WO2018176179A1 - 基于信用大数据的跨境支付快速结算方法 - Google Patents

基于信用大数据的跨境支付快速结算方法 Download PDF

Info

Publication number
WO2018176179A1
WO2018176179A1 PCT/CN2017/000351 CN2017000351W WO2018176179A1 WO 2018176179 A1 WO2018176179 A1 WO 2018176179A1 CN 2017000351 W CN2017000351 W CN 2017000351W WO 2018176179 A1 WO2018176179 A1 WO 2018176179A1
Authority
WO
WIPO (PCT)
Prior art keywords
data
merchant
transaction
neural network
credit
Prior art date
Application number
PCT/CN2017/000351
Other languages
English (en)
French (fr)
Inventor
熊伟
陈宇
芦帅
汪宁
陈鹏
Original Assignee
杭州呯嘭智能技术有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 杭州呯嘭智能技术有限公司 filed Critical 杭州呯嘭智能技术有限公司
Publication of WO2018176179A1 publication Critical patent/WO2018176179A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q20/00—Payment architectures, schemes or protocols
    • G06Q20/08—Payment architectures
    • G06Q20/085—Payment architectures involving remote charge determination or related payment systems
    • G06Q20/0855—Payment architectures involving remote charge determination or related payment systems involving a third party
    • 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
    • G06Q20/403—Solvency checks
    • G06Q20/4037—Remote solvency checks
    • 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

