WO2022068280A1

WO2022068280A1 - Data processing method and apparatus, device, and storage medium

Info

Publication number: WO2022068280A1
Application number: PCT/CN2021/101899
Authority: WO
Inventors: 唐圳; 杨涵; 刘博�; 郑文琛
Original assignee: 深圳前海微众银行股份有限公司
Priority date: 2020-09-30
Filing date: 2021-06-23
Publication date: 2022-04-07
Also published as: CN112163154A; CN112163154B

Abstract

The present application discloses a data processing method and apparatus, a device, and a storage medium. The method comprises: acquiring locally stored attribute information of a client; inputting the attribute information of the client into a preset probability estimation model of a target service, so as to obtain an estimated probability of the client; and if it is determined, according to the estimated probability of the client, that the client needs to be contacted via a target contact manner, sending contact instruction information to a terminal of service personnel, wherein the contact instruction information is used to instruct the service personnel to contact, via the target contact manner, the client about the target service. In the present application, a client in need of a target service can be quickly and accurately identified, such that said client can be targeted and contacted via a target contact manner, thereby improving conversion efficiency for the target service, and accordingly improving the performance and work efficiency of service personnel.

Description

数据处理方法、装置、设备及存储介质Data processing method, apparatus, equipment and storage medium

本申请要求于2020年09月30日提交中国专利局、申请号为202011061868.4、申请名称为“数据处理方法、装置、设备及存储介质”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。This application claims the priority of the Chinese patent application filed on September 30, 2020 with the application number 202011061868.4 and the application name "Data Processing Method, Apparatus, Equipment and Storage Medium", the entire contents of which are incorporated by reference in in this application.

技术领域technical field

本申请涉及人工智能技术领域，尤其涉及一种数据处理方法、装置、设备及存储介质。The present application relates to the technical field of artificial intelligence, and in particular, to a data processing method, apparatus, device, and storage medium.

背景技术Background technique

随着金融企业行业市场趋于饱和，新客户的挖掘成本越来越高，而维护存量客户的成本远小于挖掘新客户的成本。因此，越来越多的金融企业将目光聚焦于存量客户，而如何经营存量客户，使其尽可能多的存量客户购买金融产品对于金融企业来说至关重要。As the financial enterprise industry market tends to be saturated, the cost of finding new customers is getting higher and higher, while the cost of maintaining existing customers is far less than the cost of mining new customers. Therefore, more and more financial enterprises are focusing on existing customers, and how to manage existing customers so that as many existing customers as possible to purchase financial products is very important for financial enterprises.

目前，对于处于高度活跃期的存量客户，客户经理会针对***。对于低活跃期的存量客户，通常采用电销的方式向其推销金融产品。但是，由于该类存量客户与企业金融产品之间的互动行为较少，导致对这类存量客户的需求了解的不准确。因此，为了解该类存量客户的需求，通常通过电销的方式接触该类存量客户。At present, for existing customers in a highly active period, the account manager will provide targeted services. For existing customers in the low active period, financial products are usually sold to them by means of telemarketing. However, due to the lack of interaction between such existing customers and corporate financial products, the understanding of the needs of such existing customers is inaccurate. Therefore, in order to understand the needs of such existing customers, we usually contact such existing customers through telemarketing.

但是，电销接触方式的覆盖度小、成本高，难以挖掘出具有需求的客户，影响坐席的业务量，从而导致坐席的工作效率低。However, the electric pin contact method has small coverage and high cost, and it is difficult to dig out customers with needs, which affects the business volume of the agents, resulting in low work efficiency of the agents.

技术解决方案technical solutions

本申请的主要目的在于提供一种数据处理方法、装置、设备及存储介质，旨在解决因对客户需求了解不准确而导致的业务人员业绩差，工作效率低的问题。The main purpose of the present application is to provide a data processing method, device, equipment and storage medium, aiming to solve the problems of poor performance and low work efficiency of business personnel caused by inaccurate understanding of customer needs.

为实现上述目的，本申请提供一种数据处理方法，包括：To achieve the above purpose, the present application provides a data processing method, comprising:

获取本地存储的客户的属性信息；Get the attribute information of the customer stored locally;

将所述客户的属性信息输入到目标业务的预设概率预估模型中，获得所述客户的预估概率，所述预估概率用于表示通过目标接触方式接触所述客户后，所述客户在预设时长内针对所述目标业务转化的概率；Input the attribute information of the customer into the preset probability estimation model of the target business, and obtain the estimated probability of the customer, and the estimated probability is used to indicate that after the customer is contacted by the target contact method, the customer The probability of conversion for the target business within a preset time period;

若根据所述客户的预估概率，确定需通过目标接触方式进行接触所述客户，则向业务人员的终端发送接触提示信息，所述接触提示信息用于提示针对所述目标业务通过目标接触方式接触所述客户。If, according to the estimated probability of the customer, it is determined that the customer needs to be contacted by the target contact method, contact prompt information is sent to the terminal of the business personnel, and the contact prompt information is used to prompt the target business to be contacted by the target contact method. Contact said customer.

在一种具体实施方式中，In a specific embodiment,

所述预设概率预估模型是根据第一训练样本集和第二训练样本集训练获得的；The preset probability estimation model is obtained by training according to the first training sample set and the second training sample set;

所述第一训练样本集包含：通过目标接触方式接触的样本客户的属性信息、以所述目标接触方式接触所述样本客户的时间点以及第一标签，所述第一标签用于表示以所述目标接触方式接触所述样本客户的时间点起预设时长内是否针对所述目标产品转化；The first training sample set includes: attribute information of sample customers contacted by the target contact method, time points of contacting the sample customers by the target contact method, and a first label, where the first label is used to represent Whether the target product is converted within a preset time period from the time when the target contact method contacts the sample customer;

所述第二训练样本集包含：未通过目标接触方式接触的样本客户的属性信息、预设时间点以及第二标签，所述第二标签用于表示以预设时间点起所述预设时长内是否针对所述目标产品转化。The second training sample set includes: attribute information of sample customers who have not been contacted by the target contact method, a preset time point, and a second label, where the second label is used to indicate the preset duration from the preset time point whether to convert for the target product.

在一种具体实施方式中，所述以所述目标接触方式接触所述样本客户的时间点与所述获取所述客户的属性信息的时间点之间的间隔时长大于或等于预设时长；In a specific implementation manner, the interval duration between the time point of contacting the sample customer in the target contact manner and the time point of acquiring the attribute information of the customer is greater than or equal to a preset duration;

所述预设时间点与所述获取所述客户的属性信息的时间点之间的间隔时长大于或等于预设时长。The interval duration between the preset time point and the time point of acquiring the attribute information of the customer is greater than or equal to the preset time duration.

在一种具体实施方式中，所述将所述客户的第一特征输入到目标产品的预设概率预估模型中之前，还包括：In a specific embodiment, before the inputting the first characteristic of the customer into the preset probability estimation model of the target product, the method further includes:

获取所述第一训练样本集和所述第二训练样本集；obtaining the first training sample set and the second training sample set;

根据所述第一训练样本集和所述第二训练样本集，训练初始的概率预估模型，获得所述预设概率预估模型。According to the first training sample set and the second training sample set, an initial probability prediction model is trained to obtain the preset probability prediction model.

在一种具体实施方式中，所述根据第一训练样本集和第二训练样本集，训练初始的概率预估模型，获得所述预设概率预估模型，包括：In a specific embodiment, the initial probability prediction model is trained according to the first training sample set and the second training sample set, and the preset probability prediction model is obtained, including:

根据所述第二训练样本集，训练所述初始的概率预估模型，获得中间预设概率预估模型；According to the second training sample set, the initial probability prediction model is trained to obtain an intermediate preset probability prediction model;

根据所述第一训练样本集，训练所述中间预设概率预估模型，获得所述预设概率预估模型。According to the first training sample set, the intermediate preset probability prediction model is trained to obtain the preset probability prediction model.

在一种具体实施方式中，获取所述第二训练样本集，包括：In a specific implementation manner, acquiring the second training sample set includes:

在同一个预设时间点获取未通过目标接触方式接触的不同样本客户中每个样本客户的属性信息以及所述第二标签；或者，Obtain the attribute information and the second label of each sample customer among the different sample customers who are not contacted by the target contact at the same preset time point; or,

在多个不同预设时间点分别获取未通过目标接触方式接触的同一样本客户的属性信息以及所述第二标签。The attribute information and the second label of the same sample customer who have not been contacted by the target contact method are respectively acquired at multiple different preset time points.

在一种具体实施方式中，还包括：In a specific embodiment, it also includes:

在对所述客户通过目标接触方式进行接触后，预设时长内所述客户针对所述目标产品未转化，则将所述客户作为第一训练样本集中的样本客户，获得更新后的第一训练样本集；After the customer is contacted by the target contact method, if the customer has not converted to the target product within the preset time period, the customer will be regarded as a sample customer in the first training sample set, and the updated first training will be obtained. sample set;

根据更新后的第一训练样本集，对所述预设概率预估模型进行优化。The preset probability prediction model is optimized according to the updated first training sample set.

在一种具体实施方式中，所述将所述客户作为第一训练样本集中的样本客户，获得更新后的第一训练样本集，包括：In a specific embodiment, the obtaining the updated first training sample set by using the customer as a sample customer in the first training sample set includes:

将所述客户作为第一训练样本集中的多份样本客户，获得更新后的第一训练样本集。Taking the customer as a plurality of sample customers in the first training sample set, the updated first training sample set is obtained.

在一种具体实施方式中，所述属性信息包括如下至少一项：所属的行业信息、名称、地址、提款信息。In a specific implementation manner, the attribute information includes at least one of the following: industry information, name, address, and withdrawal information.

本申请还提供一种数据处理装置，包括：The present application also provides a data processing device, comprising:

获取模块，用于获取本地存储的客户的属性信息；The acquisition module is used to acquire the attribute information of the customer stored locally;

预估模块，用于将客户的属性信息输入到目标业务的预设概率预估模型中，获得客户的预估概率，预估概率用于表示通过目标接触方式接触客户后，客户在预设时长内针对目标业务转化的概率；The estimation module is used to input the attribute information of the customer into the preset probability estimation model of the target business, and obtain the estimated probability of the customer. The probability of target business conversion within the target business;

确定模块，用于若根据客户的预估概率，确定需通过目标接触方式进行接触客户，则向业务人员的终端发送接触提示信息，接触提示信息用于提示针对目标业务通过目标接触方式接触客户。The determining module is configured to send contact prompt information to the terminal of the business personnel if it is determined that the customer needs to be contacted by the target contact method according to the estimated probability of the customer, and the contact prompt information is used to prompt the target business to contact the customer by the target contact method.

