WO2019062021A1 - 在应用程序中推送贷款广告的方法、电子装置及介质 - Google Patents

在应用程序中推送贷款广告的方法、电子装置及介质 Download PDF

Info

Publication number
WO2019062021A1
WO2019062021A1 PCT/CN2018/077677 CN2018077677W WO2019062021A1 WO 2019062021 A1 WO2019062021 A1 WO 2019062021A1 CN 2018077677 W CN2018077677 W CN 2018077677W WO 2019062021 A1 WO2019062021 A1 WO 2019062021A1
Authority
WO
WIPO (PCT)
Prior art keywords
user
loan
application
decision tree
tree model
Prior art date
Application number
PCT/CN2018/077677
Other languages
English (en)
French (fr)
Inventor
金新
王建明
肖京
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 WO2019062021A1 publication Critical patent/WO2019062021A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0241—Advertisements
    • G06Q30/0251—Targeted advertisements
    • G06Q30/0269—Targeted advertisements based on user profile or attribute
    • G06Q30/0271—Personalized advertisement
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201—Market modelling; Market analysis; Collecting market data

Definitions

  • the present application relates to the field of computer technologies, and in particular, to a method, an electronic device, and a readable storage medium for pushing a loan advertisement in an application.
  • loan marketing method on the APP is usually a fixed location operation or a group text message, that is, loan marketing for all APP users.
  • the present application provides a method, an electronic device, and a readable storage medium for pushing a loan advertisement in an application, and aims to accurately push a loan advertisement to a user who has a loan demand among APP users.
  • a first aspect of the present application provides an electronic device including a memory, a processor, and a memory on which a system for pushing a loan advertisement in an application running on the processor is stored
  • the system for pushing a loan advertisement in an application is executed by the processor to implement the following steps:
  • the second aspect of the present application further provides a method for pushing a loan advertisement in an application, where the method for pushing a loan advertisement in an application includes:
  • Step 1 association user data on the application with a preset user image including the attribute tag to obtain associated user data; and labeling the loan-related attribute tag preset in the associated user data;
  • Step 2 extracting, from the associated user data, loan user data that has been converted into a loan user by the application, and training based on the loan-related attribute tag marked in the loan user data to obtain a loan propensity prediction Decision tree model;
  • Step 3 Using the decision tree model to predict the loan propensity of the user on the application, and pushing the loan advertisement information according to the predicted result.
  • a third aspect of the present application further provides a computer readable storage medium, where the computer readable storage medium stores a system for pushing a loan advertisement in an application, where the loan is pushed in the application
  • the system of advertisements can be executed by at least one processor to cause the at least one processor to perform the steps of the method of pushing a loan advertisement in an application as described above.
  • the method, system and readable storage medium for pushing a loan advertisement in an application according to the present application by associating user data on an application with a preset user image including an attribute tag to obtain associated user data; labeling the associated user
  • the loan-related attribute tag preset in the data so that the user data on the application also includes the attribute tag related to the loan, and the user who has been converted into the loan user through the application as the seed user, according to the seed user
  • a property tag with loan propensity to train and establish a decision tree model for loan propensity prediction, and use the decision tree model to predict the loan propensity of the user on the application, and may have a loan propensity for the user on the application.
  • the demanding users are accurately positioned to push the loan advertisement information accurately to the users with loan-oriented needs, which can improve the effectiveness of loan advertisement marketing and avoid harassment of non-loan demanders.
  • FIG. 1 is a schematic diagram of an optional application environment of each embodiment of the present application.
  • FIG. 2 is a schematic diagram of a hardware architecture of an embodiment of the electronic device of FIG. 1;
  • FIG. 3 is a schematic flowchart of an embodiment of a method for pushing a loan advertisement in an application in the application.
  • first, second and the like in the present application are for the purpose of description only, and are not to be construed as indicating or implying their relative importance or implicitly indicating the number of technical features indicated. .
  • features defining “first” and “second” may include at least one of the features, either explicitly or implicitly.
  • the technical solutions between the various embodiments may be combined with each other, but must be based on the realization of those skilled in the art, and when the combination of the technical solutions is contradictory or impossible to implement, it should be considered that the combination of the technical solutions does not exist. Nor is it within the scope of protection required by this application.
