WO2020048056A1 - 一种风险决策方法和装置 - Google Patents

一种风险决策方法和装置 Download PDF

Info

Publication number
WO2020048056A1
WO2020048056A1 PCT/CN2018/123887 CN2018123887W WO2020048056A1 WO 2020048056 A1 WO2020048056 A1 WO 2020048056A1 CN 2018123887 W CN2018123887 W CN 2018123887W WO 2020048056 A1 WO2020048056 A1 WO 2020048056A1
Authority
WO
WIPO (PCT)
Prior art keywords
risk
user
related data
standard
rule engine
Prior art date
Application number
PCT/CN2018/123887
Other languages
English (en)
French (fr)
Inventor
徐竑
张轶
何艳茹
Original Assignee
平安科技(深圳)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 平安科技(深圳)有限公司 filed Critical 平安科技(深圳)有限公司
Publication of WO2020048056A1 publication Critical patent/WO2020048056A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION 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
    • G06Q10/00Administration; Management
    • G06Q10/06Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063Operations research, analysis or management
    • G06Q10/0635Risk analysis of enterprise or organisation activities
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION 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
    • G06Q40/00Finance; Insurance; Tax strategies; Processing of corporate or income taxes

Definitions

  • the present application relates to the field of data processing, and in particular, to a risk decision method and device.
  • Risk control means that the risk manager adopts various measures and methods to eliminate or reduce the various possibilities of risk events, or the risk controller reduces the losses caused when risk events occur.
  • Internet finance many transaction payments, loans, guarantees, or other funding activities are involved.
  • risk control or anti-fraud responses it is necessary to perform risk ratings on users and make related risks. decision making.
  • the third-party risk decision service is an important service for the financial industry. It can provide decision-making references for companies to make capital loans or fund guarantees, which is vital to fund security.
  • Traditional risk decision-making services have problems such as non-localized deployment, no visual reports, no closed loops in the risk control process, and more importantly, the data for risk assessment is single and the accuracy of the risk assessment conclusions is low, and the reference value is low.
  • the embodiments of the present application provide a risk decision method and device, which can influence the generation of risk decisions through standardization of risk-related data, rule engine processing, and risk rating.
  • the parameters and rule engine adopted improve the comprehensiveness and standardization of risk ratings. Degree, and through risk rating, the accuracy and referability of risk decision-making have been improved.
  • the first aspect of the embodiments of the present application provides a risk decision method.
  • the risk decision method includes:
  • the rule engine includes a rule set that performs numerical constraints and / or conflict checking on the standard parameters; and imports the standard parameters into the Rule engine to get risk rating;
  • a second aspect of the embodiments of the present application provides a risk decision device, where the risk decision device includes:
  • An obtaining unit configured to monitor an electronic device of a user and obtain risk-related data in the electronic device
  • a standard processing unit configured to parameterize the risk-related data to obtain standard parameters applicable to a rule engine, where the rule engine includes a rule set that performs numerical constraints and / or conflict checking on the standard parameters;
  • An import unit configured to import the standard parameters into the rule engine to obtain a risk rating
  • a decision unit configured to make a risk decision according to the risk rating.
  • a third aspect of the embodiments of the present application provides an electronic device including a processor, a memory, a communication interface, and one or more programs.
  • the one or more programs are stored in the memory and configured by The processor executes, and the program includes instructions for performing steps in any method of the first aspect.
  • a fourth aspect of the embodiments of the present application provides a computer-readable storage medium storing a computer program for electronic data exchange, wherein the computer program causes a computer to execute instructions of the steps described in any of the methods of the first aspect.
  • risk-related data through the obtained risk-related data, user risks can be more comprehensively examined, the dimension of parameterized processing of risk-related data is enriched, and the parameterized processing of risk data is imported into the rule engine and obtained Risk rating improves the standardization of risk ratings. Finally, risk decisions are made through risk ratings, which improves the accuracy and referability of risk decisions. It has great reference value for risk control in the financial process.
  • FIG. 1 is a schematic flowchart of a risk decision method according to an embodiment of the present application
  • FIG. 2 is a schematic flowchart of another risk decision method according to an embodiment of the present application.
  • FIG. 3 is a schematic flowchart of another risk decision method according to an embodiment of the present application.
  • FIG. 4 is a schematic flowchart of another risk decision method provided by an embodiment of the present application.
  • FIG. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
  • FIG. 6 is a structural block diagram of a risk decision device provided by an embodiment of the present application.
  • FIG. 1 is a schematic flowchart of a risk decision method according to an embodiment of the present application. As shown in FIG. 1, the risk decision method includes:
  • Risk control means that the risk manager adopts various measures and methods to eliminate or reduce the various possibilities of risk events, or the risk controller reduces the losses caused when risk events occur.
  • Internet finance many transaction payments, loans, guarantees, or other funding activities are involved.
  • risk related data including device hardware data, device software data, related data when the device is connected to the network, and User account information and user-related data.
  • the risk control platform can provide users with services through applications, browsers, pop-up windows, or web pages, and obtain data through interaction with users to complete the purpose of monitoring user electronic devices and obtaining risk-related data.
  • the risk-related data includes device-related data
  • obtaining the risk-related data in the electronic device includes: obtaining hardware parameters of the device, including the device's physical address or unique serial number of the device; obtaining device network data, including the network to which the device is connected Name or Internet Protocol address; determine device-related data based on hardware parameters and network data.
  • acquiring device-related data is data that the device itself has and cannot be easily changed. For example, device hardware name, model, shape, color, function, physical address (MAC Address), central processing unit (CPU, Central Processing Unit) model, etc., or the Internet Protocol address (IP Address) obtained when the device is connected to the network, Wireless network (Fidelity, WIFI) name or Global Positioning System (GPS) positioning.
  • MAC Address physical address
  • CPU central processing unit
  • IP Address Internet Protocol address
  • WIFI Wireless network
  • GPS Global Positioning System
  • the risk-related data includes user-related data
  • obtaining the risk-related data in the electronic device includes: obtaining user account information, including account name or account interaction information; obtaining user activity information, including web page address of the user browsing , Operation rules, or time consuming; determine user related data according to user account information and user activity information.
  • User-related data can be divided into two types.
  • One type is static information such as user account information, including user account numbers, profiles, friend lists, communication objects, communication records, or bound bank card numbers on shopping or social networking sites.
  • the other type is dynamic information such as user activity information, including the web page address, browsing order, length of stay, frequency of clicks, or content of clicks. Account information and user activity information can get user activity-related data.