本申请还提供一种电子设备，所述电子设备包括：存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述计算机程序被所述处理器执行时实现第一方面任一实施方式所述的数据处理方法的步骤。The present application also provides an electronic device, the electronic device includes: a memory, a processor, and a computer program stored on the memory and executable on the processor, when the computer program is executed by the processor Implement the steps of the data processing method described in any embodiment of the first aspect.

本申请还提供一种计算机可读存储介质，所述计算机可读存储介质上存储有计算机程序，所述计算机程序被处理器执行时实现如第一方面任一实施方式提供的数据处理方法的步骤。The present application further provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, implements the steps of the data processing method provided by any embodiment of the first aspect .

本申请提供一种程序产品，所述程序产品包括计算机程序，所述计算机程序存储在可读存储介质中，电子设备的至少一个处理器可以从所述可读存储介质读取所述计算机程序，所述至少一个处理器执行所述计算机程序使得电子设备实施第一方面任一项提供的数据处理方法。The present application provides a program product, the program product includes a computer program, the computer program is stored in a readable storage medium, and at least one processor of an electronic device can read the computer program from the readable storage medium, Executing the computer program by the at least one processor causes the electronic device to implement the data processing method provided in any one of the first aspects.

本申请中，对于每个客户，获取每个客户的属性信息，将每个客户的属性信息输入到目标业务的预设概率预估模型中。由于客户的属性信息反映该客户的需求点，因此，通过预设概率预估模型对每个客户的属性信息进行分析，获得每个客户对目标业务的需求程度，根据需求程度给出每个客户的预估概率。根据客户的预估概率，若确定通过目标接触方式进行接触的至少一个客户，则向业务人员的终端发送接触提示信息，以对客户通过目标接触方式进行接触。因此，本实施例根据客户的属性信息以及预设概率预估模型快速且准确的确定出对目标业务有需求的客户，从而有针对性的对客户通过目标接触方式接触，提高了在对客户进行目标方式接触后，针对目标业务成功转化的效率。并且在尽可能减少通过目标接触方式接触的客户的数量的基础上，提高了针对目标产品成功转化的效率，还降低了接触成本，从而提高业务人员业绩，提高业务人员的工作效率。In this application, for each customer, the attribute information of each customer is obtained, and the attribute information of each customer is input into the preset probability estimation model of the target business. Since the customer's attribute information reflects the customer's demand point, the attribute information of each customer is analyzed through a preset probability estimation model to obtain the demand level of each customer for the target business, and each customer is given according to the demand level. the estimated probability of . According to the estimated probability of the customer, if at least one customer contacted by the target contact method is determined, contact prompt information is sent to the terminal of the business personnel, so as to contact the customer by the target contact method. Therefore, this embodiment quickly and accurately determines the customers who have needs for the target business according to the attribute information of the customers and the preset probability estimation model, so that the customers can be contacted by the target contact method in a targeted manner, which improves the performance of the customers. After the target method is contacted, the efficiency of the successful conversion of the target business. And on the basis of reducing the number of customers contacted through the target contact method as much as possible, the efficiency of successful conversion of target products is improved, and the contact cost is also reduced, thereby improving the performance of business personnel and improving the work efficiency of business personnel.

附图说明Description of drawings

图1为本申请一实施例提供的一种应用场景示意图；FIG. 1 is a schematic diagram of an application scenario provided by an embodiment of the present application;

图2为本申请一实施例提供的一种数据处理方法的流程示意图；2 is a schematic flowchart of a data processing method provided by an embodiment of the present application;

图3为本申请一实施例提供的一种数据处理方法的框图；3 is a block diagram of a data processing method provided by an embodiment of the present application;

图4为本申请一实施例提供的获取预设概率预估模型的方法流程图；4 is a flowchart of a method for obtaining a preset probability prediction model provided by an embodiment of the present application;

图5为本申请一实施例提供的预设概率预估模型训练框图；5 is a training block diagram of a preset probability estimation model provided by an embodiment of the present application;

图6为本申请一实施例提供的数据处理装置的结构示意图；FIG. 6 is a schematic structural diagram of a data processing apparatus provided by an embodiment of the present application;

图7为本申请一实施例提供的电子设备的结构示意图。FIG. 7 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

本申请目的的实现、功能特点及优点将结合实施例，参照附图做进一步说明。The realization, functional characteristics and advantages of the purpose of the present application will be further described with reference to the accompanying drawings in conjunction with the embodiments.

本发明的实施方式Embodiments of the present invention

下面将参照附图更详细地描述本公开的示例性实施例。虽然附图中显示了本公开的示例性实施例，然而应当理解，可以以各种形式实现本公开而不应被这里阐述的实施例所限制。相反，提供这些实施例是为了能够更透彻地理解本公开，并且能够将本公开的范围完整的传达给本领域的技术人员。Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided so that the present disclosure will be more thoroughly understood, and will fully convey the scope of the present disclosure to those skilled in the art.

图1为本申请一实施例提供的一种应用场景示意图。对于低活跃期的存量客户，由于这些存量客户在最近一段时间内与企业之间的互动行为较少，因此，坐席无法确定低活跃期的存量客户的业务需求。为了了解低活跃期的存量客户的需求点，坐席通常通过电销方式接触低活跃期的存量客户，向低活跃期的存量客户介绍金融产品并了解他们的需求点。FIG. 1 is a schematic diagram of an application scenario provided by an embodiment of the present application. For the existing customers in the low active period, because these existing customers have less interaction with the enterprise in the recent period, the agent cannot determine the business needs of the existing customers in the low active period. In order to understand the demand points of existing customers in the low active period, agents usually contact the existing customers in the low active period through telemarketing, introduce financial products to the existing customers in the low active period and understand their demand points.

例如如图1所示，存量客户的名单存储在数据库102中，服务器101从数据库102中获取存量客户的名单，采用随机分配或者根据经验将存量客户的名单发送到对应坐席的终端上，以使坐席根据名单中客户的信息对客户进行电销接触，从而促使存量客户对业务进行转化。For example, as shown in FIG. 1, the list of existing customers is stored in the database 102, and the server 101 obtains the list of existing customers from the database 102, and sends the list of existing customers to the terminal of the corresponding agent by random allocation or based on experience, so that the The agent conducts telemarketing contact with customers based on the information of the customers in the list, thereby prompting the existing customers to convert the business.

其中，现有技术中坐席在通过电销方式接触存量客户时，由于不了解存量客户的需求，只能盲目的向存量客户介绍产品，导致客户的转化率低，即影响坐席的业绩。因此，为了增加业务，坐席会尽可能的通过电销方式接触更多的存量客户，但是，坐席通过电销方式接触存量客户的数量是有限的，难以接触到每个存量客户，存量客户的覆盖率低，也影响存量客户的转化率，导致坐席的工作效率低。Among them, in the prior art, when an agent contacts existing customers through telemarketing, because they do not understand the needs of the existing customers, they can only blindly introduce products to the existing customers, resulting in a low customer conversion rate, which affects the performance of the agent. Therefore, in order to increase business, agents will try their best to contact more existing customers through telemarketing. However, the number of agents who can contact existing customers through telemarketing is limited, and it is difficult to reach each existing customer and cover the existing customers. The low rate also affects the conversion rate of existing customers, resulting in low work efficiency of agents.

为了解决上述问题中的至少一个问题，本申请实施例提供一种方案，将客户特征输入到预设概率预估模型中，预设概率预估模型根据每个客户的特征，确定每个客户的预估概率，所述预估概率用于表示通过目标接触方式接触每个客户后，该客户在预设时长内转化的概率，从而根据预估概率确定接下来通过目标接触方式接触的客户。由于客户特征是基于各种关于客户的信息和数据分析后获取的，可以反映客户的需求，因此，通过预设概率预估模型对客户特征的计算，可以快速、准确地获取客户需求，从而在通过目标方式接触客户时，为客户提供符合其需求的产品，提高客户转化率，从而提高坐席的工作效率。In order to solve at least one of the above problems, an embodiment of the present application provides a solution, in which customer characteristics are input into a preset probability estimation model, and the preset probability estimation model determines the characteristics of each customer based on the characteristics of each customer. Estimated probability, where the estimated probability is used to indicate the probability that each customer will be converted within a preset period of time after contacting each customer through the target contact method, so that the next customer to be contacted through the target contact method is determined according to the estimated probability. Since customer characteristics are obtained based on various information and data analysis about customers and can reflect customer needs, the calculation of customer characteristics through the preset probability estimation model can quickly and accurately obtain customer needs, so that in When reaching customers in a targeted way, provide customers with products that meet their needs, improve customer conversion rates, and thus improve the work efficiency of agents.

下面结合附图，对本申请的一些实施方式作详细说明。在各实施例之间不冲突的情况下，下述的实施例及实施例中的特征可以相互组合。Some embodiments of the present application will be described in detail below with reference to the accompanying drawings. The following embodiments and features in the embodiments may be combined with each other without conflict between the embodiments.

图2为本申请一实施例提供的一种数据处理方法的流程示意图。本实施例中方法的执行主体可以为电子设备，例如计算机或服务器。本实施例中的方法可以通过软件、硬件或者软硬件结合的方式来实现。如图2所示，所述方法可以包括：FIG. 2 is a schematic flowchart of a data processing method provided by an embodiment of the present application. The execution body of the method in this embodiment may be an electronic device, such as a computer or a server. The method in this embodiment may be implemented by software, hardware, or a combination of software and hardware. As shown in Figure 2, the method may include:

S201、获取本地存储的客户的属性信息。S201. Acquire attribute information of a customer stored locally.

在本步骤中，客户可以是个人，也可以是企业、团体组织等。下面，以客户为企业为例进行说明。In this step, the customer may be an individual, or an enterprise, a group organization, or the like. Below, the customer is an enterprise as an example to explain.

当客户接触过公司产品，并且，向公司提交过自身的属性信息后，公司的数据库中存储有这类客户的属性信息。When customers have contacted the company's products and submitted their own attribute information to the company, the company's database stores the attribute information of such customers.

由于企业之前是公司的客户，因此，企业的信息存储在公司的数据库102中，当对这些企业进行电销接触时，服务器101从数据库102中获取客户的名单，其中，客户的名单上包括客户的属性信息。Since the enterprise was the company's customer before, the information of the enterprise is stored in the company's database 102. When contacting these enterprises for telemarketing, the server 101 obtains the customer's list from the database 102, wherein the customer's list includes the customer attribute information.

其中，由于数据库中储存的客户的数量十分庞大，因此，可以每次从数据库中获取部分客户的名单，例如，随机获取数据库中的部分客户；或者，获取上一次接触该客户的时间与当前时间的时间间隔超过预设时间间隔的客户，本申请对此不限制。Among them, since the number of customers stored in the database is very large, the list of some customers can be obtained from the database each time, for example, some customers in the database can be obtained randomly; or, the time and current time of the last contact with the customer can be obtained For clients whose time interval exceeds the preset time interval, this application does not limit this.