Definitions

  • the invention relates to the field of cross-border payment technology, in particular to a fast settlement method for cross-border payment based on credit big data, which can improve the efficiency of merchants' capital transfer, shorten the time of receipt of funds, and reduce the risk of exchange rate fluctuation in the clearing process of funds.
  • the cross-border e-commerce transaction volume reached 1.6 trillion yuan in 2011, a year-on-year increase of 33%; in 2012, the cross-border transaction volume was 2 trillion yuan, a year-on-year increase of more than 25%, and the growth rate was much higher than the growth rate of foreign trade. It has undertaken an important task for the transformation and upgrading of China's foreign trade. Judging from the distribution of cross-border e-commerce and export structure, more than 90% of the transaction volume in 2012 was contributed by export e-commerce trade.
  • cross-border e-commerce requires strong cross-border payment methods as a foundation.
  • electronic payment mainly refers to the e-commerce payment platform to establish a connection between online merchants and banks through the use of standardized connectors. It solves the problems of online currency payment, cash flow, fund clearing, and query statistics from consumers to financial institutions and merchants.
  • Cross-border payment refers to the inter- and inter-regional transfer of funds between the two countries or regions, such as inter-regional trade, international investment and other aspects of international bond transactions, through certain settlement tools and payment systems.
  • the payment performance is the foreign card acquiring business, that is, the domestic merchants sell the goods to the overseas consumers through the foreign trade platform.
  • the payment institution collects the foreign currency for the domestic merchants and acts as the agent to settle the foreign exchange.
  • cross-border e-commerce companies mainly have five different business models in cross-border commodity transactions: traditional cross-border transaction information platform model, portal-type B2B integrated platform model, and comprehensive vertical cross-border micro-platform (B2C, C2C).
  • B2B traditional cross-border transaction information platform model
  • portal-type B2B integrated platform model portal-type B2B integrated platform model
  • comprehensive vertical cross-border micro-platform (B2C, C2C) B2B) mode
  • third-party service platform (generational operation) mode vertical cross-border micro-transaction website (independent B2C) mode.
  • cross-border electronic payment business will involve the settlement and sale of funds and the receipt and payment of foreign exchange.
  • cross-border e-commerce import business (including individual consumer Haitao) involves cross-border payment for foreign exchange purchase.
  • the foreign exchange purchase method generally has third-party purchase of foreign exchange, overseas e-commerce accepts RMB payment, and domestic Bank purchases and remittances, etc.
  • the cross-border e-commerce export business involves the settlement of cross-border income, and its foreign exchange settlement mainly includes the settlement of foreign exchange by third parties, remittance through domestic banks, and the split and inflow of foreign exchange in the name of foreign exchange settlement.
  • third-party payment platforms can simultaneously satisfy users' convenience and low fees for cross-border remittances.
  • the rate requirements therefore, are favored by more and more cross-border merchants.
  • most of the domestic third-party payment is through the cooperation with the bank, the currency exchange and payment process is completed by its custodian bank, due to the e-commerce platform account period, bank settlement and sales exchange audit management process
  • the payment process is complicated, the efficiency is low, and the account period is long. As shown in Figure 1, it usually takes 6 to 23 working days to receive the payment.
  • the object of the present invention is to overcome the shortcomings of the conventional process of cross-border e-commerce payment in the prior art, the long period of capital account, and the high process risk, and provide a method for improving the efficiency of the flow of merchants and shortening the time of receipt of funds.
  • a fast settlement method for cross-border payment based on credit big data comprising a third-party payment platform, a plurality of e-commerce platforms, a data interface module, a data standardization module, a smart credit evaluation module and a process control module;
  • the third-party payment platform If the evaluation result is lower than the preset threshold, there is a greater risk that the transaction will be advanced in advance, and the third-party payment platform will process according to the normal process, that is, after the end of the e-commerce platform account period, The e-commerce platform settles to the overseas account opened by the customer on the third-party payment platform. After the third-party payment platform receives the payment, after the compliance review, the cooperative bank distributes the compliance fund to the merchant's domestic bank account to complete the entire process. .
  • the data interface module is responsible for transaction screening, field clipping, and data loading, and is a data exchange interface between the system of the present invention and an e-commerce platform and a third-party payment platform, and is used for obtaining historical transaction data from an e-commerce platform (the average daily value of the merchant within 30 days) Trading volume, total transaction volume of merchants for three months, merchant registration duration, evaluation and return rate of traded goods, type of traded goods, average time of confirmation of traded goods, average time of confirmation of customers, historical evaluation of all purchased goods by customers, Customer's return rate and other information), the above data is used for smart credit evaluation
  • the module conducts the credit evaluation model training and the credit evaluation of whether the merchant order funds are settled in advance.
  • the data interface module is sent to the third party payment platform to notify the third party to pay.
  • the platform advances the merchant transaction funds with high credit evaluation in advance and settles them in the bank account of the merchant.
  • the data standardization module converts the transaction data obtained by the above data interface module into a unified format that can be recognized by the intelligent credit evaluation module. Standardization is also referred to as pre-processing and normalization processing in the input data processing described in the present invention. Due to the type of transaction data, the range of values varies greatly. The commonly used mathematical model for credit evaluation based on big data generally requires the value of the input data to be between 0 and 1. Therefore, the input data is preprocessed. In addition, in order to ensure that the established model has a certain extrapolation ability, it is preferable to make the data pre-processed value between 0.2 and 0.8.
  • the intelligent credit evaluation module calculates the relevant credit evaluation value based on the standardized transaction data, and decides whether to advance the funds of the transaction in advance.
  • merchant credit evaluation methods based on transaction big data. Commonly include rule base classification, artificial neural network, decision tree, etc.
  • the input value of this module is the standardized transaction data, and the output is the credit evaluation value of the merchant transaction ( Between 0 and 1), the higher the score, the higher the credibility of the transaction on behalf of the merchant, and the lower the risk that the third-party payment system advances the payment in advance.
  • the process control module controls the settlement process of the merchant payment based on the result of the smart credit evaluation. If the result of the smart credit evaluation is higher than a predetermined threshold (for example, 0.6), the system determines that the risk of the advance payment of the transaction is small. The corresponding funds can be advanced in advance by the third-party platform, and the system enters the advance advance payment processing process; if the result of the smart credit evaluation is lower than a predetermined threshold (for example, 0.6), the system determines the advance advance of the transaction. If there is a risk, the system will still process according to the normal process, that is, wait for the e-commerce platform to be processed after the account period has passed.
  • a predetermined threshold for example, 0.6
  • the system notifies the third-party payment platform to enter the advance advance payment process, and the third-party payment platform notifies the cooperative bank to distribute after performing the compliance audit.
  • the transaction amount will be credited to the merchant's domestic bank account in advance. As shown in Figure 2, the entire process can take only one to one business day to complete the settlement of the client's funds.
  • the system notifies the third-party payment platform to process according to the normal process, that is, after the end of the e-commerce platform account period (transaction and electronic).
  • the business platform processing process is generally 3 to 17 working days, depending on the delivery speed, user confirmation speed, and platform processing time limit. It is settled by the e-commerce platform to the overseas account opened by the customer on the third-party payment platform (3 to 5 On the working day), after the third-party payment platform receives the payment, after the compliance review, the cooperative bank will distribute the compliance funds to the merchant's domestic bank account, and complete the entire process, which takes about 6 to 23 working days to complete. The settlement of the client's funds is settled.
  • the invention is based on historical transaction data of multiple dimensions of the platform (average daily trading volume within 30 days of the merchant, total transaction volume of the merchant in three months, merchant registration duration, evaluation and return rate of the transaction commodity, type of the transaction commodity, average confirmation of the transaction commodity Time, customer's average confirmation time, customer's historical evaluation of all purchased goods, customer's return rate and other information), multi-dimensional comprehensive evaluation of current trading orders, get customer credit evaluation based on platform trading big data, for higher credit evaluation
  • the merchants simplify the cross-border liquidation process of their funds, and the system obtains the confirmation information of the cross-border e-commerce platform customers to purchase goods and pay, that is, in advance