  • the expressions of the obtained series of risk-related data are various.
  • the obtained equipment hardware name, model, shape, color, and function may be English strings, Chinese strings, Chinese and English symbols combined with strings, numbers, etc.
  • the user's activity information may be a number, a URL, or a moment. Then, these data need to be parameterized in order to import the rule engine and obtain a risk rating.
  • parameterize the risk-related data to obtain standard parameters applicable to the rule engine, including: generating a device fingerprint based on the risk-related data; matching the device fingerprint with the standard fingerprint of the device, and The matching result obtains standard parameters, wherein a parameter one is obtained when the device fingerprint matches the standard fingerprint successfully, a parameter two is obtained when the device fingerprint fails to match the standard fingerprint, and the parameter one is The two parameters mentioned above are standard parameters.
  • a device fingerprint can be generated based on risk-related data to uniquely authenticate the device.
  • Device-related data can be obtained according to a preset time interval. During this period of time, the user-device-related data does not change much or contains a certain change rule. After the obtained device-related data and change rule, a device fingerprint is generated as a standard fingerprint. Then, when judging the security of the current device, obtain the current device-related data, generate a device fingerprint, and match it with the standard fingerprint. If it cannot be completely matched, or the degree of matching is less than a preset threshold, then the device fingerprint can be determined If the match fails, set the current device fingerprint parameter 2 to 0, otherwise, set the current device fingerprint parameter 1 to 1.
  • the generation of the set fingerprint can be obtained according to the device MAC address and the application installed on the device for a long time.
  • the MAC address is composed of 20 hexadecimal characters, and the application has a User Identification (UID).
  • UID User Identification
  • each application will obtain a segment of characters.
  • the first half is the MAC address and the second half is the UID to form a longer string; or the two are inserted and connected, such as inserting the UID into the middle of the MAC address; or The two are cross-connected.
  • two UID characters are inserted after two MAC addresses, and two UID characters are inserted after three MAC addresses.
  • the generated string is used as plain text, and the two are inserted.
  • the method generates a secret key.
  • the obtained device MAC address and UID need to generate a new plaintext based on the secret key, which is compared with the plaintext of the standard fingerprint.
  • device-related data and user-related data are obtained as risk-related data, and then a device fingerprint is generated based on the risk-related data, and then the generated device fingerprint is matched with a previously generated standard fingerprint, and according to the matching result
  • this process can effectively standardize equipment risk data, make the generated standard parameters effectively indicate the risk of the equipment, and improve the accuracy and reference value of risk judgment.
  • parameterize the risk-related data to obtain standard parameters applicable to the rule engine, including: obtaining multiple characteristic values of different risk operations in the risk-related data; and clustering user-related data based on the multiple characteristic values
  • feature values that do not contain numerical values are used for keyword clustering, and feature values that contain numerical values are clustered for numerical ranges; weights are assigned to multiple user categories, and multiple user categories and Its corresponding weight.
  • user-related data including the account information and user activity information of the aforementioned users, including the account name or account interaction information, as well as the web page address, operation rule, or time-consuming time of the user.
  • user-related data you must first determine the feature values used for clustering. For example, select the web page address, user operation frequency, and time spent on the page as feature values to classify users. Then, according to the classification results, weights are assigned to users in different categories. This process is shown in Table 1:
  • Table 1 User classification table obtained based on risk-related data
  • clustering in addition to clustering according to each type of data, all data can be clustered uniformly to obtain corresponding categories and assign different weights to different categories.
  • parameterize the risk-related data to obtain standard parameters applicable to the rule engine, including: generating a user relationship map based on device-related data and user-related data; and obtaining user and relationship threatening user threats in the relationship map based on the user relationship map Correlation coefficient; determine the user's threat coefficient based on the correlation coefficient.
  • a relational map is a graph that describes individuals and their relationships.
  • the user device's IP address, wifi name, etc. can be determined.
  • the user's account, phone number, address book, call or chat history, etc. can be determined.
  • Each dimension of data can generate a user relationship map. .
  • the address book relationship can form a directed relationship graph.
  • the contact list of user A contains the contact method of user B, and there is a path from A to B. It is also possible to generate a common relationship graph based on data from all dimensions, for example, the wifi names of user C and user D are the same, and there is a call record between them, then there are two paths between C and D.
  • Threat users refer to unsafe users, who may have initiated fraudulent behaviors, or may have records of dishonesty of repayment.
  • the user finally judges the threat coefficient of the unknown user based on the number of paths between the unknown user and the threat user. Among them, each path
  • the weights can be the same or different. By evaluating the importance of the paths between users, different weights are set for the daily paths. For example, if there is a call record between two users, the path is highly important. You can set a larger weight, such as 10. However, if two users have viewed the same webpage, the path is of low importance. Small weight, such as 0.5.
  • a user relationship map is generated based on risk-related data, and then the threat coefficient of the current user is determined according to the correlation coefficient between the current user and the threat user.
  • the interpretation of social relevance has enriched the dimension of parameterized processing of risk-related data, and improved the accuracy and reference value of risk judgments.
  • the rule engine is a service that parses, invokes, and executes rule packages.
  • the rule engine can accept data input, interpret business rules, and make business decisions based on business rules.
  • the rule engine includes rules that perform numerical constraints and / or conflict checks on standard parameters. set. The data has been parameterized according to step 102, and then the parameters are imported into the rule engine to obtain the final threat value, and the risk rating is determined according to the threat value.
  • importing standard parameters into the rule engine to obtain a risk rating includes: importing one or more of the obtained standard parameters into the rule engine to obtain a rule score.
  • the rule engine includes weighted processing, binarization processing, or condition selection; according to the rules Scoring to give users a risk rating.
  • the rule engine contains a rule set composed of various rules, which is used to perform numerical constraints or conflict checks on various types of parameters obtained, so that all parameters can be uniformly used for risk rating. Assume that the rules contained in the rule engine are: sum of all parameter values. If the sum of all parameter values is less than or equal to 1, the risk is low. If it is greater than 1 and less than 1.5, then the risk is greater than or equal to 1.5. The degree of risk is high.
  • the rule engine is shown in Table 1:
  • the risk rating has been obtained according to step 103, then the server must make a corresponding risk decision according to the obtained risk rating. For example, when the risk level is low, the server can mark the user as a concerned user; if the risk level is within the range, the user can be appropriately reduced. The amount of arrears or payments, and mark the user as a potential risk user, and further check the user; if the risk is high, you can reject the user ’s loan application and mark the user as a threat user.