获取到客户的名单后，获取客户的属性信息。可选的，属性信息包括如下至少一项：待筛选客户所属的行业信息、提款信息、基础信息。After obtaining the customer's list, obtain the customer's attribute information. Optionally, the attribute information includes at least one of the following: industry information, withdrawal information, and basic information to which the customer to be screened belongs.

对于行业信息，同一行业，由于其对应的社会领域相同，因此，当该行业对应的社会领域促进该行业发展时，处于同一行业中的相关企业也会得到发展。例如，在出现疫情时，会促进医药卫生领域相关企业的发展，例如，医疗器械制造企业、生产口罩的企业。因此，行业信息可以反应该行业的发展状况，继而通过企业行业信息，从行业维度确定企业的需求，从而为客户提供符合其需求的产品。For industry information, the same industry has the same corresponding social field. Therefore, when the corresponding social field of the industry promotes the development of the industry, related enterprises in the same industry will also develop. For example, in the event of an epidemic, it will promote the development of related companies in the medical and health field, such as medical device manufacturers and mask manufacturers. Therefore, industry information can reflect the development status of the industry, and then through the enterprise industry information, the needs of the enterprise can be determined from the industry dimension, so as to provide customers with products that meet their needs.

可选的，行业信息可以具体包括：企业的经营范围。对于具有相同经营范围的企业，或者经营同一类产品的不同企业，其生产产品的原料、销售、生产设备的购买渠道等所对应的供应链几乎相同。因此，对于具有相同经营范围的企业，或者经营同一类产品的不同企业，如果一家企业受到影响，其他企业受到影响的概率也会增加。Optionally, the industry information may specifically include: the business scope of the enterprise. For enterprises with the same business scope, or different enterprises operating the same type of products, the supply chains corresponding to the raw materials, sales, and purchasing channels of production equipment are almost the same. Therefore, for companies with the same business scope, or different companies operating the same type of product, if one company is affected, the probability of other companies being affected will also increase.

并且，对于具有相同供应链的不同企业，当供应链中的其中一个环节出现问题或者发展壮大时，可能会对整个供应链上的行业产生影响。例如，当口罩企业的生产量增加时，处于口罩生产供应链上的不同行业的企业的效益也会随之增加，例如，口罩原材料提供企业、口罩生产设备企业等。当口罩生产设备企业需要生产大量的设备时，其对应的原料供应企业也会随之发展，例如，配件生产企业，其中，配件生产企业又属于机械制造行业。因此，行业信息不仅可以反应该行业的发展状况，还可以反应与该行业相关的其他行业的信息，通过一个行业的行业信息，可以推测其他相关行业的发展状况，从而确定企业的需求。Moreover, for different companies with the same supply chain, when one link in the supply chain has a problem or develops and grows, it may have an impact on the industry in the entire supply chain. For example, when the production volume of mask companies increases, the benefits of companies in different industries in the mask production supply chain will also increase, such as mask raw material suppliers, mask production equipment companies, etc. When a mask production equipment company needs to produce a large amount of equipment, its corresponding raw material supply companies will also develop accordingly. For example, accessories production companies, among which accessories production companies belong to the machinery manufacturing industry. Therefore, industry information can not only reflect the development status of the industry, but also reflect the information of other industries related to the industry. Through the industry information of an industry, the development status of other related industries can be inferred, so as to determine the needs of enterprises.

可选的，行业信息还可以具体包括：企业的经营规模。企业的经营规模不同，其对风险的抵抗能力也不同，当然，在机遇来临时，其发展潜力也不同。因此，获取企业的经营规模，根据经营规模确定企业的抗风险能力或者发展潜力，从而确定企业的需求。Optionally, the industry information may further specifically include: the business scale of the enterprise. Enterprises with different operating scales have different resistance to risks. Of course, when opportunities come, their development potential is also different. Therefore, the operation scale of the enterprise is obtained, and the anti-risk ability or development potential of the enterprise is determined according to the operation scale, so as to determine the needs of the enterprise.

对于提款信息，可选的，提款信息可以包括以下至少一项：该企业已贷款的金额、贷款时间、已有贷款的申请时间、已有贷款的提款时间、已有贷款的还款信息、信用度。With regard to the withdrawal information, optionally, the withdrawal information may include at least one of the following: the amount of the loan, the loan time, the application time of the existing loan, the withdrawal time of the existing loan, and the repayment of the existing loan. information and credibility.

通过提款信息，可以反应企业的历史经营状况，从而可以预测企业的需求以及还款能力等。Through the withdrawal information, the historical operation status of the enterprise can be reflected, so that the demand and repayment ability of the enterprise can be predicted.

其中，相同行业或者说相同类型的企业以及主要供应商等，为了减少成本以及促进公司的发展，企业设置的区域会有一定的属性，例如，电子制造业，一般聚集在一定的区域。Among them, the same industry or the same type of enterprises and major suppliers, in order to reduce costs and promote the development of the company, the area set up by the enterprise will have certain attributes. For example, the electronics manufacturing industry is generally concentrated in a certain area.

并且，对于有些行业，由于行业的限制，其相关产业也聚集在一定的区域，例如，能源产业，具体的，例如，煤炭行业，其相关行业均设置在煤炭资源丰富的地区。Moreover, for some industries, due to industry restrictions, their related industries are also concentrated in certain areas, for example, the energy industry, specifically, for example, the coal industry, and its related industries are located in areas rich in coal resources.

因此，同一区域的相关企业之间的发展会相互影响，因此，对于其中一家企业，通过该企业地址，可以获取该地址对应区域其他相关企业的需求，从而确定该企业的需求。Therefore, the development of related enterprises in the same area will affect each other. Therefore, for one of the enterprises, through the enterprise address, the needs of other related enterprises in the corresponding area of the address can be obtained, so as to determine the needs of the enterprise.

当企业在之前提交的信息中没有填写行业信息时，通过企业名称，可以分析获取企业的行业，经营范围。可选的，企业名称中会出现企业的地址信息，因此，通过企业名称也可以获取该企业的企业地址。When the enterprise does not fill in the industry information in the previously submitted information, through the enterprise name, the industry and business scope of the enterprise can be analyzed and obtained. Optionally, the address information of the enterprise will appear in the enterprise name, so the enterprise address of the enterprise can also be obtained through the enterprise name.

S202、将客户的属性信息输入到目标业务的预设概率预估模型中，获得客户的预估概率。S202. Input the attribute information of the customer into the preset probability estimation model of the target business to obtain the estimated probability of the customer.

其中，预估概率用于表示通过目标接触方式接触客户后，客户在预设时长内针对目标业务转化的概率。Among them, the estimated probability is used to indicate the probability that the customer will convert to the target business within a preset period of time after contacting the customer through the target contact method.

在本步骤中，以金融公司为例，目标业务例如可以是金融公司推出的理财产品、贷款产品等，本实施例以其中一款贷款业务为例进行说明。In this step, taking a financial company as an example, the target business may be, for example, a wealth management product or a loan product launched by the financial company, and this embodiment takes one of the loan businesses as an example for description.

目标接触方式是预先设置的接触方式，例如，邮件、电话、短信等方式。其中，本实施例以电话接触（即电销）的方式为例进行说明。The target contact method is a preset contact method, such as email, phone call, text message, and the like. Among them, the present embodiment takes the method of telephone contact (ie, electric pin) as an example for description.

贷款业务的预设概率预估模型（以下称为预设概率预估模型）为预先针对该贷款业务，通过人工智能的方法训练模型获得的，用于计算在通过电销方式接触客户后，客户在预设时长内办理贷款业务的概率。因此，通过预设概率预估模型获得预估概率，可以提高预估效率和预估的准确性。The preset probability prediction model of the loan business (hereinafter referred to as the preset probability prediction model) is obtained by training the model through artificial intelligence methods for the loan business in advance, and is used to calculate the value of the customer after contacting the customer through telemarketing. The probability of processing a loan business within a preset period of time. Therefore, obtaining the estimated probability through the preset probability estimation model can improve the estimation efficiency and the estimation accuracy.

例如，针对其中一种贷款业务训练获得的预设概率预估模型，通过该预设概率预估模型，可以计算出通过电销方式接触客户后，客户在预设时长内进行贷款的概率。其中，预设概率预估模型的其中一种训练方法在图4中详细介绍。For example, with respect to a preset probability estimation model obtained by training one of the loan businesses, through the preset probability estimation model, it is possible to calculate the probability that the customer will take out a loan within a preset period of time after contacting the customer through telemarketing. Among them, one of the training methods of the preset probability prediction model is described in detail in FIG. 4 .

预设时长为预先设置的用于评估客户是否办理贷款业务的时长，其值例如可以为一周、10天、30天等，在预设时长内客户办理贷款业务，都认为客户针对该贷款业务成功转化。The preset duration is the preset duration used to evaluate whether the customer handles the loan business, and its value can be, for example, one week, 10 days, 30 days, etc. If the client handles the loan business within the preset duration, it is considered that the client has succeeded in the loan business. transform.

通过S201获取到每个客户的属性信息后，如图3所示，将客户的属性信息输入到预设概率预估模型中，预设概率预估模型对客户的属性信息进行分析计算，输出客户的预估概率。After acquiring the attribute information of each customer through S201, as shown in Figure 3, the attribute information of the customer is input into the preset probability estimation model, and the preset probability estimation model analyzes and calculates the attribute information of the customer, and outputs the customer the estimated probability of .

可选的，S202的一种可能的实现方式为：每次将一个客户的属性信息输入到预设概率预估模型中，预设概率预估模型对该待筛选客户的特征进行分析计算，获取该待筛选客户的预估概率，然后，将另一个客户的属性信息输入到预设概率预估模型中。Optionally, a possible implementation of S202 is: each time the attribute information of a customer is input into the preset probability estimation model, the preset probability estimation model analyzes and calculates the characteristics of the customer to be screened, and obtains the information. The estimated probability of the customer to be screened, and then the attribute information of another customer is input into the preset probability estimation model.

可选的，S202的另一种可能的实现方式为：将所有客户的属性信息输入到预设概率预估模型内部中，预设概率预估模型内部分别对每个客户的属性信息进行分析计算，获取每个客户的预估概率。Optionally, another possible implementation manner of S202 is: input the attribute information of all customers into the preset probability estimation model, and analyze and calculate the attribute information of each customer in the preset probability estimation model. , to get the estimated probability for each customer.

对于业务人员来说，基于业绩需求，希望在接触客户后，客户能很快办理贷款业务，因此，预估概率为通过电销方式接触客户后，客户在预设时长内办理贷款业务的概率。For business personnel, based on performance needs, it is hoped that after contacting customers, customers can quickly handle loan business. Therefore, the estimated probability is the probability that customers will handle loan business within a preset period of time after contacting customers through telemarketing.