  • the three-party payment platform advances the corresponding payment of the merchant, and the third-party platform completes the transaction review and the cooperative bank completes the compliance audit and then distributes the settlement to the merchant bank account.
  • the whole process can be shortened to 0 to 1 working day, which is extremely large.
  • the invention patent is aimed at the problem that the cross-border merchants have complicated collection process, high cost and long account period. Data mining and analysis are carried out on the huge user and merchant transaction information accumulated by the platform, and the credit evaluation method based on transaction big data is adopted. Shorten the settlement process for high-credit transactions, carry out early lending, and achieve the effect of simplifying cross-border clearing process, reducing merchant capital cost, and greatly shortening the time of funds arrival while controlling risks.
  • the historical transaction data includes a daily average transaction volume of the merchant within 30 days, a total transaction volume of the merchant for three months, a merchant registration duration, an evaluation of the transaction commodity and a return rate, a type of the transaction commodity, an average confirmation time of the transaction commodity, and a customer.
  • the merchant credit evaluation method includes a rule base classification method, an artificial neural network method, and a decision tree method.
  • the artificial neural network method comprises the following steps:
  • step (4-1) comprises the following steps:
  • a batch of learning samples is selected to form a data sample set.
  • the data sample set includes learning samples with good credit and poor credit.
  • Each learning sample includes training data, verification data and test data, and the neural network model is trained by using the data sample set. Make nerves
  • the network model has the ability to evaluate credits;
  • the standardized historical transaction data is used as an input sample of the neural network model, and sent to the neural network for learning.
  • the training data of the data sample set is used to train the neural network, and the verification data of the data sample set is used to test the learning success of the neural network, and the data sample is used.
  • the aggregated test data tests whether the neural network has good generalization ability for credit evaluation;
  • the training is interrupted, and the neural network model outputs the Y value obtained according to the fund performance in the historical transaction.
  • step (4-2) comprises the following steps:
  • the connection weight is v j
  • the node offset coefficient of the hidden node layer is ⁇ j
  • the node offset coefficient of the output node layer is r; then the calculation formula of the neural network model output is:
  • a 0 is a constant
  • the merchant credit evaluation value y based on the transaction data is calculated.
  • the y value is normalized such that the y value is in the range of 0.2 to 0.8.
  • the credit evaluation value of the evaluation result is between 0 and 1.
  • the present invention has the following beneficial effects: the present invention is based on historical transaction big data accumulated by merchants, commodities, and customers on an e-commerce platform, and the risk of the transaction is evaluated by an artificial intelligence method, and is used for evaluating transactions with lower risks.
  • the third-party platform advance payment mode will settle the funds to the merchant's domestic account in advance, thus avoiding the need for the regular clearing mode to wait for the customer to confirm the completion of the transaction, the e-commerce platform account period is completed, the e-commerce settlement to the third-party platform, and the third-party platform
  • the cumbersome process of bank clearing to the domestic accounts of merchants under the premise of controlling the risk of advance payment, simplifying the settlement process of cross-border funds of merchants, improving the efficiency of clearing settlements of third-party payments, shortening the time for merchants to settle accounts, improving the efficiency of merchants' capital turnover, and reducing
  • Figure 1 is a schematic diagram of the process of clearing settlement of conventional cross-border e-commerce funds and processing timeliness;
  • FIG. 2 is a schematic diagram of a cross-border merchant settlement and settlement process of the present invention
  • Figure 3 is a flow chart of a working principle of the present invention.
  • Figure 4 is a structural diagram of a credit evaluation system of the present invention.
  • FIG. 5 is a diagram of a transaction credit evaluation model based on an artificial neural network of the present invention.
  • the embodiment shown in FIG. 3 and FIG. 4 is a fast settlement method for cross-border payment based on credit big data, including a third-party payment platform, a cross-border e-commerce platform, a cooperative bank, a data interface module, a data standardization module, and an intelligent method.
  • Credit evaluation module and process control module including the following steps:
  • the system when the system detects that an order for an e-commerce platform of the corresponding merchant has been generated and the customer has completed the payment, the system will initiate a credit evaluation of the corresponding fund of the transaction, and a transaction involves the merchant (seller), the commodity, the customer. (Buyer)
  • the information used includes:
  • the data in Table 1 has different expressions and cannot be directly applied to the credit evaluation model. Standardization of data is required. Standardization is also referred to as pre-processing and normalization processing in the input data processing described in the present invention, due to BP.
  • the hidden layer of the neural network generally adopts the Sigmoid conversion function. In order to improve the training speed and sensitivity and effectively avoid the saturation region of the Sigmoid function, the value of the input data is generally required to be between 0 and 1. Therefore, the input data is preprocessed. Further, in order to ensure that the established model has a certain extrapolation ability, it is preferable to make the data pre-processed value between 0.2 and 0.8.
  • variable normalization process of the present invention will be described below by taking different types of variables as an example.
  • the merchant registration date represents the operation time of the merchant on this e-commerce platform and the third-party payment platform to a certain extent, and is also an important variable to characterize the merchant's credit.
  • the segment registration method is used to register the merchant registration date. Standardize.
  • an artificial neural network model is used to evaluate merchant credit based on transaction data.
  • the artificial neural network is an artificially constructed function capable of realizing certain functions based on human understanding of the brain neural network.
  • the model does not require the modeler to have an accurate understanding of the internal mechanism of the modeled object.
  • the network itself has strong fault tolerance and association functions.
  • Figure 5 shows the schematic diagram of the merchant credit evaluation model based on transaction data of the present invention.
  • the x 1 to x 10 in Table 1 is used as the model input, and the merchant credit evaluation value Y is used as the model output.
  • the artificial neural network model of merchant credit evaluation based on transaction data is constructed, and the three-layer network structure (ie, input node layer and hidden node layer) is adopted. , output node layer).
  • a batch of learning samples is selected to train the neural network to make it have the ability of credit evaluation.
  • the learning samples contain good learning examples with good credit and poor credit.
  • the input samples of the model are standardized after historical transaction data.
  • the value, the model output sample is the Y value obtained based on the fund performance status (account period, bad debt number, return quantity, etc.) in the historical transaction, and is standardized within the range of 0.2 to 0.8.
  • the standardized sample data is sent to the neural network for learning, the neural network is trained by the training data of the data sample set, the verification data of the data sample set is used to test the learning success of the neural network, and the test data of the data sample set is used to test whether the neural network has Good generalization ability of credit evaluation.
  • the training will be interrupted and the model parameters will be output. If the credit evaluation ability of the neural network does not meet expectations, the training will continue until the model meets the requirements.
  • the credit evaluation neural network model is trained using a standard backpropagation BP algorithm.
  • the standardized transaction data is input into the credit evaluation model of the trained neural network, and the credit evaluation of the related transaction can be obtained.
  • the transfer function of the hidden layer node uses the Sigmoid function (represented by f(x)), and the conversion function of the output node uses a linear function. .
  • the connection weight is v j
  • the node offset coefficient of the hidden node layer is ⁇ j
  • the node offset coefficient of the output node layer is r; then the calculation formula of the neural network model output is:
  • a 0 is a constant
  • the merchant credit evaluation value y based on the transaction data is calculated.
  • the system considers that based on the merchant credit evaluation, the risk of advance advance payment of the transaction is small, and the transaction settlement payment settlement process of Figure 2 is used to process the transaction funds in advance, and the funds are 0 to 1 day in advance. You can reach the domestic account of the merchant.
  • the system considers that based on the merchant credit evaluation, the risk of advance payment of this transaction is large, then the routine clearing process of Figure 1 is adopted, and the funds need to take 6 to 23 days to reach the domestic account of the merchant. .
  • the corresponding order funds of the merchant A go ahead of the advance payment process, and the corresponding order funds of the merchant B go through the normal process.
  • the artificial neural network model and the rule base based decision method are adopted to evaluate the credit value of the merchant's funds according to the transaction big data.
  • the credit evaluation method based on the transaction big data in the invention Common algorithms such as support vector machine and logistic regression method can also be used. Similar methods and systems implemented by such algorithms should also be within the scope of this patent.