  • the user risk can be more comprehensively examined through the acquired risk-related data, which enriches the dimension of parameterized processing of risk-related data.
  • After parameterizing the risk data it is imported into the rule engine and Obtaining a risk rating improves the degree of risk rating standardization.
  • the risk decision is made through the risk rating, which improves the accuracy and referability of the risk decision, and has great reference value for risk control in the financial process.
  • FIG. 2 is a schematic flowchart of another risk decision method provided by an embodiment of the present application. As shown in the figure, the risk decision method in this embodiment includes:
  • Obtain device network data including a network name or an Internet Protocol address to which the device is connected;
  • FIG. 3 is a schematic flowchart of another risk decision method provided by an embodiment of the present application. As shown in the figure, the risk decision method in this embodiment includes:
  • Obtain user account information including account name or account interaction information
  • Obtain user activity information including a webpage address browsed by the user, an operation rule, or a time-consuming operation.
  • weights to the multiple user classifications obtain the multiple user classifications and their corresponding weights, and use the weights as standard parameters applicable to the rule engine;
  • FIG. 4 is a schematic flowchart of another risk decision method provided by an embodiment of the present application. As shown in the figure, the risk decision method in this embodiment includes:
  • the user's electronic device is monitored, the risk-related data in the electronic device is obtained, and a user relationship map is generated according to the risk-related data, and then the user's threat coefficient is obtained according to the user relationship map to complete the risk Relevant data is parameterized so that the obtained standard parameters are used in subsequent rule engines to obtain a risk rating, and finally a risk decision is made based on the risk rating.
  • This process enriches the dimension of parameterized processing of risk-related data, improves the standardization of risk ratings, and finally makes risk decisions through risk ratings, which improves the accuracy and referability of risk decisions.
  • FIG. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
  • the electronic device includes a processor, a memory, a communication interface, and one or more programs.
  • the one or more programs are Stored in the memory and configured to be executed by the processor, the program includes instructions for performing the following steps:
  • the electronic device in the embodiment of the present application can more comprehensively examine user risks through the acquired risk-related data, enriching the dimension of risk-related data parameterization, and after parameterizing the risk data,
  • the introduction of the rule engine and the risk rating have improved the degree of risk rating standardization.
  • the risk decision is made through the risk rating, which improves the accuracy and referability of the risk decision, and has great reference value for risk control in the financial process.
  • FIG. 6 is a block diagram of functional units of a risk decision device 600 according to an embodiment of the present application.
  • the risk decision device 600 is applied to an electronic device, and the risk decision device includes:
  • An obtaining unit 601 configured to monitor an electronic device of a user, and obtain risk-related data in the electronic device
  • a standard processing unit 602 configured to parameterize the risk-related data to obtain standard parameters applicable to a rule engine, where the rule engine includes a rule set that performs numerical constraints and / or conflict checking on the standard parameters;
  • An import unit 603, configured to import the standard parameters into the rule engine to obtain a risk rating
  • a decision unit 604 is configured to make a risk decision according to the risk rating.
  • the risk decision device in the embodiment of the present application can more comprehensively examine user risks through the acquired risk-related data, enriching the dimension of risk-related data parameterization, and parameterizing risk data. Later, the rule engine was imported and risk ratings were obtained, which improved the standardization of risk ratings. Finally, risk decisions were made through risk ratings, which improved the accuracy and referability of risk decisions, and has a great reference for risk control in the financial process. value.
  • the risk-related data includes device-related data
  • the obtaining unit 601 is specifically configured to: obtain hardware parameters of the device, including the physical address of the device or a unique serial number of the device; obtain network data of the device, including the device connection A network name or an internet protocol address; determining the device-related data according to the hardware parameters and network data.
  • the risk-related data includes user-related data
  • the obtaining unit 601 is specifically configured to: obtain account information of the user, including account name or account interaction information; and obtain activity information of the user, including The webpage address browsed by the user, the operation rule, or the time-consuming process; determining the user-related data according to the user's account information and the user's activity information.
  • the standard processing unit 602 is specifically configured to: generate a device fingerprint according to the risk-related data; match the device fingerprint with a standard fingerprint of the device, and obtain standard parameters according to the matching result, where A parameter one is obtained when the device fingerprint matches the standard fingerprint successfully; a parameter two is obtained when the device fingerprint fails to match the standard fingerprint; both the parameter one and the parameter two are applicable to Standard parameters of the rules engine.
  • the standard processing unit 602 is specifically configured to: obtain multiple feature values of different risk operations in the risk-related data; and cluster the user-related data according to the multiple feature values to obtain A plurality of user classifications, in which feature values that do not include numerical values are subjected to keyword clustering and feature values that include numerical values are clustered in numerical ranges; weights are assigned to the plurality of user classifications to obtain the plurality of users Classify and their corresponding weights, and use the weights as standard parameters applicable to the rule engine.
  • the standard processing unit 602 is specifically configured to: generate a user relationship graph according to the device-related data and user-related data; and obtain an association coefficient between the user and a threat user in the relationship graph according to the user relationship graph Determining the threat coefficient of the user according to the correlation coefficient, and the threat coefficient is a standard parameter applicable to the rule engine.
  • the importing unit 603 is specifically configured to: import the obtained one or more standard parameters into a rule engine to obtain a rule score, and the rule engine includes weighting processing, binarization processing, or condition selection;
  • the rule score is a risk rating performed by the user. The higher the rule score, the higher the level of the risk rating.
  • An embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program for electronic data exchange, and the computer program causes a computer to execute any method described in the foregoing method embodiments. Part or all of the steps, the computer includes a mobile terminal.
  • An embodiment of the present application further provides a computer program product.
  • the computer program product includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute any one of the methods described in the foregoing method embodiments. Part or all of the steps of a method.
  • the computer program product may be a software installation package, and the computer includes a mobile terminal.
  • the program may be stored in a computer-readable memory, and the memory may include a flash disk. , ROM, RAM, disk or disc, etc.