因此，可选的，在将客户的属性信息输入到预设概率预估模型时，还需要输入当前时间；或者，预设概率预估模型接收到客户的特征后，记录当前时间，并基于当前时间计算预估概率。其中，当前时间可以为当天的日期。Therefore, optionally, when inputting the attribute information of the customer into the preset probability estimation model, the current time also needs to be input; or, after the preset probability estimation model receives the customer's characteristics, the current time is recorded, and the Time to calculate the estimated probability. Wherein, the current time can be the date of the current day.

S203、若根据客户的预估概率，确定需通过目标接触方式进行接触客户，则向业务人员的终端发送接触提示信息。S203 , if it is determined that the customer needs to be contacted by the target contact method according to the estimated probability of the customer, send contact prompt information to the terminal of the business person.

其中，接触提示信息用于提示针对目标业务通过目标接触方式接触客户。Among them, the contact prompt information is used to prompt the target business to contact the customer through the target contact method.

在本步骤中，获取到客户的预估概率后，根据每个客户的预估概率，确是否通过电销方式接触该客户，若确定通过电销方式接触该客户，则向业务人员的终端发送接触提示信息，以使业务人员通过电销方式接触该客户。其中，接触提示信息例如可以是客户的名单。In this step, after obtaining the estimated probability of the customer, according to the estimated probability of each customer, whether to contact the customer through telemarketing, if it is determined that the customer is contacted by telemarketing, send a message to the terminal of the salesperson. Contact prompt information to enable business personnel to contact the customer through telemarketing. The contact prompt information may be, for example, a list of customers.

可选的，S203的一种实现方式为：根据每个客户的预估概率，确定客户中预估概率大于预设概率的客户，对预估概率大于预设概率的客户通过目标接触方式进行接触。Optionally, an implementation manner of S203 is: according to the estimated probability of each customer, determine the customers whose estimated probability is greater than the preset probability among the customers, and contact the customers whose estimated probability is greater than the preset probability through the target contact method. .

预设概率为预先设置的概率，用于确定是否有必要通过电销方式接触客户。当S202中通过预设概率预估模型计算获得的客户的预估概率大于该预设概率时，说明在通过电销方式接触该客户后，在预设时长内该客户办理贷款业务的可能性较大。The preset probability is a preset probability used to determine whether it is necessary to contact the customer through telemarketing. When the estimated probability of the client calculated by the preset probability estimation model in S202 is greater than the preset probability, it means that after contacting the client through telemarketing, the client is more likely to handle the loan business within the preset time period. big.

因此，获取每个客户的预估概率后，将每个客户的预估概率与预估概率记性比较，对预估概率大于预设概率的客户，通过电销方式进行接触。Therefore, after obtaining the estimated probability of each customer, compare the estimated probability of each customer with the estimated probability memory, and contact customers whose estimated probability is greater than the preset probability through telemarketing.

可选的，S203的另一种实现方式为：根据每个客户的预估概率的大小，确定预估概率最大的前N个客户，对预估概率最大的前N个客户通过目标接触方式进行接触。其中，N为正整数。Optionally, another implementation manner of S203 is: according to the estimated probability of each customer, determine the top N customers with the largest estimated probability, and conduct the target contact method for the top N customers with the largest estimated probability. touch. Among them, N is a positive integer.

例如，客户1-客户10的预估概率分别为：0.65、0.32、0.98、0.75、0.54、0.12、0.86、0.90、0.78、0.48，N为4，则通过电销方式进行接触的客户为：客户3、客户7、客户8以及客户9。For example, the estimated probabilities of customer 1-customer 10 are: 0.65, 0.32, 0.98, 0.75, 0.54, 0.12, 0.86, 0.90, 0.78, 0.48, and N is 4, then the customers who are contacted by telemarketing are: customers 3. Customer 7, Customer 8 and Customer 9.

本实施例提供的数据处理方法，对于每个客户，获取每个客户的属性信息，将每个客户的属性信息输入到目标业务的预设概率预估模型中。由于客户的属性信息反映该客户的需求点，因此，通过预设概率预估模型对每个客户的属性信息进行分析，获得每个客户对目标业务的需求程度，根据需求程度给出每个客户的预估概率。根据客户的预估概率，若确定通过目标接触方式进行接触的至少一个客户，则向业务人员的终端发送接触提示信息，以对客户通过目标接触方式进行接触。因此，本实施例根据客户的属性信息以及预设概率预估模型快速且准确的确定出对目标业务有需求的客户，从而有针对性的对客户通过目标接触方式接触，提高了在对客户进行目标方式接触后，针对目标业务成功转化的效率。并且在尽可能减少通过目标接触方式接触的客户的数量的基础上，提高了针对目标产品成功转化的效率，还降低了接触成本，从而提高业务人员业绩，提高业务人员的工作效率。In the data processing method provided by this embodiment, for each customer, the attribute information of each customer is acquired, and the attribute information of each customer is input into the preset probability estimation model of the target business. Since the customer's attribute information reflects the customer's demand point, the attribute information of each customer is analyzed through a preset probability estimation model to obtain the demand level of each customer for the target business, and each customer is given according to the demand level. the estimated probability of . According to the estimated probability of the customer, if at least one customer contacted by the target contact method is determined, contact prompt information is sent to the terminal of the business personnel, so as to contact the customer by the target contact method. Therefore, this embodiment quickly and accurately determines the customers who have needs for the target business according to the attribute information of the customers and the preset probability estimation model, so that the customers can be contacted by the target contact method in a targeted manner, which improves the performance of the customers. After the target method is contacted, the efficiency of the successful conversion of the target business. And on the basis of reducing the number of customers contacted through the target contact method as much as possible, the efficiency of successful conversion of target products is improved, and the contact cost is also reduced, thereby improving the performance of business personnel and improving the work efficiency of business personnel.

图4为本申请一实施例提供的获取预设概率预估模型的方法流程图。如图4所示，本实施例的方法包括：FIG. 4 is a flowchart of a method for obtaining a preset probability prediction model provided by an embodiment of the present application. As shown in Figure 4, the method of this embodiment includes:

S401、获取第一训练样本集和第二训练样本集。S401. Obtain a first training sample set and a second training sample set.

本步骤中，由于本申请是通过预设概率预估模型，预估通过目标接触方式接触客户后，客户在预设时长内针对目标产品转化的概率。因此，需要将通过目标方式接触过的客作为样本户，将样本客户的信息作为训练样本。In this step, since the application uses a preset probability estimation model to estimate the probability that the customer will convert to the target product within a preset time period after contacting the customer through the target contact method. Therefore, it is necessary to take the customers who have been in contact with the target method as the sample households, and the information of the sample customers as the training samples.

根据通过电销方式接触过的样本客户，获取每个样本客户的属性信息、以所述目标接触方式接触所述样本客户的时间点以及第一标签，从而获得训练样本集，即第一训练样本集。According to the sample customers who have been contacted through telemarketing, obtain the attribute information of each sample customer, the time point of contacting the sample customer in the target contact method, and the first label, so as to obtain a training sample set, that is, the first training sample set.

其中，样本客户的属性信息包括以下至少一项：样本客户所属的行业信息、提款信息、样本客户的名称、样本客户的地址。以电销接触方式接触样本客户的时间例如可以为最近一次业务人员通过电销接触方式接触样本客户的日期，例如，业务人员2020年8月30日电话联系样本客户，则以电销接触方式接触样本客户的时间为2020年8月30日。The attribute information of the sample customer includes at least one of the following: industry information to which the sample customer belongs, withdrawal information, the name of the sample customer, and the address of the sample customer. The time of contacting the sample customer by means of telemarketing contact can be, for example, the date when the salesperson contacted the sample customer by means of telemarketing contact. The time for the sample customer is August 30, 2020.

每个样本客户的标签用于表示以电销接触方式接触样本客户的时间点起，在预设时长内样本客户是否办理贷款业务。例如，业务人员2020年8月30日电话联系样本客户，业务人员将贷款产品介绍给样本客户，如果样本客户了解贷款产品后，以2020年8月30日为开始时间，在30天内办理贷款业务，则说明样本客户成功转化，此时，标签的值例如可以为1；否则，样本客户未成功转化，此时，标签的值例如可以为0。The label of each sample customer is used to indicate whether the sample customer handles loan business within a preset period of time from the time when the sample customer is contacted by telemarketing. For example, the business staff contacted the sample customer by phone on August 30, 2020, and the business staff introduced the loan product to the sample customer. If the sample customer understands the loan product, the start time will be August 30, 2020, and the loan business will be processed within 30 days. , it means that the sample customer is successfully converted, and the value of the tag can be, for example, 1; otherwise, the sample customer has not been successfully converted, and the value of the tag can be 0, for example.

可选的，由于通过电销方式接触的客户的数量有限，因此，获取到的通过电销方式接触的客户的属性信息较少，即通过第一训练样本集提取到的信息量较少。例如，电销方式接触到的行业较少，行业覆盖率低，难以挖掘出具有贷款意愿的行业。这样，训练获得的预设概率预估模型在计算预估概率时，模型质量差，准确性较低。因此，为了提高模型的质量，可以利用未通过电销接触方式接触的样本客户来弥补通过电销方式接触的客户的数量少的问题，即获取未通过电销接触方式接触的样本客户，根据未通过电销接触方式接触的样本客户，获得训练样本集，即第二训练样本集。Optionally, since the number of customers contacted by telemarketing is limited, the acquired attribute information of customers contacted by telemarketing is less, that is, the amount of information extracted from the first training sample set is small. For example, the telemarketing method has less exposure to industries, and the industry coverage rate is low, making it difficult to find industries that are willing to lend. In this way, when the preset probability prediction model obtained by training is used to calculate the estimated probability, the model quality is poor and the accuracy is low. Therefore, in order to improve the quality of the model, it is possible to use the sample customers who are not contacted by electrical pins to make up for the problem of the small number of customers contacted by electrical pins, that is, to obtain sample customers who are not contacted by electrical pins. The sample customers contacted by means of telepin contact obtain a training sample set, that is, the second training sample set.

第二训练样本集中包括每个未通过电销接触方式接触的样本客户的属性信息、预设时间点以及第二标签。对于属性信息，其包括以下至少一项：样本客户所属的行业信息、样本客户的名称、样本客户的地址。The second training sample set includes attribute information, a preset time point, and a second label of each sample customer who is not contacted by electrical pins. For attribute information, it includes at least one of the following: industry information to which the sample customer belongs, the name of the sample customer, and the address of the sample customer.

预设时间点为预先设置的时间点，可以为任一时间点，例如，下一个周一，或者当天所对应的日期等。The preset time point is a preset time point, which may be any time point, for example, the next Monday, or the date corresponding to the current day.