Landscapes

  • Business, Economics & Management (AREA)
  • Accounting & Taxation (AREA)
  • Engineering & Computer Science (AREA)
  • Finance (AREA)
  • Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Security & Cryptography (AREA)
  • Development Economics (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Technology Law (AREA)
  • Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)

Abstract

本发明公开了一种基于信用大数据的跨境支付快速结算方法，本发明基于商户、商品、客户在电子商务平台所积累的历史交易大数据，对该交易的风险采用人工智能方法进行评估，对于评估风险较低的交易，采用第三方平台提前垫付的模式提前将资金结算至商户的国内账户，从而避免了常规清结算模式需要等待客户确认交易完成、电商平台账期完成、电商结算至第三方平台、第三方平台通过银行清算至商户国内账户的繁琐过程，达到在控制垫付风险的前提下，简化商户跨境资金清结算流程、提高第三方支付清结算效率、缩短商户资金到账时间、提高商户资金周转效率、降低常规结算周期过长带来的汇率变动风险、降低商户经营风险的有益效果。

Description

基于信用大数据的跨境支付快速结算方法 技术领域
本发明涉及跨境支付技术领域，尤其是涉及一种可提高商户资金流转效率、缩短资金到账时间、降低资金清结算流程汇率变动风险的基于信用大数据的跨境支付快速结算方法。
背景技术
近年来，我国进出口增速趋缓。海关统计数据显示，2010年，我国传统贸易增长率为34.7％，2011年下降为22.5％，至2012年，我国进出口额为38667.6亿美元，同比增长6.2％，2013年进出口额为41603亿美元，同比增长7.6％。出口业务也逐渐出现减速。在传统外贸出口增长乏力之际，同期我国跨境电子商务发展势头却呈现迅猛增长的势头，一跃成为我国外贸新的增长点，成为国际贸易的新方式和新手段，改变着传统国际贸易格局。
国家发改委的数据显示，2011年跨境电商交易额达到1.6万亿元，同比增长33％；2012年跨境交易额2万亿元，同比增长超过25％，增速远高于外贸增速，承担了我国外贸转型升级的重要任务。从跨境电商进、出口结构分布情况来看，2012年超过90％的交易规模由出口电商贸易贡献。
跨境电商的高速发展，需要强有力的跨境支付手段作为基础支撑。电子支付作为电子商务的重要组成之一，主要是指电子商务支付平台通过采用规范的连接器，在网上商家和银行之间建立起连接，从 而解决从消费者到金融机构、商家现金的在线货币支付、现金流转、资金清算、查询统计等问题。
跨境支付指两个或以上国家者地区之间因际贸易、国际投资及其他方面所发生的国际间债券务借助一定结算工具和支付系统实现资金跨和跨地区转移的行为。如中国消费者在网上购买外商家产品或品时，由于币种的不一样就需要通过定结算工具和支付系统实现两个国家或地区之间的资金转换，最终完成交易。对出口跨境电商来说，支付表现为外卡收单业务，即国内的商家通过外贸平台将商品销售给境外消费者，消费者付款后，由支付机构为国内商家收取外币并代理结汇。
目前，我国跨境电商企业在跨境商品交易中主要有五种不同的商业模式：传统跨境交易咨讯平台模式、门户型B2B综合平台模式、综合性垂直跨境小额平台(B2C、C2C、B2B)模式、第三方服务平台(代运营)模式和垂直型跨境小额交易网站(独立B2C)模式。
跨境电子商务的业务模式不同，采用的支付结算方式也存在差异。跨境电子支付业务会涉及资金结售汇与收付汇。从支付资金的流向来看，跨境电商进口业务(包括个人消费者海淘)涉及到跨境支付购汇，购汇途径一般有第三方购汇支付、境外电商接受人民币支付、通过国内银行购汇汇出等。跨境电商出口业务涉及到跨境收入结汇，其结汇途径主要包括第三方收结汇、通过国内银行汇款，以结汇个人名义拆分结汇流入等。相较于银行较高的费率和专业汇款公司的有限覆盖网点，第三方支付平台能同时满足用户对跨境汇款便捷性和低费 率的要求，因此，受到越来越多的跨境商户的青睐。然而，出于合规性的要求，国内第三方支付多数是通过和银行合作，将货币兑换和付款流程由其托管银行完成，由于受到电子商务平台账期、银行结售汇审核管理流程的约束，支付流程复杂，效率低，账期长，如图1所示，一般需要6到23个工作日才能收到货款。
由于商户接到订单后需要立刻发货，而相应的货款经过多家机构的多个结算环节，需要等待平台、银行和第三方支付的复杂审核与监管流程完成后才能到账，如此繁杂的流程使得资金流转缓慢，并且每个环节都要承担一定的费用，增加了企业经营成本；同时，漫长低效的资金清结算过程对跨境商户造成了极大的压力，带来了一定的汇率风险，严重影响了资金的使用效率，阻碍了商户扩大经营规模，增加了跨境商户的运营风险。
发明内容
本发明的发明目的是为了克服现有技术中的跨境电子商务支付常规流程环节繁多、资金账期长、过程风险大的不足，提供了一种可提高商户资金流转效率、缩短资金到账时间、降低资金清结算流程的汇率变动风险的基于信用大数据的跨境支付快速结算方法。
为了实现上述目的，本发明采用以下技术方案：
一种基于信用大数据的跨境支付快速结算方法，包括第三方支付平台、若干个电子商务平台、数据接口模块，数据标准化模块，智能信用评价模块和流程控制模块；
(1-1)第三方支付平台接到商户关于某笔交易的提现申请后，调 用数据接口模块从交易发生的电子商务平台获得智能信用评价模块所需要的大数据信息；
(1-2)调用数据标准化模块对数据进行清洗预处理和归一化，得到能够被智能信用评价模块接受的标准化数据；
(1-3)将标准化后的交易数据送入智能信用评价模块进行信用评价结果计算，将计算后的评价结果送入流程控制模块；
(1-4)若评价结果高于预先设定的阈值，则该笔交易进行提前垫付的风险很小，第三方支付平台进入提前垫付流程，第三方支付平台进行合规审核后通知合作银行进行分发，将交易款提前打入商户在国内的银行账户；
(1-5)若评价结果低于预先设定的阈值，则该笔交易进行提前垫付存在较大的风险，第三方支付平台按照正常流程进行处理，即在电子商务平台账期结束后，由电子商务平台结算至客户在第三方支付平台开设的海外账户，第三方支付平台收到货款后，进行合规审核后，经合作银行将合规资金分发至商户在国内的银行账户，完成整个流程。
数据接口模块，负责交易筛选、字段裁剪、数据加载，是本发明系统与电商平台及第三方支付平台之间的数据交换接口，用于从电商平台获取历史交易数据(商户30天内日均交易量，商户三个月总交易量，商户注册时长，交易商品的评价及退货率，交易商品所属类型，交易商品的平均确认时间，客户的平均确认时间，客户对所有购买商品的历史评价，客户的退货率等信息)，上述数据用于智能信用评价 模块进行信用评价模型的训练与商户订单资金是否提前结算的信用评价，同时，待智能信用评价模块完成商户订单资金的信用评价之后，通过数据接口模块下发至第三方支付平台，通知第三方支付平台提前将信用评价高的商户交易资金提前垫付，结算至商户的银行账户中。
数据标准化模块，将上述数据接口模块获得的交易数据转换为统一的，能够被智能信用评价模块识别的格式。标准化在本发明所描述的输入数据处理中也被称为预处理和归一化处理。由于交易数据的种类，取值范围差别很大，而常用的基于大数据的信用评价数学模型一般要求输入数据的值在0～1之间，因此，要对输入数据进行预处理。此外，为保证建立的模型具有一定的外推能力，最好使数据预处理后的值在0.2～0.8之间。
智能信用评价模块，基于标准化后的交易数据，计算相关的信用评价值，决定是否对该笔交易的资金进行提前垫付。基于交易大数据的商户信用评价方法有很多种，常有的包括规则库分类、人工神经网络、决策树等，该模块的输入值为标准化后的交易数据，输出为商户交易的信用评价值(0～1之间)，分值越高，代表该商户该笔交易的可信度越高，第三方支付系统提前垫付货款的风险越小。
流程控制模块基于智能信用评价的结果对商户货款清结算流程进行控制，若智能信用评价的结果高于某一事先给定阈值(例如0.6)，则系统判定该笔交易提前垫付的风险很小，对应的资金可以由第三方平台提前垫付，则系统进入提前垫付处理流程；若智能信用评价的结果低于某一事先给定阈值(例如0.6)，则系统判定该笔交易提前垫 付存在风险，则系统仍然按照正常流程进行处理，即等待电子商务平台账期过后再进行处理。
若评价结果高于系统预先设定的阈值，则该笔交易进行提前垫付的风险很小，系统通知第三方支付平台进入提前垫付流程，第三方支付平台进行合规审核后通知合作银行进行分发，将交易款提前打入商户在国内的银行账户，如图2所示，整个流程仅需0到1个工作日就能完成客户资金的清结算到账。若评价结果低于系统预先设定的阈值，则该笔交易进行提前垫付存在较大的风险，系统通知第三方支付平台按照正常流程进行处理，即在电子商务平台账期结束后(交易及电子商务平台处理流程，一般是3～17个工作日，视配送速度、用户确认速度，平台处理时效而定)，由电子商务平台结算至客户在第三方支付平台开设的海外账户(3～5个工作日)，第三方支付平台收到货款后，进行合规审核后，经合作银行将合规资金分发至商户在国内的银行账户，完成整个流程，大约共需要6到23个工作日才能完成客户资金的清结算到账。
本发明基于平台多个维度的历史交易数据(商户30天内日均交易量，商户三个月总交易量，商户注册时长，交易商品的评价及退货率，交易商品所属类型，交易商品的平均确认时间，客户的平均确认时间，客户对所有购买商品的历史评价，客户的退货率等信息)，对当前交易订单进行多维综合评价，得到基于平台交易大数据的客户信用评价，对于信用评价较高的商户，简化其资金的跨境清算流程，系统得到跨境电商平台客户购买货品并付款的确认信息后，即提前由第 三方支付平台对商户相应货款进行先行垫付，并由第三方平台完成交易审核并由合作银行完成合规审核后结汇分发至商户银行账户，整个过程可缩短至0到1个工作日完成，极大的缩短了商户的资金到账时间，提高了资金周转效率，降低了商户资金成本和汇率、运营风险。
本发明专利针对跨境商户收款流程复杂、成本居高不下、账期长的问题，通过对平台积累的庞大用户、商户交易信息进行数据挖掘和分析，采用基于交易大数据的信用度评价方法，对高信用度的交易缩短清结算流程，进行提前放款，在控制风险的同时，达到简化跨境清结算流程、减少商户资金成本、大大缩短资金到账时间的效果。
作为优选，所述历史交易数据包括商户30天内日均交易量，商户三个月总交易量，商户注册时长，交易商品的评价及退货率，交易商品所属类型，交易商品的平均确认时间，客户的平均确认时间，客户对所有购买商品的历史评价，客户的退货率。