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Theoretical Computer Science (AREA)
  • Development Economics (AREA)
  • Marketing (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Technology Law (AREA)
  • Finance (AREA)
  • Educational Administration (AREA)
  • Accounting & Taxation (AREA)
  • Game Theory and Decision Science (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

一种风险决策方法和装置,风险决策方法包括:监测用户的电子设备,获取电子设备中的风险相关数据(101);对风险相关数据进行参数化处理,获得适用于规则引擎的标准参数,规则引擎包括对标准参数进行数值约束和/或冲突检验的规则集合(102);将标准参数导入规则引擎,获得风险评级(103);根据风险评级做出风险决策(104)。通过对风险相关数据的标准化,规则引擎处理和风险评级,影响风险决策的生成,采用的参数和规则引擎提升了风险评级的全面性和规范程度,进而提升了风险决策的准确性和可参考性。

Description

一种风险决策方法和装置
本申请要求于2018年9月6日提交中国专利局、申请号为2018110378170、申请名称为“一种风险决策方法和装置”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及数据处理领域,具体涉及一种风险决策方法和装置。
背景技术
风险控制是指风险管理者采取各种措施和方法,消灭或减少风险事件发生的各种可能性,或风险控制者减少风险事件发生时造成的损失。在互联网金融中,涉及到很多交易支付、贷款、担保或其他资金活动,为了识别资金交易中的风险,并提前做出风险控制或反欺诈应对,需要对用户进行风险评级,并做出相关风险决策。
第三方的风险决策服务是金融行业的重要服务,能够为企业做出资金贷款或资金担保时提供决策参考,对资金安全至关重要。传统风险决策服务存在不可本地化部署、无可视化报表、风控流程无闭环等等问题,更重要的是,进行风险评估的数据单一,得出的风险评估结论准确性低,可参考价值低。
发明内容
本申请实施例提供一种风险决策方法和装置,能够通过对风险相关数据的标准化,规则引擎处理和风险评级,影响风险决策的生成,采用的参数和规则引擎提升了风险评级的全面性和规范程度,且通过风险评级,提升了风险决策的准确性和可参考性。
本申请实施例的第一方面提供了一种风险决策方法,所述风险决策方法包括:
监测用户的电子设备,获取所述电子设备中的风险相关数据;
对所述风险相关数据进行参数化处理,获得适用于规则引擎的标准参数,所述规则引擎包括对所述标准参数进行数值约束和/或冲突检验的规则集合;将所述标准参数导入所述规则引擎,获得风险评级;
根据所述风险评级做出风险决策。
本申请实施例的第二方面提供了一种风险决策装置,所述风险决策装置包 括:
获取单元,用于监测用户的电子设备,获取所述电子设备中的风险相关数据;
标准处理单元,用于对所述风险相关数据进行参数化处理,获得适用于规则引擎的标准参数,所述规则引擎包括对所述标准参数进行数值约束和/或冲突检验的规则集合;
导入单元,用于将所述标准参数导入所述规则引擎,获得风险评级;
决策单元,用于根据所述风险评级做出风险决策。
本申请实施例第三方面提供了一种电子装置,包括处理器、存储器、通信接口,以及一个或多个程序,所述一个或多个程序被存储在所述存储器中,并且被配置由所述处理器执行,所述程序包括用于执行第一方面任一方法中的步骤的指令。
本申请实施例第四方面提供了一种计算机可读存储介质,存储用于电子数据交换的计算机程序,其中,所述计算机程序使得计算机执行第一方面任一方法所述的步骤的指令。
在本申请实施例中,通过获取到的风险相关数据,能够更全面地对用户风险进行考察,丰富了风险相关数据参数化处理的维度,将风险数据进行参数化处理后,导入规则引擎并获得风险评级,提升了风险评级规范程度,最后通过风险评级做出风险决策,提升了风险决策的准确性和可参考性,对于金融过程中的风险控制具有很大的参考价值。
附图说明
为了更清楚地说明本申请实施例或现有技术中的技术方案,下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍。
图1是本申请实施例提供的一种风险决策方法流程示意图;
图2是本申请实施例提供的另一种风险决策方法流程示意图;
图3是本申请实施例提供的另一种风险决策方法的流程示意图;
图4是本申请实施例提供的另一种风险决策方法的流程示意图;
图5是本申请实施例提供的一种电子装置的结构示意图;
图6是本申请实施例提供的一种风险决策装置的结构框图。
具体实施方式
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。
下面对本申请实施例进行详细介绍。
请参阅图1,图1为本申请实施例中一种风险决策方法流程示意图,如图1所示,所述风险决策方法包括:
101、监测用户的电子设备,获取所述电子设备中的风险相关数据。
风险控制是指风险管理者采取各种措施和方法,消灭或减少风险事件发生的各种可能性,或风险控制者减少风险事件发生时造成的损失。在互联网金融中,涉及到很多交易支付、贷款、担保或其他资金活动,为了识别资金交易中的风险,并提前做出风险控制或反欺诈应对,需要对用户进行风险评级,并做出相关风险决策。进一步地,因为用户的资金活动是通过终端设备来完成的,所以终端设备可获取或记录的相关数据都是风险相关数据,包括设备硬件数据,设备软件数据,设备连接网络时的相关数据,以及用户账户信息和用户相关数据等。风险控制平台可以通过应用程序、浏览器、弹窗或网页等形式,为用户提供服务,并通过与用户的互动获得数据,完成监测用户电子设备,获得风险相关数据的目的。
可选的,风险相关数据包括设备相关数据,获取所述电子设备中的风险相关数据包括:获取设备的硬件参数,包括设备物理地址或设备唯一序列号;获取设备网络数据,包括设备连接的网络名称或网际协议地址;根据硬件参数和网络数据,确定设备相关数据。
具体地,获取设备相关数据,是设备本身具有的,不容易更改的数据。例如设备硬件的名称、型号、形状、颜色、功能,物理地址(MAC Address),中央处理器(CPU,Central Processing Unit)型号等,或者设备连接网络时获取到的网际协议地址(IP Address)、无线网络(Wireless Fidelity,WIFI)名称或全球定位系统(Global Positioning System,GPS)定位等。获取到设备相关数据后,可以发现用户的设备是否进行了改装,或者常用的物理地址发生了改变, 或者WIFI名称与某个风险用户的WIFI名称相同等,这些参数都可以用于判断用户的设备具有风险。结合设备硬件参数和设备连接网络时的相关参数,即可确定设备相关的数据。
可选的,风险相关数据包括用户相关数据,获取所述电子设备中的风险相关数据包括:获取用户的账户信息,包括账户名称或账户交互信息;获取用户的活动信息,包括用户浏览的网页地址、操作规律或耗费时长等;根据用户的账户信息和用户的活动信息,确定用户相关数据。
具体地,用户在使用终端设备时,会有相关的操作,对应用户相关数据,记录并分析相关数据,可以得出用户的风险程度。例如用户在每个页面停留时间极短,页面与页面之间的切换时长小于1s(秒),这明显是不符合人的操作习惯和规律的,很有可能是机器进行的刷单操作。因此可根据该数据判断当前设备用户属于风险用户。用户相关数据可以分为两类,一类是用户的账户信息这类静态信息,包括用户在购物网站或社交网站的账号、简介、好友列表、沟通交流对象、沟通交流记录或绑定的银行卡号、手机号码,或者用户常用应用程序的名称、账户等,另一类是用户的活动信息这类动态信息,包括浏览的网页地址、浏览顺序、停留时长、点击频率或点击内容等,结合用户的账户信息和用户的活动信息,可得到用户的活动相关数据。
102、对所述风险相关数据进行参数化处理,获得适用于规则引擎的标准参数,所述规则引擎包括对所述标准参数进行数值约束和/或冲突检验的规则集合。