第二标签用于表示以预设时间点起，未通过电销方式接触过的样本客户在预设时长内是否办理贷款业务。例如，预设时间点为当天所对应的日期，则从当天开始，如果客户在预设时长内自己办理贷款业务，则说明样本客户成功转化，此时，标签的值例如可以为1；否则，样本客户未成功转化，此时，标签的值例如可以为0。一般的，预设时长为30天。The second label is used to indicate whether the sample customers who have not been contacted by telemarketing have applied for loan business within the preset time period from the preset time point. For example, if the preset time point is the date corresponding to the current day, starting from the current day, if the customer handles the loan business by himself within the preset time period, it means that the sample customer has been successfully converted. At this time, the value of the tag can be, for example, 1; otherwise, The sample customer has not been successfully converted. In this case, the value of the tag can be, for example, 0. Generally, the preset duration is 30 days.

例如，对于未通过电销接触方式接触的样本客户，以2020年9月1日为预设时间点，也就是说2020年9月1日为开始时间，预设时长例如为30天。在2020年9月1日-2020年9月30日，业务人员没有对该样本客户通过电销接触方式接触的前提下，如果该样本客户在2020年9月1日-2020年9月30日中的任一天办理了贷款业务，则说明样本客户成功转化，此时，标签的值例如可以为1；否则，样本客户为成功转化，此时，标签的值例如可以为0。For example, for a sample customer who has not been contacted through telemarketing contact, September 1, 2020 is the preset time point, that is, September 1, 2020 is the start time, and the preset duration is, for example, 30 days. On the premise that from September 1, 2020 to September 30, 2020, the business personnel did not contact the sample customer through telemarketing contact, if the sample customer was contacted from September 1, 2020 to September 30, 2020 If the loan business is processed on any of the days, it means that the sample customer has been successfully converted. In this case, the value of the tag can be, for example, 1; otherwise, the sample customer is successfully converted, and the value of the tag can be 0, for example.

可选的，由于预估概率对应的是通过电销接触方式接触待筛选客户后，待筛选客户在预设时长内办理贷款业务的概率。也就是说在通过电销方式接触客户后，需要经过预设时长才能否确定客户是否办理贷款业务。因此，以电销接触方式接触样本客户的时间点与获取客户的属性信息的时间点之间间隔的时长需要大于或等于预设时长。Optionally, since the estimated probability corresponds to the probability that the to-be-screened customer will handle the loan business within a preset time period after contacting the to-be-screened customer through telemarketing contact. That is to say, after contacting a customer through telemarketing, it takes a preset period of time to determine whether the customer has applied for a loan business. Therefore, the time interval between the time point when the sample customer is contacted with the electric pin and the time point when the attribute information of the customer is acquired needs to be greater than or equal to the preset time period.

例如，获取客户的属性信息的时间点为2020年9月1日，预设时长为30，则以电销接触方式接触样本客户的时间点最迟应该是2020年8月2日。如果选取的以电销接触方式接触样本客户的时间点为2020年8月2日之后的日期，例如，以电销接触方式接触样本客户的时间点为2020年8月4日，根据预设时长，在2020年9月1日时，还未到达预设时长，也就是说样本客户在2020年9月1日时还未购买产品，但是，样本客户在2020年9月2日或者2020年9月3日有可能会购买产品。因此，在2020年9月1日时还不确定该客户是否针对贷款产品进行贷款。因此，如果将2020年8月4日以电销接触方式接触客户作为样本客户，导致训练结果不准确，即通过预设概率预估模型获得的预估概率不准确。For example, if the time point for obtaining customer attribute information is September 1, 2020, and the preset time period is 30, the time point for contacting sample customers through telemarketing contact should be August 2, 2020 at the latest. If the selected time point to contact sample customers by telemarketing is a date after August 2, 2020, for example, the time point to contact sample customers by telemarketing contact is August 4, 2020, according to the preset duration , on September 1, 2020, the preset time has not been reached, that is to say, the sample customer has not purchased the product on September 1, 2020, but the sample customer has not purchased the product on September 2, 2020 or September 2020. May 3rd to purchase products. Therefore, as of September 1, 2020, it is not certain whether the customer is lending against the loan product. Therefore, if the customer contacted by telemarketing on August 4, 2020 is used as a sample customer, the training result will be inaccurate, that is, the estimated probability obtained by the preset probability estimation model will be inaccurate.

可选的，同理可知，预设时间点与获取待筛选客户的第一特征的时间点之间间隔的时长需要大于或等于预设时长。Optionally, in the same way, it can be known that the time interval between the preset time point and the time point at which the first feature of the customer to be screened is obtained needs to be greater than or equal to the preset time length.

可选的，S401中获取第二训练样本集的可能的实现方式为：Optionally, a possible implementation manner of obtaining the second training sample set in S401 is:

在同一个预设时间点获取未通过目标接触方式接触的不同样本客户中每个样本客户的属性信息以及第二标签；或者，Obtain the attribute information and the second label of each sample customer among different sample customers who have not been contacted by the target contact at the same preset time point; or,

在多个不同预设时间点分别获取未通过目标接触方式接触的同一样本客户的属性信息以及第二标签。The attribute information and the second label of the same sample customer who have not been contacted by the target contact method are respectively acquired at a plurality of different preset time points.

具体的，为了增加训练样本的数量，提高通过预设概率预估模型获取的预估概率的准确性，在获取第二训练样本集时，设置不同的预设时间点，分别获取与每个预设时间点对应的未通过电销方式接触的样本客户。例如，预设时间点可以为每周一，在第一个周一时，获取一次未通过电销方式接触的样本客户，在第二个周一时，获取一次未通过电销方式接触的样本客户，依次获取对每个周一对应的未通过电销方式接触的样本客户。Specifically, in order to increase the number of training samples and improve the accuracy of the estimated probability obtained through the preset probability prediction model, when obtaining the second training sample set, different preset time points are set, and the corresponding Set up sample customers who have not been contacted by telemarketing at the corresponding time point. For example, the preset time point can be every Monday. On the first Monday, a sample customer who has not been contacted through telemarketing is obtained, and on the second Monday, a sample customer who has not been contacted through telemarketing is obtained. Get sample customers who have not been contacted by telemarketing for each Monday.

其中，对于同一未通过电销方式接触的样本客户，以其中一个周一为开始时间，在预设时长内可能没有针对贷款产品进行贷款，但是，在以下一个周一为开始时间，在预设时长内可能会针对贷款产品进行贷款。例如，对于一个未通过电销方式接触的样本客户，当以预设时间点2020年8月3日为开始时间，在30天内，即在2020年9月1日之前没有针对贷款产品进行贷款，但是，该样本客户在2020年9月4日针对贷款产品进行贷款，因此，当以预设时间点2020年8月10日为开始时间，在30天内，即在2020年9月8日之前针对贷款产品进行贷款。Among them, for the same sample customer who has not been contacted through telemarketing, one of the Mondays is the starting time, and there may be no loans for loan products within the preset time period. However, the next Monday is the starting time, within the preset time period. Lending may be made against loan products. For example, for a sample customer who has not been contacted through telemarketing, when starting from the preset time point August 3, 2020, within 30 days, that is, before September 1, 2020, there is no loan for the loan product, However, the sample customer made a loan for the loan product on September 4, 2020. Therefore, when starting from the preset time point August 10, 2020, within 30 days, that is, before September 8, 2020 Loan products for lending.

需要说明的是，第二训练样本集中的所有预设时间点包含多个不同的时间点时，任意相邻两个预设时间点之间的间隔可以相同，也可以不同。例如，每隔固定时长获取不同的预设时间点，或者，根据需求选择不同的时间点作为预设时间点，本申请对此不限制。It should be noted that when all the preset time points in the second training sample set include multiple different time points, the interval between any two adjacent preset time points may be the same or different. For example, different preset time points may be acquired at fixed time intervals, or different time points may be selected as preset time points according to requirements, which is not limited in this application.

S402、根据第一训练样本集和第二训练样本集，训练初始的概率预估模型，获得预设概率预估模型。S402. Train an initial probability prediction model according to the first training sample set and the second training sample set to obtain a preset probability prediction model.

本步骤中，初始的概率预估模型例如为LGBM模型。In this step, the initial probability prediction model is, for example, an LGBM model.

如图5所示，将第一训练样本集的属性信息、以所述目标接触方式接触所述样本客户的时间点以及第一标签，以及第二训练样本集的属性信息、预设时间点和第二标签输入到LGBM模型中，对LGBM模型进行训练，获得预设概率预估模型。As shown in FIG. 5 , the attribute information of the first training sample set, the time point of contacting the sample customer in the target contact manner and the first label, and the attribute information of the second training sample set, the preset time point and The second label is input into the LGBM model, and the LGBM model is trained to obtain a preset probability prediction model.

可选的，S402的一种可能的实现方式为：Optionally, a possible implementation manner of S402 is:

S4021、根据第二训练样本集，训练初始的概率预估模型，获得中间预设概率预估模型；S4021. Train an initial probability prediction model according to the second training sample set, and obtain an intermediate preset probability prediction model;

本步骤中，由于未通过电销方式接触的样本客户的数量多于通过电销方式接触的样本客户的数量。因此，首先，将第二训练样本集中各样本客户的属性信息、预设时间点和第二标签输入到LGBM模型，对LGBM模型进行训练。In this step, the number of sample customers who are not contacted by means of telemarketing is more than the number of sample customers who are contacted by means of telemarketing. Therefore, first, the attribute information, preset time point and second label of each sample customer in the second training sample set are input into the LGBM model to train the LGBM model.

其中，训练时，对于每个样本客户，LGBM模型根据该样本客户的属性信息，计算样本客户未通过电销方式接触时在预设时长内是否办理贷款业务，将计算结果与该样本客户对应的第二标签进行比较，根据比较结果训练LGBM模型，获得中间预设概率预估模型。Among them, during training, for each sample customer, the LGBM model calculates whether the sample customer handles loan business within a preset period of time when the sample customer is not contacted by telemarketing according to the attribute information of the sample customer, and compares the calculation result with the corresponding sample customer. The second tag is compared, and the LGBM model is trained according to the comparison result to obtain an intermediate preset probability prediction model.

S4022、根据第一训练样本集，训练中间预设概率预估模型，获得预设概率预估模型。S4022. Train an intermediate preset probability prediction model according to the first training sample set to obtain a preset probability prediction model.