作为优选，商户信用评价方法包括规则库分类方法、人工神经网络方法和决策树方法。
作为优选，人工神经网络方法包括如下步骤：
(4-1)基于神经网络的信用评估模型训练；
(4-2)基于神经网络的信用评估。
作为优选，步骤(4-1)包括如下步骤：
挑选一批学习样本组成数据样本集合，数据样本集合包含了信用良好和信用较差的学习样本，每个学习样本均包括训练数据、验证数据和测试数据，利用数据样本集合对神经网络模型进行训练，使神经 网络模型具备信用评价的能力；
将标准化后的历史交易数据作为神经网络模型的输入样本，送到神经网络当中学习，用数据样本集合的训练数据训练神经网络，数据样本集合的验证数据检验神经网络的学习成功性，并用数据样本集合的测试数据测试神经网络是否具有信用评价的良好泛化能力；
若神经网络模型的信用评估能力未达到预期，则继续训练直到模型满足要求为止；
若神经网络模型已经具备良好的信用评估能力，则中断训练，神经网络模型输出根据历史交易中的资金履约情况得到的Y值。
作为优选，步骤(4-2)包括如下步骤：
设神经网络模型包括输入节点层、隐节点层和输出节点层，三层节点分别为：ui(i＝1，2，…，m)，oj(j＝1，2，…k)，y；第i个输入节点ui与第j个隐节点oj间的连接权值为wij，输入节点层的节点偏置系数为qj；第j个隐节点oj与输出节点y间的连接权值为vj，隐节点层的节点偏置系数为θj，输出节点层的节点偏置系数为r；则神经网络模型输出的计算公式为：
Figure PCTCN2017000351-appb-000001
Figure PCTCN2017000351-appb-000002
其中，
Figure PCTCN2017000351-appb-000003
a0为常数；
基于公式(2)和(3)，计算得到基于交易数据的商户信用评估值y。
作为优选，对y值进行标准化，使y值在0.2至0.8的范围内。
作为优选，评价结果的信用评价值在0至1之间，分值越高，代表该商户该笔交易的可信度越高，第三方支付系统提前垫付货款的风险越小。
因此，本发明具有如下有益效果：本发明基于商户、商品、客户在电子商务平台所积累的历史交易大数据，对该交易的风险采用人工智能方法进行评估，对于评估风险较低的交易，采用第三方平台提前垫付的模式提前将资金结算至商户的国内账户，从而避免了常规清结算模式需要等待客户确认交易完成、电商平台账期完成、电商结算至第三方平台、第三方平台通过银行清算至商户国内账户的繁琐过程，达到在控制垫付风险的前提下，简化商户跨境资金清结算流程、提高第三方支付清结算效率、缩短商户资金到账时间、提高商户资金周转效率、降低常规结算周期过长带来的汇率变动风险、降低商户经营风险的有益效果。
附图说明
图1是常规跨境电商资金清结算过程及处理时效的一种原理图；
图2是本发明的一种跨境商户资金清结算过程及处理时效原理图；
图3是本发明的一种工作原理流程图；
图4本发明的一种信用评价系统结构图；
图5是本发明的一种基于人工神经网络的交易信用评价模型图。
具体实施方式
下面结合附图和具体实施方式对本发明做进一步的描述。
如图3、图4所示的实施例是一种基于信用大数据的跨境支付快速结算方法，包括第三方支付平台、跨境电商平台、合作银行、数据接口模块、数据标准化模块、智能信用评价模块和流程控制模块；包括如下步骤：
(一)获取交易信息数据
首先，当系统监测到对应商户某电子商务平台的一笔订单已生成且客户已完成付款时，系统会启动该笔交易对应资金的信用评估，一笔交易涉及到商户(卖家)、商品、客户(买家)三方，本发明中基于上述三个方面的信息对资金快速垫付的风险进行评估，采用数据接口模块对所需信息进行采集，所用到的信息包括：
表1交易数据信息类别
Figure PCTCN2017000351-appb-000004
(二)变量标准化
表1中的数据具有不同的表达方式，不能直接应用于信用评估模型，需要进行数据的标准化，标准化在本发明所描述的输入数据处理中也被称为预处理和归一化处理，由于BP神经网络的隐层一般采用Sigmoid转换函数，为提高训练速度和灵敏性以及有效避开Sigmoid函数的饱和区，一般要求输入数据的值在0～1之间。因此，要对输入数据进行预处理。进一步，为保证建立的模型具有一定的外推能力，最好使数据预处理后的值在0.2～0.8之间。
下面以不同类型的变量为例，介绍本发明的变量标准化过程。
(1)离散变量
以商品类型为例，根据不同商品大类的退换货频率，得出预先给定的商品所属类型的标准化输入，见表2所示：
表2商品类型离散值标准化
Figure PCTCN2017000351-appb-000005
(2)日期变量
商户注册日期在一定程度上代表了商户在此电子商务平台和第三方支付平台的运营时间，也是表征商户信用度的一个重要变量，本例中，采用分段赋值的方式对商户注册日期这一变量进行标准化。
Figure PCTCN2017000351-appb-000006
Figure PCTCN2017000351-appb-000007
(3)连续变量
连续变量采用以下公式将其映射到0.2～0.8的数值区间：
Figure PCTCN2017000351-appb-000008
以商品评价为例，若一个商品的历史评价为4.5分，评分最高为5分，最低为1分，则该商品评价经标准化之后的值为0.2+0.6*(4.5-1)/4＝0.725。
基于以上三种方式，实现了表1中的数据的标准化。
(三)基于交易数据的商户信用评估
本实施例中，采用人工神经网络模型对基于交易数据的商户信用进行评估，人工神经网络(Artificial Neural Network)是在人类对自身大脑神经网络认识理解的基础上人工构造的能够实现某种功能的模型，不需要建模者对建模对象的内在机理有精确了解，通过输入、输出数据训练神经网络的参数(包含连接权值和偏置系数)，就能使神经网络模型准确地反映实际的过程，它具有以下一些优点：
①通过非线性映射，学习系统的特性具有近似地表示任意非线性函数及其逆的能力；
②人工神经网络是由许多相同的简单处理单元采用并行分布处理结构组合而成，具有强大的信息处理能力。
③网络自身具有很强的容错性和联想功能。
基于上述优点，人工神经网络模型近年来在工程领域得到了广泛 的成功应用，图5给出了本发明基于交易数据的商户信用评估模型的原理图。即将表1中的x1到x10作为模型输入，商户信用评估值Y作为模型输出，构建基于交易数据的商户信用评估人工神经网络模型，采用三层网络结构(即输入节点层、隐节点层，输出节点层)。
基于神经网络的信用评估模型训练
首先挑选一批学习样本对神经网络进行训练，以使其具备信用评价的能力，该批学习样本中，包含了信用良好和信用较差的学习范例，模型的输入样本为历史交易数据标准化后的值，模型输出样本为根据历史交易中的资金履约情况(账期，坏账数，退换货数等)得到的Y值，并将其在0.2～0.8的范围内标准化。将标准化后的样本数据送到神经网络当中学习，用数据样本集合的训练数据训练神经网络，数据样本集合的验证数据检验神经网络的学习成功性，并用数据样本集合的测试数据测试神经网络是否具有信用评价的良好泛化能力，若训练所得的神经网络已经具备良好的信用评估能力，则中断训练，输出模型参数，若神经网络的信用评估能力未达到预期，则继续训练直到模型满足要求为止。本发明中，采用标准的反向传播BP算法对信用评估神经网络模型进行训练。
(1)基于神经网络的信用评估
随后，将标准化后的交易数据输入训练好的神经网络的信用评估模型，即可得到相关交易的信用评估。以如图5中所示的采用三层(带有一个隐层)的神经网络为例，隐层节点的转换函数采用Sigmoid函数(用f(x)表示)，输出节点的转换函数采用线性函数。
设神经网络模型包括输入节点层、隐节点层和输出节点层，三层节点分别为：ui(i＝1，2，…，m)，oj(j＝1，2，…k)，y；第i个输入节点ui与第j个隐节点oj间的连接权值为wij，输入节点层的节点偏置系数为qj；第j个隐节点oj与输出节点y间的连接权值为vj，隐节点层的节点偏置系数为θj，输出节点层的节点偏置系数为r；则神经网络模型输出的计算公式为：
Figure PCTCN2017000351-appb-000009
Figure PCTCN2017000351-appb-000010
其中，
Figure PCTCN2017000351-appb-000011
a0为常数；
基于公式(2)和(3)，计算得到基于交易数据的商户信用评估值y。
(四)基于商户信用评估的交易资金提前垫付
最后，基于商户信用评估y，我们可以根据提前设定的阈值决定是否提前进行资金垫付，即：
如果y≥0.6，系统提前垫付，合作银行分发该笔交易资金至商户国内账户；
如果y＜0.6，提前垫付有风险，拒绝提前垫付，进入常规支付流程。
若信用评估值大于等于阈值0.6，则系统认为基于商户信用评估，此笔交易提前垫付的风险较小，则采用图2的交易垫付清结算流程，对交易资金进行提前处理，0～1天资金即可到达商户的国内账户。
若信用评估值小于阈值0.6，则系统认为基于商户信用评估，此笔交易提前垫付的风险较大，则采用图1的常规清结算流程，资金需要长达6～23天才能到达商户的国内账户。
基于规则库的商户信用评估实施例：
(1)若(X1≥3万美金∪X2≥250万美金)∩X3≥24个月∩X4∈[箱包，消费电子，家居，图书，家用电器，化妆品，文具，鞋靴，食品]，同意提前垫付；
(2)若(3万美金＞X1≥2万美金∪250万美金＞X2≥150万美金)∩24个月＞X3≥12个月∩X4∈[箱包，消费电子，家居，图书，家用电器，化妆品，文具]∩X5≤15∩X6≥4∩X7≤0.01，同意提前垫付；
(3)若(2万美金＞X1≥1万美金∪150万美金＞X2≥80万美金)∩12个月＞X3≥6个月∩X4∈[箱包，消费电子，家居，图书，家用电器，化妆品，文具]∩X5≤8∩X6≥4∩X7≤0.005，同意提前垫付；
(4)若(1万美金＞X1≥0.5万美金∪80万美金＞X2≥50万美金)∩6个月＞X3≥3个月∩X4∈[家用电器，图书，文具]∩X5≤5∩X6≥4.5∩X7≤0.003∩X8≤10∩X9≥4.5∩X10≤0.01，同意提前垫付；
(5)其他情况，拒绝提前垫付，走常规流程。
如有以下两单交易，
例1：
商户A，
X1＝5万美金，X2＝350万美金，X3＝30个月，X4＝家用电器，X5＝5，X6＝4.9，X7＝0.0003，X8＝8，X9＝4.8，X10＝0.001，
例2：
商户B，
X1＝1.5万美金，X2＝100万美金，X3＝8个月，X4＝化妆品，X5＝10，X6＝3.8，X7＝0.02，X8＝12，X9＝4.3，X10＝0.03
根据上述规则，商户A相应订单资金走提前垫付流程，商户B的相应订单资金走常规流程。
本实施例中，为了说明的方便起见，采用了给人工神经网络模型和基于规则库的决策方法依据交易大数据对商户资金信用度进行评价，实际上，发明中的基于交易大数据的信用评价方法 也可以采用支持向量机、Logistic回归方法等常用算法，采用此类算法所实现的类似方法和系统也应在本专利保护范围之内。