获得的一系列风险相关数据的表现形式是多种多样的,例如获得的设备硬件的名称、型号、形状、颜色、功能可能是英文字符串、中文字符串、中英文符号结合字符串、数字等,用户的活动信息,可能是一个数字、一个网址或一个时刻,那么,需要对这些数据进行参数化处理,才能导入规则引擎并获得风险评级。
可选的,对风险相关数据进行参数化处理,获得适用于规则引擎的标准参数,包括:根据所述风险相关数据生成设备指纹;将所述设备指纹与该设备的标准指纹进行匹配,并根据匹配结果,获得标准参数,其中,所述设备指纹与所述标准指纹匹配成功时获得参数一,所述设备指纹与所述标准指纹匹配失败 时获得参数二,所述参数一和所述参数二都为标准参数。
具体地,由于设备相关的数据是不容易更改的,因此,可以根据风险相关数据生成设备指纹,用来唯一认证设备。可以根据预设时间间隔获取设备相关数据,在该段时间内,用户设备相关数据变化不大,或者包含某种变化规律,获得的设备相关数据和变化规律后,生成设备指纹,作为标准指纹。那么,对于当前设备的安全性进行判断时,获取当前的设备相关数据,生成设备指纹,并与标准指纹进行匹配,如果不能完全匹配,或者匹配程度小于某个预设阈值,那么可判定设备指纹匹配失败,设置当前设备指纹参数二为0,否则,设置当前设备指纹参数一为1。
其中,对于设置指纹的生成,可以根据设备MAC地址和设备长期安装的应用程序获得,其中MAC地址是由20位的16进制字符组成,而应用程序具有的用户识别(User Identification,UID),是电子装置中安装应用程序时,每一个应用程序将所获取的一段字符。可以将MAC地址和UID直接进行首尾连接,如前半截为MAC地址,后半截为UID,组成一个更长的字符串;或者将两者进行插入连接,例如将UID插入到MAC地址中间;或者将两者进行交叉连接,例如两位MAC地址后插入一个UID字符,然后3位MAC地址后再插入两个UID字符,直到MAC地址和UID完全混合,生成的字符串作为明文,而两者的插入方法则生成一个秘钥,当对当前设备的安全性进行判定时,获取的设备MAC地址和UID需要根据秘钥生成新的明文,在于标准指纹的明文进行对比。
可见,在本申请实施例中,通过获取设备相关数据和用户相关数据作为风险相关数据,然后根据风险相关数据生成设备指纹,再将生成的设备指纹与之前生成的标准指纹进行匹配,根据匹配结果获得标准参数,这个过程能够有效对设备风险数据进行标准化处理,使得生成的标准参数有效指示设备的风险性,提升了风险判断的准确性和可参考价值。
可选的,对风险相关数据进行参数化处理,获得适用于规则引擎的标准参数,包括:获取风险相关数据中不同风险操作的多个特征值;根据多个特征值对用户相关数据进行聚类,获得多个用户分类,其中,将不包含数值的特征值进行关键字聚类,将包含数值的特征值进行数值范围的聚类;为多个用户分类 赋予权值,获得多个用户分类及其对应的权值。
用户相关数据的种类非常多,包括前述提到的用户的账户信息和用户的活动信息,具体又包括账户名称或账户交互信息,以及用户浏览的网页地址、操作规律或耗费时长等,如果要对用户相关数据进行聚类,首先要确定用于聚类的特征值,例如选取用户浏览的网页地址、用户操作频率和页面耗费时长作为特征值,对用户进行分类。然后根据分类结果,为不同类别的用户赋予权值。这个过程如表1所示:
表1 根据风险相关数据获取的用户分类表
Figure PCTCN2018123887-appb-000001
由表1可知,根据聚类条件将用户分类两类,一类为风险用户类,另一类为安全用户类,而不同的聚类条件得到不同的风险用户类,为不同的风险用户类赋予不同的权值,而对于所有的安全用户类,都赋予0的权值。如果同一个用户同时被划分为多个风险用户类,则其权值为多个风险用户类权值之和。
另外,在聚类时,除了按照每一类数据进行聚类外,还可以将所有数据进行统一聚类,获得对应类别,并为不同类别赋予不同权值。
可见,在本申请实施例中,通过根据用户风险数据中的风险操作相关数据对用户进行分类,然后对不同的用户分类赋予不同的权值,对于风险相关数据的参数化处理从用户操作相关数据方面进行诠释,丰富了风险相关数据参数化处理的维度,提升了风险判断的准确性和可参考价值。
可选的,对风险相关数据进行参数化处理,获得适用于规则引擎的标准参数,包括:根据设备相关数据和用户相关数据生成用户关系图谱;根据用户关 系图谱获取用户与关系图谱中威胁用户的关联系数;根据关联系数确定用户的威胁系数。
关系图谱是描述个体及个体之间关系的图。根据设备相关数据可确定用户设备的IP地址、wifi名称等,根据用户相关数据可确定用户的账户、电话号码、通讯录和通话或聊天记录等,每一个维度的数据都可以生成一个用户关系图谱。例如通讯录关系可形成一个有向关系图谱,用户A的通讯录中包含用户B的联系方式,则存在从A到B的路径。也可以根据所有维度的数据生成一个共同的关系图谱,例如用户C和用户D的wifi名称相同,且两者之间有通话记录,那么C和D之间存在两条路径。威胁用户是指不安全的用户,可能曾经发起过欺诈行为,也可能有还款不诚信记录,用户最后根据未知用户与威胁用户之间的路径数判断未知用户的威胁系数,其中,每条路径的权值可以相同,也可以不同,通过评估用户之间的路径重要程度,为每天路径设置不同的权值。例如两个用户之间有通话记录,该路径重要程度高,可以设置一个较大的权值,例如10,而两个用户之间浏览过同一个网页,该路径重要程度低,可以设置一个较小的权值,例如0.5。
可见,在本申请实施例中,通过风险相关数据生成用户关系图谱,然后根据关系图谱中,当前用户与威胁用户的关联系数,确定当前用户的威胁系数,对于风险相关数据的参数化处理从用户社交关联性方面进行诠释,丰富了风险相关数据参数化处理的维度,提升了风险判断的准确性和可参考价值。
103、将所述标准参数导入所述规则引擎,获得风险评级。
规则引擎是解析、调用、执行规则包的服务,规则引擎可以接受数据输入,解释业务规则,并根据业务规则做出业务决策,规则引擎中包括对标准参数进行数值约束和/或冲突检验的规则集合。根据步骤102已经将所述数据都进行了参数化,那么把参数导入规则引擎,则可获得最终威胁值,并根据威胁值确定风险评级。
可选的,将标准参数导入规则引擎,获得风险评级,包括:将获得的一个或多个标准参数导入规则引擎,获取规则评分,规则引擎包括加权处理、二值化处理或条件选择;根据规则评分,为用户进行风险评级。
规则引擎中包含各类规则组成的规则集合,用以对获得的各类参数进行数 值约束或冲突检验,使所有参数能够统一用于风险评级。假设规则引擎中包含的规则为:对所有参数值求和,若所有参数值的和小于或等于1,则风险程度低,若大于1且小于1.5,则风险程度中,若大于或等于1.5,则风险程度高。该规则引擎如表1所示:
表2 规则引擎表
Figure PCTCN2018123887-appb-000002
那么根据该规则引擎可获得用户的风险评级。例如已知用户S的设备指纹参数是1,用户聚类获得的参数值是0.5,关系图谱获得的威胁系数是0.2,那么可知用户S的规则评分为1+0.5+0.2=1.7≥1.5,那么可知用户S的风险评级为“风险程度高”。
104、根据所述风险评级做出风险决策。
根据步骤103已经得出风险评级,那么服务器要根据获得的风险评级做出对应的风险决策,例如风险程度低时,服务器可以标记该用户为关注用户;如果风险程度中,则可以适当降低该用户的欠款额度或支付额度,并标记该用户为潜在风险用户,还可以对该用户进行进一步的核查;如果风险程度高,则可以拒绝该用户的贷款申请,并将该用户标记为威胁用户。
可见,在发明实施例中,通过获取到的风险相关数据,能够更全面地对用户风险进行考察,丰富了风险相关数据参数化处理的维度,将风险数据进行参数化处理后,导入规则引擎并获得风险评级,提升了风险评级规范程度,最后通过风险评级做出风险决策,提升了风险决策的准确性和可参考性,对于金融过程中的风险控制具有很大的参考价值。
请参阅图2,图2是本申请实施例提供的另一种风险决策方法流程示意图,如图所示,本实施例中的风险决策方法包括:
201、获取设备的硬件参数,包括设备物理地址或设备唯一序列号;
202、获取设备网络数据,包括设备连接的网络名称或网际协议地址;
203、根据所述硬件参数和网络数据,确定所述设备相关数据,并将所述设备相关数据作为风险相关数据;
204、根据所述风险相关数据生成设备指纹;
