anti-money laundering identification method and device, computer equipment and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 王成;王昌琪;: "一种面向网络支付反欺诈的自动化特征工程方法", 计算机学报, no. 10, 15 October 2020 (2020-10-15), pages 23 - 26* |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN119444227A (en)* | 2024-10-23 | 2025-02-14 | 中国平安财产保险股份有限公司 | Suspicious transaction assessment method and device, electronic device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP6913241B2 (en) | Systems and methods for issuing loans to consumers who are determined to be creditworthy | |
| WO2019196546A1 (en) | Method and apparatus for determining risk probability of service request event | |
| CN109544190A (en) | A kind of fraud identification model training method, fraud recognition methods and device | |
| CN109598331A (en) | A kind of fraud identification model training method, fraud recognition methods and device | |
| US20230088840A1 (en) | Dynamic assessment of cryptocurrency transactions and technology adaptation metrics | |
| CN112200660B (en) | Bank counter business supervision method, device and equipment | |
| CN111325619A (en) | A method and device for updating a credit card fraud detection model based on joint learning | |
| CN111553701A (en) | Session-based risk transaction determination method and device | |
| CN112150153A (en) | Telecommunication fraud user identification method and device | |
| US20220245426A1 (en) | Automatic profile extraction in data streams using recurrent neural networks | |
| CN111932130A (en) | Service type identification method and device | |
| CN113887214B (en) | Willingness presumption method based on artificial intelligence and related equipment thereof | |
| US20220327542A1 (en) | Self Learning Machine Learning Transaction Scores Adjustment via Normalization Thereof Accounting for Underlying Transaction Score Bases | |
| CN110930038A (en) | Loan demand identification method, loan demand identification device, loan demand identification terminal and loan demand identification storage medium | |
| CN111582878A (en) | A transaction risk prediction method, device and system | |
| CN112330328A (en) | Credit card fraud detection method based on feature extraction | |
| Nasr et al. | A proposed fraud detection model based on e-Payments attributes a case study in Egyptian e-Payment gateway | |
| WO2019023406A1 (en) | System and method for detecting and responding to transaction patterns | |
| CN112669154B (en) | Foreign currency exchange business development prediction implementation method and device and computer equipment | |
| CN114819975A (en) | Transaction risk identification method and device | |
| CN115545886A (en) | Overdue risk identification method, overdue risk identification device, overdue risk identification equipment and storage medium | |
| Shaik et al. | Customer loan eligibility prediction using machine learning | |
| WO2025139251A1 (en) | Method and apparatus for training risk identification model | |
| CN111553685B (en) | Method, device, electronic equipment and storage medium for determining transaction routing channel | |
| CN113344581B (en) | Service data processing 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 |