Claims (8)

  1. 一种基于信用大数据的跨境支付快速结算方法，其特征是，包括第三方支付平台、若干个电子商务平台、数据接口模块，数据标准化模块，智能信用评价模块和流程控制模块；
    (1-1)第三方支付平台接到商户关于某笔交易的提现申请后，调用数据接口模块从交易发生的电子商务平台获得智能信用评价模块所需要的大数据信息；
    (1-2)调用数据标准化模块对数据进行清洗预处理和归一化，得到能够被智能信用评价模块接受的标准化数据；
    (1-3)将标准化后的交易数据送入智能信用评价模块进行信用评价结果计算，将计算后的评价结果送入流程控制模块；
    (1-4)若评价结果高于预先设定的阈值，则该笔交易进行提前垫付的风险很小，第三方支付平台进入提前垫付流程，第三方支付平台进行合规审核后通知合作银行进行分发，将交易款提前打入商户在国内的银行账户；
    (1-5)若评价结果低于预先设定的阈值，则该笔交易进行提前垫付存在较大的风险，第三方支付平台按照正常流程进行处理，即在电子商务平台账期结束后，由电子商务平台结算至客户在第三方支付平台开设的海外账户，第三方支付平台收到货款后，进行合规审核后，经合作银行将合规资金分发至商户在国内的银行账户，完成整个流程。
  2. 根据权利要求1所述的基于信用大数据的跨境支付快速结算方法，其特征是，所述历史交易数据包括商户30天内日均交易量， 商户三个月总交易量，商户注册时长，交易商品的评价及退货率，交易商品所属类型，交易商品的平均确认时间，客户的平均确认时间，客户对所有购买商品的历史评价，客户的退货率。
  3. 根据权利要求1所述的基于信用大数据的跨境支付快速结算方法，其特征是，商户信用评价方法包括规则库分类方法、人工神经网络方法和决策树方法。
  4. 根据权利要求1所述的基于信用大数据的跨境支付快速结算方法，其特征是，人工神经网络方法包括如下步骤：
    (4-1)基于神经网络的信用评估模型训练；
    (4-2)基于神经网络的信用评估。
  5. 根据权利要求4所述的基于信用大数据的跨境支付快速结算方法，其特征是，步骤(4-1)包括如下步骤：
    挑选一批学习样本组成数据样本集合，数据样本集合包含了信用良好和信用较差的学习样本，每个学习样本均包括训练数据、验证数据和测试数据，利用数据样本集合对神经网络模型进行训练，使神经网络模型具备信用评价的能力；
    将标准化后的历史交易数据作为神经网络模型的输入样本，送到神经网络当中学习，用数据样本集合的训练数据训练神经网络，数据样本集合的验证数据检验神经网络的学习成功性，并用数据样本集合的测试数据测试神经网络是否具有信用评价的良好泛化能力；
    若神经网络模型的信用评估能力未达到预期，则继续训练直到模型满足要求为止；
    若神经网络模型已经具备良好的信用评估能力，则中断训练，神经网络模型输出根据历史交易中的资金履约情况得到的Y值。
  6. 根据权利要求4所述的基于信用大数据的跨境支付快速结算方法，其特征是，步骤(4-2)包括如下步骤：
    设神经网络模型包括输入节点层、隐节点层和输出节点层，三层节点分别为：ui(i＝1，2，…，m)，oj(j＝1，2，…k)，y；第i个输入节点ui与第j个隐节点oj间的连接权值为wij，输入节点层的节点偏置系数为qj；第j个隐节点oj与输出节点y间的连接权值为vj，隐节点层的节点偏置系数为θj，输出节点层的节点偏置系数为r；则神经网络模型输出的计算公式为：
    Figure PCTCN2017000351-appb-100001
    Figure PCTCN2017000351-appb-100002
    其中，
    Figure PCTCN2017000351-appb-100003
    a0为常数；
    基于公式(2)和(3)，计算得到基于交易数据的商户信用评估值y。
  7. 根据权利要求5所述的基于信用大数据的跨境支付快速结算方法，其特征是，对y值进行标准化，使y值在0.2至0.8的范围内。
  8. 根据权利要求1或2或3所述的基于信用大数据的跨境支付快速结算方法，其特征是，评价结果的信用评价值在0至1之间，分值越高，代表该商户该笔交易的可信度越高，第三方支付系统提前垫付货款的风险越小。
PCT/CN2017/000351 2017-03-27 2017-05-12 基于信用大数据的跨境支付快速结算方法 WO2018176179A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710192684.3A CN107103460A (zh) 2017-03-27 2017-03-27 基于信用大数据的跨境支付快速结算方法
CN201710192684.3 2017-03-27