205、将所述设备指纹与该设备的标准指纹进行匹配,并根据匹配结果,获得标准参数,其中,所述设备指纹与所述标准指纹匹配成功时获得参数一,所述设备指纹与所述标准指纹匹配失败时获得参数二,所述参数一和所述参数二都为适用于规则引擎的标准参数;
206、将所述标准参数导入所述规则引擎,获得风险评级;
207、根据所述风险评级做出风险决策。
可见,在本申请实施例中,通过获取设备相关数据作为风险相关数据,然后根据风险相关数据生成设备指纹,并根据设备指纹的匹配结果获得标准参数,完成对风险相关数据的参数化处理,使获得的标准参数用于后续规则引擎中,以获得风险评级,最终根据风险评级进行风险决策。这个过程丰富了风险相关数据参数化处理的维度,提升了风险评级规范程度,最后通过风险评级做出风险决策,提升了风险决策的准确性和可参考性。
请参阅图3,图3是本申请实施例提供的另一种风险决策方法的流程示意图,如图所示,本实施例中的风险决策方法包括:
301、获取用户的账户信息,包括账户名称或账户交互信息;
302、获取用户的活动信息,包括用户浏览的网页地址、操作规律或耗费时长;
303、根据所述用户的账户信息和所述用户的活动信息,确定所述用户相关数据,并将所述用户相关数据作为风险相关数据;
304、获取所述风险相关数据中不同风险操作的多个特征值;
305、根据所述多个特征值对所述用户相关数据进行聚类,获得多个用户分类,其中,将不包含数值的特征值进行关键字聚类,将包含数值的特征值进行数值范围的聚类;
306、为所述多个用户分类赋予权值,获得所述多个用户分类及其对应的 权值,并将所述权值作为适用于规则引擎的标准参数;
307、将所述标准参数导入所述规则引擎,获得风险评级;
308、根据所述风险评级做出风险决策。
可见,本申请实施例中,通过获取用户相关数据作为风险相关数据,然后根据风险相关数据进行聚类,获得多个用户分类,并对多个分类赋予不同的权值,完成对风险相关数据的参数化处理,使获得的标准参数用于后续规则引擎中,以获得风险评级,最终根据风险评级进行风险决策。这个过程丰富了风险相关数据参数化处理的维度,提升了风险评级规范程度,最后通过风险评级做出风险决策,提升了风险决策的准确性和可参考性。
请参阅图4,图4是本申请实施例提供的另一种风险决策方法的流程示意图,如图所示,本实施例中的风险决策方法包括:
401、监测用户的电子设备,获取所述电子设备中的风险相关数据;
402、根据所述风险相关数据生成用户关系图谱;
403、根据所述用户关系图谱获取所述用户与关系图谱中威胁用户的关联系数;
404、根据所述关联系数确定所述用户的威胁系数,所述威胁系数即为适用于规则引擎的标准参数;
405、将获得的一个或多个标准参数导入规则引擎,获取规则评分,所述规则引擎包括加权处理、二值化处理或条件选择;
406、根据所述规则评分为所述用户进行风险评级,所述规则评分越高,所述风险评级的等级越高;
407、根据所述风险评级做出风险决策。
可见,在本申请实施例中,监测用户的电子设备,获取所述电子设备中的风险相关数据,并根据风险相关数据生成用户关系图谱,然后根据用户关系图谱获得用户的威胁系数,完成对风险相关数据的参数化处理,使获得的标准参数用于后续规则引擎中,以获得风险评级,最终根据风险评级进行风险决策。这个过程丰富了风险相关数据参数化处理的维度,提升了风险评级规范程度,最后通过风险评级做出风险决策,提升了风险决策的准确性和可参考性。
图5是本申请实施例提供的一种电子装置的结构示意图,如图5所示,该电子装置包括处理器、存储器、通信接口以及一个或多个程序,其中,上述一个或多个程序被存储在上述存储器中,并且被配置由上述处理器执行,上述程序包括用于执行以下步骤的指令:
监测用户的电子设备,获取所述电子设备中的风险相关数据;
对所述风险相关数据进行参数化处理,获得适用于规则引擎的标准参数,所述规则引擎包括对所述标准参数进行数值约束和/或冲突检验的规则集合;
将所述标准参数导入所述规则引擎,获得风险评级;
根据所述风险评级做出风险决策。
可以看出,本申请实施例中的电子装置通过获取到的风险相关数据,能够更全面地对用户风险进行考察,丰富了风险相关数据参数化处理的维度,将风险数据进行参数化处理后,导入规则引擎并获得风险评级,提升了风险评级规范程度,最后通过风险评级做出风险决策,提升了风险决策的准确性和可参考性,对于金融过程中的风险控制具有很大的参考价值。
图6是本申请实施例中所涉及的风险决策装置600的功能单元组成框图。该风险决策装置600应用于电子装置,所述风险决策装置包括:
获取单元601,用于监测用户的电子设备,获取所述电子设备中的风险相关数据;
标准处理单元602,用于对所述风险相关数据进行参数化处理,获得适用于规则引擎的标准参数,所述规则引擎包括对所述标准参数进行数值约束和/或冲突检验的规则集合;
导入单元603,用于将所述标准参数导入所述规则引擎,获得风险评级;
决策单元604,用于根据所述风险评级做出风险决策。
可以看出,在本申请实施例中的风险决策装置通过获取到的风险相关数据,能够更全面地对用户风险进行考察,丰富了风险相关数据参数化处理的维度,将风险数据进行参数化处理后,导入规则引擎并获得风险评级,提升了风险评级规范程度,最后通过风险评级做出风险决策,提升了风险决策的准确性 和可参考性,对于金融过程中的风险控制具有很大的参考价值。
在可选情况下,所述风险相关数据包括设备相关数据,所述获取单元601具体用于:获取设备的硬件参数,包括设备物理地址或设备唯一序列号;获取设备网络数据,包括设备连接的网络名称或网际协议地址;根据所述硬件参数和网络数据,确定所述设备相关数据。
在可选情况下,所述风险相关数据包括用户相关数据,所述获取单元601具体用于:获取所述用户的账户信息,包括账户名称或账户交互信息;获取所述用户的活动信息,包括用户浏览的网页地址、操作规律或耗费时长;根据所述用户的账户信息和所述用户的活动信息,确定所述用户相关数据。
在可选情况下,所述标准处理单元602具体用于:根据所述风险相关数据生成设备指纹;将所述设备指纹与该设备的标准指纹进行匹配,并根据匹配结果,获得标准参数,其中,所述设备指纹与所述标准指纹匹配成功时获得参数一,所述设备指纹与所述标准指纹匹配失败时获得参数二,所述参数一和所述参数二都为适用于规则引擎的标准参数。
在可选情况下,所述标准处理单元602具体用于:获取所述风险相关数据中不同风险操作的多个特征值;根据所述多个特征值对所述用户相关数据进行聚类,获得多个用户分类,其中,将不包含数值的特征值进行关键字聚类,将包含数值的特征值进行数值范围的聚类;为所述多个用户分类赋予权值,获得所述多个用户分类及其对应的权值,并将所述权值作为适用于规则引擎的标准参数。
在可选情况下,所述标准处理单元602具体用于:根据所述设备相关数据和用户相关数据生成用户关系图谱;根据所述用户关系图谱获取所述用户与关系图谱中威胁用户的关联系数;根据所述关联系数确定所述用户的威胁系数,所述威胁系数即为适用于规则引擎的标准参数。
在可选情况下,所述导入单元603具体用于:将获得的一个或多个标准参数导入规则引擎,获取规则评分,所述规则引擎包括加权处理、二值化处理或条件选择;根据所述规则评分为所述用户进行风险评级,所述规则评分越高,所述风险评级的等级越高。
本申请实施例还提供一种计算机可读存储介质,其中,该计算机可读存储 介质存储用于电子数据交换的计算机程序,该计算机程序使得计算机执行如上述方法实施例中记载的任一方法的部分或全部步骤,上述计算机包括移动终端。
本申请实施例还提供一种计算机程序产品,上述计算机程序产品包括存储了计算机程序的非瞬时性计算机可读存储介质,上述计算机程序可操作来使计算机执行如上述方法实施例中记载的任一方法的部分或全部步骤。该计算机程序产品可以为一个软件安装包,上述计算机包括移动终端。
本领域普通技术人员可以理解上述实施例的各种方法中的全部或部分步骤是可以通过程序来指令相关的硬件来完成,该程序可以存储于一计算机可读存储器中,存储器可以包括:闪存盘、ROM、RAM、磁盘或光盘等。
以上对本申请实施例进行了详细介绍,本文中应用了具体个例对本申请的原理及实施方式进行了阐述,以上实施例的说明只是用于帮助理解本申请的方法及其核心思想;同时,对于本领域的一般技术人员,依据本申请的思想,在具体实施方式及应用范围上均会有改变之处,综上所述,本说明书内容不应理解为对本申请的限制。