本步骤中，通过第二训练样本集训练获得的中间预设概率预估模型，可以用于预测未通过电销方式接触客户时，在预设时长内针对贷款产品进行贷款概率高的客户，而本申请需要获取的是通过电销方式接触客户后，在预设时长内针对贷款产品进行贷款概率高的客户。因此，对于中间预设概率预估模型，需要根据第一训练样本集进行训练优化，以使最终获得的预设概率预估模型计算获得的预估概率尽可能的准确。In this step, the intermediate preset probability estimation model obtained through the training of the second training sample set can be used to predict customers who have a high probability of obtaining loans for loan products within a preset period of time when they have not contacted customers through telemarketing. This application needs to acquire customers who have a high probability of lending for loan products within a preset period of time after contacting customers through telemarketing. Therefore, for the intermediate preset probability prediction model, it is necessary to perform training optimization according to the first training sample set, so that the predicted probability obtained by the calculation of the finally obtained preset probability prediction model is as accurate as possible.

其中，根据第一训练样本集中各样本客户，训练中间预设概率预估模型的具体实现方式可参考S4021，此处不再赘述。Wherein, according to each sample customer in the first training sample set, the specific implementation manner of training the intermediate preset probability estimation model may refer to S4021, which will not be repeated here.

可选的，在S402之后，方法还包括：Optionally, after S402, the method further includes:

S403、在对客户通过目标接触方式进行接触后，预设时长内客户针对目标产品未转化，则将客户作为第一训练样本集中的样本客户，获得更新后的第一训练样本集。S403 , after contacting the customer through the target contact method, if the customer has not converted to the target product within a preset time period, the customer is regarded as a sample customer in the first training sample set, and an updated first training sample set is obtained.

本步骤中，通过预设概率预估模型获取的是客户的预估概率，因此，实际中，即使预估概率的值较大，通过电销方式接触该客户后，该客户也可能在预设时长内不办理贷款业务。因此，需要对预设概率预估模型进行持续的训练优化，提高模型的质量。例如，可以将根据预估概率确定的需通过电销接触方式进行接触的，但是在预设时长内没有办理贷款业务的客户添加到第一训练样本集中，作为第一训练样本集中的样本客户，获得更新后的第一训练样本。In this step, the estimated probability of the customer is obtained through the preset probability estimation model. Therefore, in practice, even if the value of the estimated probability is large, after contacting the customer through telemarketing, the customer may still be in the preset probability. The loan business will not be processed for a long period of time. Therefore, it is necessary to continuously train and optimize the preset probability prediction model to improve the quality of the model. For example, customers determined according to the estimated probability who need to be contacted by telemarketing but who have not handled loan business within a preset period of time can be added to the first training sample set as sample customers in the first training sample set, Obtain the updated first training sample.

例如，对于预估概率大于预设概率的客户，在2020年8月1日通过电销方式接触后，如果在到达2020年8月31日后仍然没有办理贷款业务，则将该客户作为第一训练样本集中的样本客户，对预设概率预估模型进行优化。For example, for a customer whose estimated probability is greater than the preset probability, after being contacted by telemarketing on August 1, 2020, if the loan business is still not processed after reaching August 31, 2020, the customer will be used as the first training The sample customers in the sample set optimize the preset probability prediction model.

可选的，S403的一种可能的实现方式为：将客户作为第一训练样本集中的多份样本客户，获得更新后的第一训练样本集。Optionally, a possible implementation manner of S403 is: taking the customer as a plurality of sample customers in the first training sample set, and obtaining the updated first training sample set.

本步骤中，为了提高对预设概率预估模型训练优化后，预设概率预估模型的质量，对于实际转化结果与预估概率有差异的客户，将其作为多份样本客户添加至第一训练样本集中。即根据该客户的属性信息以及标签，多次训练优化预设概率预估模型，从而提高预设概率预估模型对客户的属性信息的分析处理能力，从而使获得的预估概率所反映的客户的转化情况与实际转化情况更一致。In this step, in order to improve the quality of the preset probability prediction model after training and optimization of the preset probability prediction model, for customers whose actual conversion results are different from the estimated probability, they are added as multiple sample customers to the first training sample set. That is, according to the attribute information and labels of the customer, the preset probability prediction model is trained and optimized for many times, so as to improve the analysis and processing ability of the preset probability prediction model on the attribute information of the customer, so that the obtained estimated probability reflects the customer. conversions are more in line with actual conversions.

S404、根据更新后的第一训练样本集，对预设概率预估模型进行优化。S404. Optimize the preset probability prediction model according to the updated first training sample set.

本步骤中，在对预设概率预估模型进行优化时，针对每个样本客户，预设概率预估模型根据该样本客户的属性信息，计算该样本客户的预估概率。该预估概率与该样本客户的标签一致，例如，如果该样本客户的标签指示该样本客户电销方式接触后的预设时长内办理贷款业务，而根据该预估概率，可以说明该样本客户电销方式接触后的预设时长内办理贷款业务。或者，该预估概率与该样本客户的标签不一致，例如，如果该样本客户的标签指示该样本客户电销方式接触后的预设时长内没有办理贷款业务，而根据该预估概率，可以说明该样本客户电销方式接触后的预设时长内办理贷款业务。也就是说，预设概率预估模型输出的预估概率并不能与实际上客户是否办理贷款业务的真实情况之间存在差异。因此，通过预设概率预估模型输出的预估概率与实际上客户是否办理贷款业务的真实情况的差异对预设概率预估模型进行优化。In this step, when optimizing the preset probability estimation model, for each sample customer, the preset probability estimation model calculates the estimated probability of the sample customer according to the attribute information of the sample customer. The estimated probability is consistent with the sample customer's label. For example, if the sample customer's label instructs the sample customer to handle the loan business within a preset period of time after contact with the sample customer via telemarketing, the sample customer can be identified according to the estimated probability. The loan business will be processed within a preset period of time after the contact with the telemarketing method. Or, the estimated probability is inconsistent with the label of the sample customer. For example, if the label of the sample customer indicates that the sample customer has not handled the loan business within a preset period of time after the sample customer is contacted by telemarketing, and according to the estimated probability, it can be explained that The sample customer will handle the loan business within a preset period of time after the contact with the sample customer. That is to say, there is no difference between the estimated probability output by the preset probability estimation model and the actual situation of whether the customer handles the loan business. Therefore, the preset probability prediction model is optimized according to the difference between the estimated probability output by the preset probability prediction model and the actual situation of whether the customer handles the loan business.

具体的，获取预设概率预估模型输出的预估概率与实际上客户是否办理贷款业务的真实情况之间的差异，根据差异和预设概率预估模型的损失函数确定损失值，根据损失值对预设概率预估模型进行优化。Specifically, the difference between the estimated probability output by the preset probability prediction model and the actual situation of whether the customer handles the loan business is obtained, the loss value is determined according to the difference and the loss function of the preset probability prediction model, and the loss value is determined according to the difference and the loss function of the preset probability prediction model. Optimize the preset probability prediction model.

其中，在对客户通过电销接触方式进行接触后，预设时长内客户没有办理贷款业务，说明预设概率预估模型预估的该客户的预估概率不准确，因此，将该客户作为样本客户对预设概率预估模型进行优化。Among them, after the customer is contacted through telemarketing contact, the customer has not handled the loan business within the preset period of time, indicating that the estimated probability of the customer estimated by the preset probability estimation model is inaccurate. Therefore, this customer is used as a sample. The customer optimizes the preset probability prediction model.

其中，为了提高预设概率预估模型的质量，提高这些样本客户的属性信息对预设概率预估模型的影响程度。因此，可以提高这些样本客户对应的损失值的权重，提高预设概率预估模型对客户的属性信息的分析处理能力，从而使获得的预估概率所反映的客户的转化情况与实际转化情况更一致。Among them, in order to improve the quality of the preset probability prediction model, the influence degree of the attribute information of these sample customers on the preset probability prediction model is improved. Therefore, it is possible to increase the weight of the loss value corresponding to these sample customers, and improve the analysis and processing ability of the preset probability estimation model for the customer's attribute information, so that the conversion situation of the customer reflected by the obtained estimated probability is better than the actual conversion situation. Consistent.

需要说明的是，图4实施例中方法的执行主体可以与图2实施例中的执行主体为同一执行主体，或者为不同的执行主体，本申请对此不限制。It should be noted that, the execution body of the method in the embodiment of FIG. 4 may be the same execution body as the execution body in the embodiment of FIG. 2 , or a different execution body, which is not limited in this application.

图6为本申请一实施例提供的数据处理装置的结构示意图。如图6所示，数据处理装置可以包括：FIG. 6 is a schematic structural diagram of a data processing apparatus according to an embodiment of the present application. As shown in Figure 6, the data processing apparatus may include:

获取模块601，用于获取本地存储的客户的属性信息；Obtaining module 601, for obtaining the attribute information of the customer stored locally;

预估模块602，用于将客户的属性信息输入到目标业务的预设概率预估模型中，获得客户的预估概率，预估概率用于表示通过目标接触方式接触客户后，客户在预设时长内针对目标业务转化的概率；The estimation module 602 is used to input the attribute information of the customer into the preset probability estimation model of the target business, and obtain the estimated probability of the customer. The probability of conversion of the target business within the time period;

确定模块603，用于若根据客户的预估概率，确定需通过目标接触方式进行接触客户，则向业务人员的终端发送接触提示信息，接触提示信息用于提示针对目标业务通过目标接触方式接触客户。The determining module 603 is configured to send contact prompt information to the terminal of the business personnel if it is determined that the customer needs to be contacted by the target contact method according to the estimated probability of the customer, and the contact prompt information is used to prompt the target business to contact the customer through the target contact method .

本实施例提供的数据处理装置，可以用于执行前述任一方法实施例提供的技术方案，其实现原理和技术效果类似，通过对客户进行筛选，获取通过目标方式接触后，在预设时长内针对目标业务转化率高的客户，对这些客户通过目标方式接触，减少通过目标方式接触的客户的数量的基础上，提高客户的转化率。The data processing apparatus provided in this embodiment can be used to implement the technical solutions provided by any of the foregoing method embodiments. The implementation principle and technical effect are similar. For customers with a high conversion rate of the target business, contact these customers through targeted methods, and on the basis of reducing the number of customers contacted through targeted methods, the conversion rate of customers is improved.

在一种可能的实现方式中，预设概率预估模型是根据第一训练样本集和第二训练样本集训练获得的；In a possible implementation manner, the preset probability estimation model is obtained by training according to the first training sample set and the second training sample set;

第一训练样本集包含：通过目标接触方式接触的样本客户的属性信息、以目标接触方式接触样本客户的时间点以及第一标签，第一标签用于表示以目标接触方式接触样本客户的时间点起预设时长内是否针对目标产品转化；The first training sample set includes: attribute information of the sample customers contacted by the target contact method, a time point of contacting the sample customers by the target contact method, and a first label, where the first label is used to indicate the time point of contacting the sample customers by the target contact method. Whether the target product has been converted within a preset period of time;

第二训练样本集包含：未通过目标接触方式接触的样本客户的属性信息、预设时间点以及第二标签，第二标签用于表示以预设时间点起预设时长内是否针对目标产品转化。The second training sample set includes: attribute information of sample customers who have not been contacted through the target contact method, a preset time point, and a second label. The second label is used to indicate whether the target product has been converted within a preset time period from the preset time point. .