Publications (1)

Publication Number Publication Date
WO2018176179A1 true WO2018176179A1 (zh) 2018-10-04

Family

ID=59675756

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/000351 WO2018176179A1 (zh) 2017-03-27 2017-05-12 基于信用大数据的跨境支付快速结算方法

Country Status (2)

Country Link
CN (1) CN107103460A (zh)
WO (1) WO2018176179A1 (zh)

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109858894A (zh) * 2019-01-16 2019-06-07 深圳壹账通智能科技有限公司 一种支付结果通知方法、装置、可读存储介质及服务器
CN110956543A (zh) * 2019-11-06 2020-04-03 上海应用技术大学 异常交易检测的方法
CN111311356A (zh) * 2020-01-20 2020-06-19 广西东信金服信息技术有限公司 一种用于跨境互市贸易背景真实性及唯一性的校验系统及方法
CN112862298A (zh) * 2020-07-09 2021-05-28 北京睿知图远科技有限公司 一种针对用户画像的信用评估方法
CN113537731A (zh) * 2021-06-25 2021-10-22 中国海洋大学 基于强化学习的设计资源能力评估方法
CN113743677A (zh) * 2021-09-16 2021-12-03 成都数融科技有限公司 基于联邦学习的个人信用评价模型训练方法及评价方法
CN113888147A (zh) * 2021-10-13 2022-01-04 重庆葵海数字科技有限公司 关于服务电商平台第三方监管的方法
CN113988890A (zh) * 2021-11-08 2022-01-28 大路网络科技有限公司 一种跨境电商企业评估方法、装置及设备
CN116187766A (zh) * 2023-04-24 2023-05-30 北京信联数安科技有限公司 基于动态风险评估的数据出境安全合规管控方法及系统
EP4246406A1 (en) * 2022-03-17 2023-09-20 Affirm, Inc. Method and apparatus for facilitating merchant cash advances on the basis of forecasted return volume
CN117592979A (zh) * 2024-01-18 2024-02-23 快捷通支付服务有限公司 一种用于收单结算的方法、设备及介质

Families Citing this family (25)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20200327549A1 (en) * 2017-11-08 2020-10-15 Paypal, Inc. Robust and Adaptive Artificial Intelligence Modeling
CN108932622B (zh) * 2017-12-08 2021-07-30 翟红鹰 在线交易的风险控制方法、系统及计算机可读存储介质
CN108764862A (zh) * 2018-05-17 2018-11-06 阿里巴巴集团控股有限公司 支付数据处理方法及装置
CN109345048B (zh) * 2018-07-27 2023-02-10 创新先进技术有限公司 预测方法、装置、电子设备及计算机可读存储介质
CN108681957B (zh) * 2018-08-13 2021-04-23 杭州乒乓智能技术股份有限公司 跨境电商交易订单还原方法及系统及设备及存储介质
CN109034915B (zh) * 2018-09-03 2022-04-22 王昆 一种可应用数字资产或积分作为交易媒介的人工智能型电子商务系统
CN109493067A (zh) * 2018-09-29 2019-03-19 浙江口碑网络技术有限公司 针对餐饮订单的支付方法及装置，存储介质和电子设备
CN109636406B (zh) * 2018-10-17 2023-09-08 创新先进技术有限公司 资源结算方法及装置和电子设备
CN109727016A (zh) * 2018-12-17 2019-05-07 泰康保险集团股份有限公司 信用支付处理方法、装置、介质及电子设备
CN111489147A (zh) * 2019-01-25 2020-08-04 中国银联股份有限公司 资格确认的方法、支付系统和支付方法
CN110175914A (zh) * 2019-04-17 2019-08-27 苏宁易购集团股份有限公司 网络交易处理方法及装置
CN110490694B (zh) * 2019-07-30 2022-04-29 创新先进技术有限公司 一种信用服务的配置方法、装置及电子设备
CN112561520B (zh) * 2019-09-10 2023-09-08 腾讯科技（深圳）有限公司 支付数据处理方法、装置、设备及存储介质
CN110910109B (zh) * 2019-11-25 2022-03-25 腾讯科技（深圳）有限公司 基于区块链的电子资源转移方法及装置
CN111028075B (zh) * 2019-12-12 2024-05-14 腾讯科技（深圳）有限公司 虚拟资源转移方法、装置及设备
CN111353877A (zh) * 2020-03-31 2020-06-30 中国建设银行股份有限公司 两级跨境交易系统及方法
CN111967779B (zh) * 2020-08-19 2022-12-16 支付宝(杭州)信息技术有限公司 一种风险评估方法、装置及设备
CN112037038B (zh) * 2020-09-02 2024-05-28 中国银行股份有限公司 银行信贷风险预测方法及装置
CN112541802A (zh) * 2020-12-09 2021-03-23 安徽理工大学 一种电子商务交易系统
CN113298512A (zh) * 2020-12-29 2021-08-24 阿里巴巴集团控股有限公司 资金数据信息处理方法、装置及电子设备
CN112837045A (zh) * 2021-01-30 2021-05-25 银盛通信有限公司 一种基于互联网的移动第三方互联网支付系统
CN117709955A (zh) * 2021-07-09 2024-03-15 支付宝(杭州)信息技术有限公司 基于信用的交易处理方法及装置
CN114708090A (zh) * 2022-06-06 2022-07-05 广州番禺职业技术学院 基于大数据的银行支付业务风险识别装置
CN116611898B (zh) * 2023-07-20 2023-09-22 南京可码软件科技有限公司 一种基于电商平台的线上支付优化系统及方法
CN117422306A (zh) * 2023-10-30 2024-01-19 广州金财智链数字科技有限公司 基于动态神经网络的跨境电商风险控制方法及系统

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101937541A (zh) * 2009-06-30 2011-01-05 商文彬 一种用于评价客户信用度的方法及设备
CN103413242A (zh) * 2013-07-24 2013-11-27 通联支付网络服务股份有限公司 一种非面对面交易中小企业信誉及融资信用标准系统
CN103839183A (zh) * 2014-03-19 2014-06-04 江苏苏大大数据科技有限公司 智能授信方法及智能授信装置
CN105894089A (zh) * 2016-04-21 2016-08-24 百度在线网络技术（北京）有限公司 一种征信模型的建立方法、征信确定方法及对应装置