Claims (20)

  1. 一种风险决策方法,其特征在于,所述方法包括:
    监测用户的电子设备,获取所述电子设备中的风险相关数据;
    对所述风险相关数据进行参数化处理,获得适用于规则引擎的标准参数,所述规则引擎包括对所述标准参数进行数值约束和/或冲突检验的规则集合;
    将所述标准参数导入所述规则引擎,获得风险评级;
    根据所述风险评级做出风险决策。
  2. 根据权利要求1所述的方法,其特征在于,所述风险相关数据包括设备相关数据,所述获取所述电子设备中的风险相关数据包括:
    获取设备的硬件参数,包括设备物理地址或设备唯一序列号;
    获取设备网络数据,包括设备连接的网络名称或网际协议地址;
    根据所述硬件参数和网络数据,确定所述设备相关数据。
  3. 根据权利要求1或2所述的方法,其特征在于,所述风险相关数据包括用户相关数据,所述获取所述电子设备中的风险相关数据包括:
    获取所述用户的账户信息,包括账户名称或账户交互信息;
    获取所述用户的活动信息,包括用户浏览的网页地址、操作规律或耗费时长;
    根据所述用户的账户信息和所述用户的活动信息,确定所述用户相关数据。
  4. 根据权利要求3所述的方法,其特征在于,所述对所述风险相关数据进行参数化处理,获得适用于规则引擎的标准参数,包括:
    根据所述风险相关数据生成设备指纹;
    将所述设备指纹与该设备的标准指纹进行匹配,并根据匹配结果,获得标准参数,其中,所述设备指纹与所述标准指纹匹配成功时获得参数一,所述设备指纹与所述标准指纹匹配失败时获得参数二,所述参数一和所述参数二都为适用于规则引擎的标准参数。
  5. 根据权利要求4所述的方法,其特征在于,所述对所述风险相关数据进行参数化处理,获得适用于规则引擎的标准参数,包括:
    获取所述风险相关数据中不同风险操作的多个特征值;
    根据所述多个特征值对所述用户相关数据进行聚类,获得多个用户分类,其中,将不包含数值的特征值进行关键字聚类,将包含数值的特征值进行数值范围的聚类;
    为所述多个用户分类赋予权值,获得所述多个用户分类及其对应的权值,并将所述权值作为适用于规则引擎的标准参数。
  6. 根据权利要求5所述的方法,其特征在于,所述对所述风险相关数据进行参数化处理,获得适用于规则引擎的标准参数,包括:
    根据所述风险相关数据生成用户关系图谱;
    根据所述用户关系图谱获取所述用户与关系图谱中威胁用户的关联系数;
    根据所述关联系数确定所述用户的威胁系数,所述威胁系数即为适用于规则引擎的标准参数。
  7. 根据权利要求1-6所述的方法,其特征在于,所述将所述标准参数导入所述规则引擎,获得风险评级,包括:
    将获得的一个或多个标准参数导入规则引擎,获取规则评分,所述规则引擎包括加权处理、二值化处理或条件选择;
    根据所述规则评分为所述用户进行风险评级,所述规则评分越高,所述风险评级的等级越高。
  8. 一种风险决策装置,其特征在于,所述风险决策装置包括:
    获取单元,用于监测用户的电子设备,获取所述电子设备中的风险相关数据;
    标准处理单元,用于对所述风险相关数据进行参数化处理,获得适用于规则引擎的标准参数;
    导入单元,用于将所述标准参数导入所述规则引擎,获得风险评级;
    决策单元,用于根据所述风险评级做出风险决策。
  9. 根据权利要求8所述的装置,其特征在于,所述风险相关数据包括设备相关数据,所述获取单元具体用于:
    获取设备的硬件参数,包括设备物理地址或设备唯一序列号;
    获取设备网络数据,包括设备连接的网络名称或网际协议地址;
    根据所述硬件参数和网络数据,确定所述设备相关数据。
  10. 根据权利要求8或9所述的装置,其特征在于,所述风险相关数据包括用户相关数据,所述获取单元具体用于:
    获取所述用户的账户信息,包括账户名称或账户交互信息;
    获取所述用户的活动信息,包括用户浏览的网页地址、操作规律或耗费时长;
    根据所述用户的账户信息和所述用户的活动信息,确定所述用户相关数据。
  11. 根据权利要求10所述的装置,其特征在于,所述标准处理单元具体用于:
    根据所述风险相关数据生成设备指纹;
    将所述设备指纹与该设备的标准指纹进行匹配,并根据匹配结果,获得标准参数,其中,所述设备指纹与所述标准指纹匹配成功时获得参数一,所述设备指纹与所述标准指纹匹配失败时获得参数二,所述参数一和所述参数二都为适用于规则引擎的标准参数。
  12. 根据权利要求11所述的装置,其特征在于,所述标准处理单元具体用于:
    获取所述风险相关数据中不同风险操作的多个特征值;
    根据所述多个特征值对所述用户相关数据进行聚类,获得多个用户分类,其中,将不包含数值的特征值进行关键字聚类,将包含数值的特征值进行数值范围的聚类;
    为所述多个用户分类赋予权值,获得所述多个用户分类及其对应的权值,并将所述权值作为适用于规则引擎的标准参数。
  13. 根据权利要求12所述的装置,其特征在于,所述标准处理单元具体用于:
    根据所述风险相关数据生成用户关系图谱;
    根据所述用户关系图谱获取所述用户与关系图谱中威胁用户的关联系数;
    根据所述关联系数确定所述用户的威胁系数,所述威胁系数即为适用于规则引擎的标准参数。
  14. 根据权利要求8-13所述的装置,其特征在于,所述导入单元具体用于:
    将获得的一个或多个标准参数导入规则引擎,获取规则评分,所述规则引擎包括加权处理、二值化处理或条件选择;
    根据所述规则评分为所述用户进行风险评级,所述规则评分越高,所述风险评级的等级越高。
  15. 一种终端,其特征在于,包括处理器、输入设备、输出设备和存储器,所述处理器、输入设备、输出设备和存储器相互连接,其中,所述存储器用于存储计算机程序,所述计算机程序包括程序指令,所述处理器被配置用于调用所述程序指令,执行:
    监测用户的电子设备,获取所述电子设备中的风险相关数据;
    对所述风险相关数据进行参数化处理,获得适用于规则引擎的标准参数,所述规则引擎包括对所述标准参数进行数值约束和/或冲突检验的规则集合;
    将所述标准参数导入所述规则引擎,获得风险评级;
    根据所述风险评级做出风险决策。
  16. 根据权利要求15所述的终端,其特征在于,所述风险相关数据包括设备相关数据,在所述获取所述电子设备中的风险相关数据方面,所述处理器具体用于:
    获取设备的硬件参数,包括设备物理地址或设备唯一序列号;
    获取设备网络数据,包括设备连接的网络名称或网际协议地址;
    根据所述硬件参数和网络数据,确定所述设备相关数据。
  17. 根据权利要求15或16所述的终端,其特征在于,所述风险相关数据包括用户相关数据,在所述获取所述电子设备中的风险相关数据方面,所述处理器具体用于:
    获取所述用户的账户信息,包括账户名称或账户交互信息;
    获取所述用户的活动信息,包括用户浏览的网页地址、操作规律或耗费时长;
    根据所述用户的账户信息和所述用户的活动信息,确定所述用户相关数据。
  18. 根据权利要求17所述的终端,其特征在于,在所述对所述风险相关数据进行参数化处理,获得适用于规则引擎的标准参数方面,所述处理器具体用于:
    根据所述风险相关数据生成设备指纹;
    将所述设备指纹与该设备的标准指纹进行匹配,并根据匹配结果,获得标准参数,其中,所述设备指纹与所述标准指纹匹配成功时获得参数一,所述设备指纹与所述标准指纹匹配失败时获得参数二,所述参数一和所述参数二都为适用于规则引擎的标准参数。
  19. 根据权利要求18所述的终端,其特征在于,在所述对所述风险相关数据进行参数化处理,获得适用于规则引擎的标准参数,方面,所述处理器具体用于:
    获取所述风险相关数据中不同风险操作的多个特征值;
    根据所述多个特征值对所述用户相关数据进行聚类,获得多个用户分类,其中,将不包含数值的特征值进行关键字聚类,将包含数值的特征值进行数值范围的聚类;
    为所述多个用户分类赋予权值,获得所述多个用户分类及其对应的权值,并将所述权值作为适用于规则引擎的标准参数。
  20. 一种计算机可读存储介质,存储用于电子数据交换的计算机程序,其中,所述计算机程序使得计算机执行权利要求1-7中任一方法所述的步骤的指令。
PCT/CN2018/123887 2018-09-06 2018-12-26 一种风险决策方法和装置 WO2020048056A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201811037817.0 2018-09-06
CN201811037817.0A CN109214683A (zh) 2018-09-06 2018-09-06 一种风险决策方法和装置