在一种可能的实现方式中，以目标接触方式接触样本客户的时间点与获取客户的属性信息的时间点之间的间隔时长大于或等于预设时长；In a possible implementation manner, the interval time between the time point of contacting the sample customer in the target contact mode and the time point of acquiring the attribute information of the customer is greater than or equal to a preset time period;

预设时间点与获取客户的属性信息的时间点之间的间隔时长大于或等于预设时长。The interval time between the preset time point and the time point of acquiring the attribute information of the customer is greater than or equal to the preset time length.

在一种可能的实现方式中，数据处理装置还包括：训练模块604；In a possible implementation manner, the data processing apparatus further includes: a training module 604;

获取模块601，还用于获取第一训练样本集和第二训练样本集；The acquisition module 601 is further configured to acquire a first training sample set and a second training sample set;

训练模块604，用于根据第一训练样本集和第二训练样本集，训练初始的概率预估模型，获得预设概率预估模型。The training module 604 is configured to train an initial probability prediction model according to the first training sample set and the second training sample set to obtain a preset probability prediction model.

在一种可能的实现方式中，训练模块604根据第一训练样本集和第二训练样本集，训练初始的概率预估模型，获得预设概率预估模型时，具体用于：In a possible implementation manner, the training module 604 trains an initial probability prediction model according to the first training sample set and the second training sample set, and when obtaining a preset probability prediction model, it is specifically used for:

根据第二训练样本集，训练初始的概率预估模型，获得中间预设概率预估模型；According to the second training sample set, the initial probability prediction model is trained, and the intermediate preset probability prediction model is obtained;

根据第一训练样本集，训练中间预设概率预估模型，获得预设概率预估模型。According to the first training sample set, an intermediate preset probability prediction model is trained to obtain a preset probability prediction model.

在一种可能的实现方式中，获取模块601获取第二训练样本集，具体用于：In a possible implementation manner, the obtaining module 601 obtains the second training sample set, which is specifically used for:

在一种可能的实现方式中，获取模块601，还用于：In a possible implementation manner, the acquiring module 601 is further configured to:

在对客户通过目标接触方式进行接触后，预设时长内客户针对目标产品未转化，则将客户作为第一训练样本集中的样本客户，获得更新后的第一训练样本集；After the customer is contacted through the target contact method, if the customer does not convert the target product within the preset time period, the customer is regarded as the sample customer in the first training sample set, and the updated first training sample set is obtained;

训练模块604，还用于：The training module 604 is also used for:

根据更新后的第一训练样本集，对预设概率预估模型进行优化。According to the updated first training sample set, the preset probability prediction model is optimized.

在一种可能的实现方式中，获取模块601将客户作为第一训练样本集中的样本客户，获得更新后的第一训练样本集时，具体用于：In a possible implementation manner, the obtaining module 601 takes the customer as a sample customer in the first training sample set, and when obtaining the updated first training sample set, it is specifically used for:

将客户作为第一训练样本集中的多份样本客户，获得更新后的第一训练样本集。Taking the customer as a plurality of sample customers in the first training sample set, the updated first training sample set is obtained.

在一种可能的实现方式中，属性信息包括如下至少一项：所属的行业信息、名称、地址、提款信息。In a possible implementation manner, the attribute information includes at least one of the following: industry information, name, address, and withdrawal information.

前述任一实施例提供的数据处理装置，用于执行前述任一方法实施例的技术方案，其实现原理和技术效果类似，在此不再赘述。The data processing apparatus provided in any of the foregoing embodiments is used to execute the technical solutions of any of the foregoing method embodiments, and the implementation principles and technical effects thereof are similar, and details are not described herein again.

图7为本申请一实施例提供的电子设备的结构示意图，如图7所示，该电子设备包括：存储器72、处理器71及存储在所述存储器72上并可在所述处理器71上运行的计算机程序，所述计算机程序被所述处理器71执行时实现前述任一方法实施例提供的数据处理方法的步骤。FIG. 7 is a schematic structural diagram of an electronic device provided by an embodiment of the present application. As shown in FIG. 7 , the electronic device includes: a memory 72 , a processor 71 , and a memory 72 and a processor 71 that are stored on the memory 72 and can be stored on the processor 71 . A running computer program, when the computer program is executed by the processor 71, implements the steps of the data processing method provided by any of the foregoing method embodiments.

可选的，电子设备还可以包括显示器73。Optionally, the electronic device may further include a display 73 .

该电子设备的上述各个器件之间可以通过总线连接。The above-mentioned various components of the electronic device can be connected through a bus.

存储器72可以是单独的存储单元，也可以是集成在处理器71中的存储单元。处理器71的数量为一个或者多个。The memory 72 may be a separate storage unit, or may be a storage unit integrated in the processor 71 . The number of processors 71 is one or more.

在上述在电子设备的实现中，存储器和处理器之间直接或间接地电性连接，以实现数据的传输或交互，也就是存储器和处理器可以通过接口连接，也可以集成在一起。例如，这些元件相互之间可以通过一条或者多条通信总线或信号线实现电性连接，如可以通过总线连接。存储器可以是，但不限于，随机存取存储器(Random Access Memory，简称：RAM)，只读存储器(Read Only Memory，简称：ROM)，可编程只读存储器(Programmable Read-Only Memory，简称：PROM)，可擦除只读存储器(Erasable Programmable Read-Only Memory，简称：EPROM)，电可擦除只读存储器(Electric Erasable Programmable Read-Only Memory，简称：EEPROM)等。其中，存储器用于存储程序，处理器在接收到执行指令后，执行程序。进一步地，上述存储器内的软件程序以及模块还可包括操作***，其可包括各种用于管理***任务(例如内存管理、存储设备控制、电源管理等)的软件组件和/或驱动，并可与各种硬件或软件组件相互通信，从而提供其他软件组件的运行环境。In the above implementation of the electronic device, the memory and the processor are electrically connected directly or indirectly to realize data transmission or interaction, that is, the memory and the processor can be connected through an interface or integrated together. For example, these elements can be electrically connected to each other through one or more communication buses or signal lines, such as can be connected through a bus. The memory can be, but is not limited to, random access memory (Random Access Memory) Access Memory, referred to as: RAM), read-only memory (Read Only Memory, referred to as: ROM), Programmable Read-Only Memory (Programmable Read-Only Memory) Read-Only Memory, referred to as: PROM), Erasable Programmable Read-Only Memory (Erasable Programmable Read-Only Memory, referred to as: EPROM), Electrically Erasable Read-Only Memory (Electric Erasable Programmable Read-Only Memory, referred to as: EEPROM) and so on. The memory is used to store the program, and the processor executes the program after receiving the execution instruction. Further, the software programs and modules in the above-mentioned memory may also include an operating system, which may include various software components and/or drivers for managing system tasks (such as memory management, storage device control, power management, etc.), and may Intercommunicate with various hardware or software components to provide the operating environment for other software components.

处理器可以是一种集成电路芯片，具有信号的处理能力。上述的处理器可以是通用处理器，包括中央处理器(Central Processing Unit，简称：CPU)、图像处理器等，可以实现或者执行本申请实施例中的公开的各方法、步骤及逻辑框图。The processor may be an integrated circuit chip with signal processing capability. The above-mentioned processor can be a general-purpose processor, including a central processing unit (Central A Processing Unit, CPU for short), an image processor, etc., can implement or execute the methods, steps, and logical block diagrams disclosed in the embodiments of the present application.

本申请还提供一种计算机可读存储介质，所述计算机可读存储介质上存储有计算机程序，所述计算机程序被处理器执行时实现如前述任一方法实施例提供的数据处理方法的步骤The present application also provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, implements the steps of the data processing method provided by any of the foregoing method embodiments

需要说明的是，在本文中，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者装置不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者装置所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括该要素的过程、方法、物品或者装置中还存在另外的相同要素。It should be noted that, herein, the terms "comprising", "comprising" or any other variation thereof are intended to encompass non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements, It also includes other elements not expressly listed or inherent to such a process, method, article or apparatus. Without further limitation, an element qualified by the phrase "comprising a..." does not preclude the presence of additional identical elements in a process, method, article or apparatus that includes the element.

上述本申请实施例序号仅仅为了描述，不代表实施例的优劣。The above-mentioned serial numbers of the embodiments of the present application are only for description, and do not represent the advantages or disadvantages of the embodiments.

通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到上述实施例方法可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件，但很多情况下前者是更佳的实施方式。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质（如ROM/RAM、磁碟、光盘）中，包括若干指令用以使得一台终端设备（可以是手机，计算机，服务器，空调器，或者网络设备等）执行本申请各个实施例所述的方法。From the description of the above embodiments, those skilled in the art can clearly understand that the method of the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is better implementation. Based on this understanding, the technical solution of the present application can be embodied in the form of a software product in essence or the part that contributes to the prior art, and the computer software product is stored in a storage medium (such as ROM/RAM, magnetic disk, CD-ROM), including several instructions to make a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) execute the methods described in the various embodiments of this application.

以上仅为本申请的优选实施例，并非因此限制本申请的专利范围，凡是利用本申请说明书及附图内容所作的等效结构或等效流程变换，或直接或间接运用在其他相关的技术领域，均同理包括在本申请的专利保护范围内。The above are only the preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made by using the contents of the description and drawings of the present application, or directly or indirectly applied in other related technical fields , are similarly included within the scope of patent protection of this application.