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1912909A (zh) * 2005-08-10 2007-02-14 蒋妍妍 电子商务诚信交易方法
CN102214337A (zh) * 2010-04-07 2011-10-12 上海钧宸投资有限公司 商品交易结算方法和系统
CN102982479A (zh) * 2012-12-21 2013-03-20 江苏乐买到网络科技有限公司 一种网络购物中退款的方法
CN103295156A (zh) * 2013-01-17 2013-09-11 厦门蓝象网络科技有限公司 一种网络借贷平台
CN103310333A (zh) * 2013-03-22 2013-09-18 涂志敏 一种线下终端刷卡实现资金划至第三方支付平台的方法
CN104408610A (zh) * 2014-12-03 2015-03-11 苏州贝多环保技术有限公司 一种基于风险评估的第三方支付平台业务处理方法
CN105989462A (zh) * 2015-02-02 2016-10-05 卓望数码技术（深圳）有限公司 一种支持不同支付渠道的结算方法及结算系统
CN106469377A (zh) * 2015-08-21 2017-03-01 阿里巴巴集团控股有限公司 退货信息处理方法及装置
CN105809430A (zh) * 2016-03-03 2016-07-27 中南林业科技大学 一种林业用电子商务交易控制系统和控制方法
CN106296175A (zh) * 2016-08-12 2017-01-04 中国银行股份有限公司 支付方法、商户终端、银行交易系统、客户终端及系统

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101937541A (zh) * 2009-06-30 2011-01-05 商文彬 一种用于评价客户信用度的方法及设备
CN103413242A (zh) * 2013-07-24 2013-11-27 通联支付网络服务股份有限公司 一种非面对面交易中小企业信誉及融资信用标准系统
CN103839183A (zh) * 2014-03-19 2014-06-04 江苏苏大大数据科技有限公司 智能授信方法及智能授信装置
CN105894089A (zh) * 2016-04-21 2016-08-24 百度在线网络技术（北京）有限公司 一种征信模型的建立方法、征信确定方法及对应装置

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
CHEN, JING: "Alibaba : Technology Connects the World", ECONOMIC DAILY, vol. 13, no. 11566, 18 March 2015 (2015-03-18) *

Cited By (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109858894A (zh) * 2019-01-16 2019-06-07 深圳壹账通智能科技有限公司 一种支付结果通知方法、装置、可读存储介质及服务器
CN110956543A (zh) * 2019-11-06 2020-04-03 上海应用技术大学 异常交易检测的方法
CN111311356B (zh) * 2020-01-20 2023-10-20 广西联合征信有限公司 一种用于跨境互市贸易背景真实性及唯一性的校验系统及方法
CN111311356A (zh) * 2020-01-20 2020-06-19 广西东信金服信息技术有限公司 一种用于跨境互市贸易背景真实性及唯一性的校验系统及方法
CN112862298A (zh) * 2020-07-09 2021-05-28 北京睿知图远科技有限公司 一种针对用户画像的信用评估方法
CN112862298B (zh) * 2020-07-09 2024-02-27 北京睿知图远科技有限公司 一种针对用户画像的信用评估方法
CN113537731B (zh) * 2021-06-25 2023-10-27 中国海洋大学 基于强化学习的设计资源能力评估方法
CN113537731A (zh) * 2021-06-25 2021-10-22 中国海洋大学 基于强化学习的设计资源能力评估方法
CN113743677B (zh) * 2021-09-16 2023-06-30 成都数融科技有限公司 基于联邦学习的个人信用评价模型训练方法及评价方法
CN113743677A (zh) * 2021-09-16 2021-12-03 成都数融科技有限公司 基于联邦学习的个人信用评价模型训练方法及评价方法
CN113888147A (zh) * 2021-10-13 2022-01-04 重庆葵海数字科技有限公司 关于服务电商平台第三方监管的方法
CN113988890A (zh) * 2021-11-08 2022-01-28 大路网络科技有限公司 一种跨境电商企业评估方法、装置及设备
EP4246406A1 (en) * 2022-03-17 2023-09-20 Affirm, Inc. Method and apparatus for facilitating merchant cash advances on the basis of forecasted return volume
CN116187766A (zh) * 2023-04-24 2023-05-30 北京信联数安科技有限公司 基于动态风险评估的数据出境安全合规管控方法及系统
CN117592979A (zh) * 2024-01-18 2024-02-23 快捷通支付服务有限公司 一种用于收单结算的方法、设备及介质
CN117592979B (zh) * 2024-01-18 2024-03-19 快捷通支付服务有限公司 一种用于收单结算的方法、设备及介质

Also Published As

Publication number Publication date
CN107103460A (zh) 2017-08-29

Similar Documents

Publication Publication Date Title
WO2018176179A1 (zh) 基于信用大数据的跨境支付快速结算方法
US7296734B2 (en) Systems and methods for scoring bank customers direct deposit account transaction activity to match financial behavior to specific acquisition, performance and risk events defined by the bank using a decision tree and stochastic process
US6216115B1 (en) Method for multi-directional consumer purchasing, selling, and transaction management
Afaha Electronic payment systems (E-payments) and Nigeria economic growth
CN106529923A (zh) 一种一点接入全网通的收款服务方法
CN110415097A (zh) 一种汇票拆分、合并系统及实现汇票交易的方法
Agyemang et al. The role of mobile money banking service in financial development: evidence from Ghana
KR20170037299A (ko) 통합계좌정보 기반의 실시간 신용리스크 분석 방법 및 시스템
TWM613536U (zh) 對於基金商品的投資風險評分系統
Clarke et al. Methodology| RDP 2012-05: Payment System Design and Participant Operational Disruptions
CN102915508A (zh) 基于交易记录的信用贷款风险控制的系统及方法
Asimakopoulos et al. The efficiency of the Greek banking system and its determinants
KR20110049226A (ko) 대용량 처리 전용 데이터베이스를 이용한 신용위험도 산출 시스템
KR101137305B1 (ko) 온라인 증권계좌 담보대출 시스템 및 이를 이용한 온라인 증권계좌 담보대출 거래방법
Vovchak et al. Bank business modeling and levels of non-performing loans: Perspectives of international risk factors in Ukraine
CN108537547A (zh) 一种基于互联网的多市场证券账号在线货币兑换方法
CN108985924A (zh) 用于校园债务清偿的管理系统
Mulyani et al. The effect of e-money on the money supply and inflation di Indonesia year 2019-2021
Chen The crucial role of the five-year Treasury in the US yield curve
TWI776370B (zh) 對於基金商品的投資風險評分方法及系統
CN117422576B (zh) 基于云计算的商品销售的业务财务一体化管理系统及方法
KR20170004934A (ko) 통합계좌정보 기반의 실시간 신용리스크 분석 방법 및 시스템
Tan et al. Comparison of traditional payment and blockchain payment methods in the same payment point
Zhang et al. ANN-Based Integrated Multi-Objective Investment Decision Model
KR20170037298A (ko) 통합계좌정보 기반의 실시간 신용리스크 분석 방법 및 시스템

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17903068

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 17903068

Country of ref document: EP

Kind code of ref document: A1