Publications (1)

Publication Number Publication Date
WO2020048056A1 true WO2020048056A1 (zh) 2020-03-12

Family

ID=64987206

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/123887 WO2020048056A1 (zh) 2018-09-06 2018-12-26 一种风险决策方法和装置

Country Status (2)

Country Link
CN (1) CN109214683A (zh)
WO (1) WO2020048056A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114386858A (zh) * 2022-01-14 2022-04-22 深圳前海环融联易信息科技服务有限公司 一种智能风险决策平台
CN114386858B (zh) * 2022-01-14 2024-05-31 深圳前海环融联易信息科技服务有限公司 一种智能风险决策平台

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111104664B (zh) * 2019-11-29 2022-03-15 北京云测信息技术有限公司 一种电子设备的风险识别方法及服务器
CN111641594B (zh) * 2020-05-09 2021-11-30 同济大学 基于页面行为的欺诈用户检测方法、系统、介质及设备
CN111724069A (zh) * 2020-06-22 2020-09-29 百度在线网络技术(北京)有限公司 用于处理数据的方法、装置、设备及存储介质
CN112508677A (zh) * 2020-11-06 2021-03-16 无锡艺界科技有限公司 一种基于大数据风控的金融系统
CN113505823B (zh) * 2021-07-02 2023-06-23 中国联合网络通信集团有限公司 供应链安全分析方法及计算机可读存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1926569A (zh) * 2004-02-23 2007-03-07 自动金融系统股份有限公司 风险管理系统和方法
CN101930585A (zh) * 2010-09-01 2010-12-29 中国建设银行股份有限公司 一种授信业务风险监测系统及其方法
CN104392391A (zh) * 2014-11-14 2015-03-04 国家电网公司 一种电网运行安全风险量化方法
CN107025509A (zh) * 2016-02-01 2017-08-08 腾讯科技(深圳)有限公司 基于业务模型的决策系统和方法

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8739278B2 (en) * 2006-04-28 2014-05-27 Oracle International Corporation Techniques for fraud monitoring and detection using application fingerprinting
CN201974854U (zh) * 2011-03-24 2011-09-14 国网电力科学研究院 用于电力二次系统的信息安全风险测量装置
CN103235983A (zh) * 2013-04-23 2013-08-07 浪潮集团山东通用软件有限公司 一种多维度的供应商自动评价方法
CN103530772A (zh) * 2013-09-30 2014-01-22 深圳钱盒信息技术有限公司 一种移动互联支付风险控制方法及系统
CN106682906B (zh) * 2015-11-10 2021-03-19 创新先进技术有限公司 一种风险识别、业务处理方法和设备
CN107067157A (zh) * 2017-03-01 2017-08-18 北京奇艺世纪科技有限公司 业务风险评估方法、装置及风控系统
CN108205783A (zh) * 2017-12-13 2018-06-26 南京农纷期电子商务有限公司 一种农业信贷领域的自动化信用评分系统
CN108038700A (zh) * 2017-12-22 2018-05-15 上海前隆信息科技有限公司 一种反欺诈数据分析方法与系统
CN108492173A (zh) * 2018-03-23 2018-09-04 上海氪信信息技术有限公司 一种基于双模网络图挖掘算法的信用卡反欺诈预测方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1926569A (zh) * 2004-02-23 2007-03-07 自动金融系统股份有限公司 风险管理系统和方法
CN101930585A (zh) * 2010-09-01 2010-12-29 中国建设银行股份有限公司 一种授信业务风险监测系统及其方法
CN104392391A (zh) * 2014-11-14 2015-03-04 国家电网公司 一种电网运行安全风险量化方法
CN107025509A (zh) * 2016-02-01 2017-08-08 腾讯科技(深圳)有限公司 基于业务模型的决策系统和方法

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114386858A (zh) * 2022-01-14 2022-04-22 深圳前海环融联易信息科技服务有限公司 一种智能风险决策平台
CN114386858B (zh) * 2022-01-14 2024-05-31 深圳前海环融联易信息科技服务有限公司 一种智能风险决策平台

Also Published As

Publication number Publication date
CN109214683A (zh) 2019-01-15

Similar Documents

Publication Publication Date Title
US11985037B2 (en) Systems and methods for conducting more reliable assessments with connectivity statistics
WO2020048056A1 (zh) 一种风险决策方法和装置
US11886555B2 (en) Online identity reputation
US10678798B2 (en) Method and system for scoring credibility of information sources
US8380607B2 (en) Predicting economic trends via network communication mood tracking
Kaur et al. A comparative analysis of structural graph metrics to identify anomalies in online social networks
US11238169B2 (en) Privacy score
US10956522B1 (en) Regular expression generation and screening of textual items
CN106874253A (zh) 识别敏感信息的方法及装置
CN109118051A (zh) 基于网络舆情的风险商户识别及处置方法、装置及服务器
WO2014207753A1 (en) Assessing value of brand based on online content
Firmansyah et al. The influence of efficacy, credibility, and normative pressure to M-banking adoption level in Indonesia
US20230093540A1 (en) System and Method for Detecting Anomalous Activity Based on a Data Distribution
CN110197426B (zh) 一种信用评分模型的建立方法、装置及可读存储介质
CN108280644B (zh) 群组成员关系数据可视化方法及系统
CN114399396A (zh) 保险产品推荐方法、装置、计算机设备及存储介质
Löbner et al. Explainable machine learning for default privacy setting prediction
US20140180765A1 (en) Web-based survey verification
CN116318974A (zh) 站点风险识别方法、装置、计算机可读介质及电子设备
Sen et al. Analysis of media bias in policy discourse in india
US20230039039A1 (en) Process for determining a degree of data exposure
CN109636568A (zh) 电话号码的风险检测方法、装置、设备及存储介质
CN110610378A (zh) 产品需求分析方法、装置、计算机设备和存储介质
US11410447B2 (en) Information security assessment translation engine
US20150340026A1 (en) Extracting candidate answers for a knowledge base from conversational sources

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: 18932475

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: 18932475

Country of ref document: EP

Kind code of ref document: A1