Claims

一种数据处理方法，其特征在于，包括：A data processing method, comprising:

获取本地存储的客户的属性信息；Get the attribute information of the customer stored locally;

将所述客户的属性信息输入到目标业务的预设概率预估模型中，获得所述客户的预估概率，所述预估概率用于表示通过目标接触方式接触所述客户后，所述客户在预设时长内针对所述目标业务转化的概率；Input the attribute information of the customer into the preset probability estimation model of the target business, and obtain the estimated probability of the customer, and the estimated probability is used to indicate that after the customer is contacted by the target contact method, the customer The probability of conversion for the target business within a preset time period;

若根据所述客户的预估概率，确定需通过目标接触方式进行接触所述客户，则向业务人员的终端发送接触提示信息，所述接触提示信息用于提示针对所述目标业务通过目标接触方式接触所述客户。If, according to the estimated probability of the customer, it is determined that the customer needs to be contacted by the target contact method, contact prompt information is sent to the terminal of the business personnel, and the contact prompt information is used to prompt the target business to be contacted by the target contact method. Contact said customer.
根据权利要求1所述的方法，其特征在于，所述预设概率预估模型是根据第一训练样本集和第二训练样本集训练获得的；The method according to claim 1, wherein the preset probability estimation model is obtained by training according to the first training sample set and the second training sample set;

所述第一训练样本集包含：通过目标接触方式接触的样本客户的属性信息、以所述目标接触方式接触所述样本客户的时间点以及第一标签，所述第一标签用于表示以所述目标接触方式接触所述样本客户的时间点起预设时长内是否针对所述目标产品转化；The first training sample set includes: attribute information of sample customers contacted by the target contact method, time points of contacting the sample customers by the target contact method, and a first label, where the first label is used to represent Whether the target product is converted within a preset time period from the time when the target contact method contacts the sample customer;

所述第二训练样本集包含：未通过目标接触方式接触的样本客户的属性信息、预设时间点以及第二标签，所述第二标签用于表示以预设时间点起所述预设时长内是否针对所述目标产品转化。The second training sample set includes: attribute information of sample customers who have not been contacted by the target contact method, a preset time point, and a second label, where the second label is used to indicate the preset duration from the preset time point whether to convert for the target product.
根据权利要求2所述的方法，其特征在于，所述以所述目标接触方式接触所述样本客户的时间点与所述获取所述客户的属性信息的时间点之间的间隔时长大于或等于预设时长；The method according to claim 2, wherein the time interval between the time point of contacting the sample customer in the target contact manner and the time point of acquiring the attribute information of the customer is greater than or equal to preset duration;

所述预设时间点与所述获取所述客户的属性信息的时间点之间的间隔时长大于或等于预设时长。The interval duration between the preset time point and the time point of acquiring the attribute information of the customer is greater than or equal to the preset time duration.
根据权利要求2所述的方法，其特征在于，所述将所述客户的第一特征输入到目标产品的预设概率预估模型中之前，还包括：The method according to claim 2, wherein before the inputting the first characteristic of the customer into the preset probability estimation model of the target product, the method further comprises:

获取所述第一训练样本集和所述第二训练样本集；obtaining the first training sample set and the second training sample set;

根据所述第一训练样本集和所述第二训练样本集，训练初始的概率预估模型，获得所述预设概率预估模型。According to the first training sample set and the second training sample set, an initial probability prediction model is trained to obtain the preset probability prediction model.
根据权利要求4所述的方法，其特征在于，所述根据第一训练样本集和第二训练样本集，训练初始的概率预估模型，获得所述预设概率预估模型，包括：The method according to claim 4, wherein, according to the first training sample set and the second training sample set, training an initial probability prediction model to obtain the preset probability prediction model, comprising:

根据所述第二训练样本集，训练所述初始的概率预估模型，获得中间预设概率预估模型；According to the second training sample set, the initial probability prediction model is trained to obtain an intermediate preset probability prediction model;

根据所述第一训练样本集，训练所述中间预设概率预估模型，获得所述预设概率预估模型。According to the first training sample set, the intermediate preset probability prediction model is trained to obtain the preset probability prediction model.
根据权利要求4所述的方法，其特征在于，获取所述第二训练样本集，包括：The method according to claim 4, wherein obtaining the second training sample set comprises:

在同一个预设时间点获取未通过目标接触方式接触的不同样本客户中每个样本客户的属性信息以及所述第二标签。At the same preset time point, the attribute information and the second label of each sample customer among the different sample customers who are not contacted by the target contact method are acquired.
根据权利要求4所述的方法，其特征在于，获取所述第二训练样本集，包括：The method according to claim 4, wherein obtaining the second training sample set comprises:

在多个不同预设时间点分别获取未通过目标接触方式接触的同一样本客户的属性信息以及所述第二标签。The attribute information and the second label of the same sample customer who have not been contacted by the target contact method are respectively acquired at multiple different preset time points.
根据权利要求5至7任一项所述的方法，其特征在于，还包括：The method according to any one of claims 5 to 7, further comprising:

在对所述客户通过目标接触方式进行接触后，预设时长内所述客户针对所述目标产品未转化，则将所述客户作为第一训练样本集中的样本客户，获得更新后的第一训练样本集；After the customer is contacted by the target contact method, if the customer has not converted to the target product within the preset time period, the customer will be regarded as a sample customer in the first training sample set, and the updated first training will be obtained. sample set;

根据更新后的第一训练样本集，对所述预设概率预估模型进行优化。The preset probability prediction model is optimized according to the updated first training sample set.
根据权利要求8所述的方法，其特征在于，所述将所述客户作为第一训练样本集中的样本客户，获得更新后的第一训练样本集，包括：The method according to claim 8, wherein the obtaining the updated first training sample set by using the customer as a sample customer in the first training sample set comprises:

将所述客户作为第一训练样本集中的多份样本客户，获得更新后的第一训练样本集。Taking the customer as a plurality of sample customers in the first training sample set, the updated first training sample set is obtained.
根据权利要求1-7任一项所述的方法，其特征在于，根据所述客户的预估概率，确定需通过目标接触方式进行接触所述客户，包括：The method according to any one of claims 1-7, wherein, according to the estimated probability of the customer, it is determined that the customer needs to be contacted by a target contact method, comprising:

若所述客户的预估概率大于预设概率，则确定需通过目标接触方式进行接触所述客户。If the estimated probability of the customer is greater than the preset probability, it is determined that the customer needs to be contacted by a target contact method.
根据权利要求1-7任一项所述的方法，其特征在于，所述属性信息包括如下至少一项：所属的行业信息、名称、地址、提款信息。The method according to any one of claims 1-7, wherein the attribute information includes at least one of the following: industry information, name, address, and withdrawal information.
一种数据处理装置，其特征在于，包括：A data processing device, comprising:

获取模块，用于获取本地存储的客户的属性信息；The acquisition module is used to acquire the attribute information of the customer stored locally;

预估模块，用于将所述客户的属性信息输入到目标业务的预设概率预估模型中，获得所述客户的预估概率，所述预估概率用于表示通过目标接触方式接触所述客户后，所述客户在预设时长内针对所述目标业务转化的概率；The estimation module is used to input the attribute information of the customer into the preset probability estimation model of the target business, and obtain the estimated probability of the customer, and the estimated probability is used to indicate that the target contact method is used to contact the After the customer, the probability that the customer will convert to the target business within a preset time period;

确定模块，若根据所述客户的预估概率，确定需通过目标接触方式进行接触所述客户，则向业务人员的终端发送接触提示信息，所述接触提示信息用于提示针对所述目标业务通过目标接触方式接触所述客户。The determining module, if it is determined according to the estimated probability of the customer, that the customer needs to be contacted by the target contact method, then send contact prompt information to the terminal of the business personnel, and the contact prompt information is used to prompt the target business to pass through. The target contact method reaches the customer.
根据权利要求12所述的装置，其特征在于，所述预设概率预估模型是根据第一训练样本集和第二训练样本集训练获得的；The device according to claim 12, wherein the preset probability estimation model is obtained by training according to the first training sample set and the second training sample set;

所述第一训练样本集包含：通过目标接触方式接触的样本客户的属性信息、以所述目标接触方式接触所述样本客户的时间点以及第一标签，所述第一标签用于表示以所述目标接触方式接触所述样本客户的时间点起预设时长内是否针对所述目标产品转化；The first training sample set includes: attribute information of sample customers contacted by the target contact method, time points of contacting the sample customers by the target contact method, and a first label, where the first label is used to represent Whether the target product is converted within a preset time period from the time when the target contact method contacts the sample customer;

所述第二训练样本集包含：未通过目标接触方式接触的样本客户的属性信息、预设时间点以及第二标签，所述第二标签用于表示以预设时间点起所述预设时长内是否针对所述目标产品转化。The second training sample set includes: attribute information of sample customers who have not been contacted by the target contact method, a preset time point, and a second label, where the second label is used to indicate the preset duration from the preset time point whether to convert for the target product.
根据权利要求13所述的装置，其特征在于，所述以所述目标接触方式接触所述样本客户的时间点与所述获取所述客户的属性信息的时间点之间的间隔时长大于或等于预设时长；The device according to claim 13, wherein an interval duration between the time point when the sample customer is contacted in the target contact manner and the time point when the attribute information of the customer is acquired is greater than or equal to preset duration;

所述预设时间点与所述获取所述客户的属性信息的时间点之间的间隔时长大于或等于预设时长。The interval duration between the preset time point and the time point of acquiring the attribute information of the customer is greater than or equal to the preset time duration.
根据权利要求13所述的装置，其特征在于，所述装置还包括：The apparatus of claim 13, wherein the apparatus further comprises:

训练模块，用于获取所述第一训练样本集和所述第二训练样本集，根据所述第一训练样本集和所述第二训练样本集，训练初始的概率预估模型，获得所述预设概率预估模型。A training module, configured to obtain the first training sample set and the second training sample set, train an initial probability prediction model according to the first training sample set and the second training sample set, and obtain the Preset probability prediction models.
根据权利要求15所述的装置，其特征在于，所述训练模块具体用于：The apparatus according to claim 15, wherein the training module is specifically used for:

根据所述第二训练样本集，训练所述初始的概率预估模型，获得中间预设概率预估模型；According to the second training sample set, the initial probability prediction model is trained to obtain an intermediate preset probability prediction model;

根据所述第一训练样本集，训练所述中间预设概率预估模型，获得所述预设概率预估模型。According to the first training sample set, the intermediate preset probability prediction model is trained to obtain the preset probability prediction model.
根据权利要求15所述的装置，其特征在于，所述训练模块具体用于：The apparatus according to claim 15, wherein the training module is specifically used for:

在同一个预设时间点获取未通过目标接触方式接触的不同样本客户中每个样本客户的属性信息以及所述第二标签。At the same preset time point, the attribute information and the second label of each sample customer among the different sample customers who are not contacted by the target contact method are acquired.
一种电子设备，其特征在于，所述电子设备包括：存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述计算机程序被所述处理器执行时实现如权利要求1至11中任一项所述的数据处理方法的步骤。An electronic device, characterized in that the electronic device comprises: a memory, a processor, and a computer program stored on the memory and executable on the processor, when the computer program is executed by the processor Steps for implementing a data processing method as claimed in any one of claims 1 to 11.
一种计算机可读存储介质，其特征在于，所述计算机可读存储介质上存储有计算机程序，所述计算机程序被处理器执行时实现如权利要求1至11中任一项所述的数据处理方法的步骤。A computer-readable storage medium, characterized in that a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the data processing according to any one of claims 1 to 11 is realized steps of the method.
一种计算机程序产品，其特征在于，包括计算机程序，所述计算机程序被处理器执行时实现如权利要求1至11中任一项所述的数据处理方法的步骤。A computer program product, characterized in that it includes a computer program, which implements the steps of the data processing method according to any one of claims 1 to 11 when the computer program is executed by a processor.