  • FIG. 1 is a schematic diagram of an optional application environment of each embodiment of the present application.
  • the present application is applicable to an application environment including, but not limited to, an electronic device 1, a terminal device 2, and a network 3.
  • the electronic device 1 is an apparatus capable of automatically performing numerical calculation and/or information processing in accordance with an instruction set or stored in advance.
  • the electronic device 1 may be a computer, a single network server, a server group composed of a plurality of network servers, or a cloud-based cloud composed of a large number of hosts or network servers, wherein the cloud computing is a type of distributed computing, A super virtual computer consisting of a loosely coupled set of computers.
  • the terminal device 2 includes, but is not limited to, any electronic product that can interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device, for example, a personal computer, a tablet computer, a smart phone, or an individual.
  • Digital Assistant (PDA) game console, Internet Protocol Television (IPTV), smart wearable device, etc.
  • the network 3 may be an intranet, an Internet, a global mobile communication system (global S-type communication of GSM in an application, GSM), and a Wideband Code Division (Wideband Code Division). Multiple Access, WCDMA), 4G network, 5G network, Bluetooth, Wi-Fi, etc.
  • the electronic device 1 is communicatively connected to one or more of the terminal devices 2 via the network 3 respectively.
  • FIG. 2 is a schematic diagram of an optional hardware architecture of the electronic device 1 of FIG. 1.
  • the electronic device 1 may include, but is not limited to, a memory 11 and a processor 12 that are communicably connected to each other through a system bus. Network interface 13. It is to be noted that FIG. 2 only shows the electronic device 1 having the components 11-13, but it should be understood that not all illustrated components may be implemented, and more or fewer components may be implemented instead.
  • the memory 11 includes at least one type of readable storage medium including a flash memory, a hard disk, a multimedia card, a card type memory (eg, SD or DX memory, etc.), and a random access memory (RAM). , static random access memory (SRAM), read only memory (ROM), electrically erasable programmable read only memory (EEPROM), programmable read only memory (PROM), magnetic memory, magnetic disk, optical disk, and the like.
  • the memory 11 may be an internal storage unit of the electronic device 1, such as a hard disk or a memory of the electronic device 1.
  • the memory 11 may also be an external storage device of the electronic device 1, such as a plug-in hard disk equipped on the electronic device 1, a smart memory card (SMC), and a secure digital device. (Secure Digital, SD) card, flash card, etc.
  • SMC smart memory card
  • SD Secure Digital
  • the memory 11 can also include both the internal storage unit of the electronic device 1 and its external storage device.
  • the memory 11 is generally used to store an operating system installed in the electronic device 1 and various types of application software, such as the program code of the system 10 for pushing a loan advertisement in an application. Further, the memory 11 can also be used to temporarily store various types of data that have been output or are to be output.
  • the processor 12 may be a Central Processing Unit (CPU), controller, microcontroller, microprocessor, or other data processing chip in some embodiments.
  • the processor 12 is typically used to control the overall operation of the electronic device 1, such as performing control and processing related to data interaction or communication with the terminal device 2.
  • the processor 12 is configured to run program code or process data stored in the memory 11, such as running the system 10 for pushing a loan advertisement in an application.
  • the network interface 13 may comprise a wireless network interface or a wired network interface, which is typically used to establish a communication connection between the electronic device 1 and other electronic devices.
  • the network interface 13 is mainly used to connect the electronic device 1 to one or more of the terminal devices 2 through the network 3, in the electronic device 1 and one or more of the terminals. A data transmission channel and a communication connection are established between the devices 2.
  • the system 10 for pushing a loan advertisement in an application includes at least one computer readable instruction stored in the memory 11, the at least one computer readable instruction being executable by the processor 12 to implement various embodiments of the present application.
  • step S1 the user data on the application is associated with the preset user image including the attribute tag to obtain the associated user data; and the loan-related attribute tag preset in the associated user data is marked.
  • the user data on the APP that needs to be pushed by the loan advertisement is associated with the preset user image to obtain the associated user data.
  • the user portrait is a target user model based on a series of real data, and is a tagged user model abstracted based on information such as user social attributes, living habits and consumer behavior.
  • the core task of building user portraits is to label users with "high-profile" features that are analyzed by user information.
  • the preset user portrait in this embodiment may directly call the established user portrait related to the loan personnel, or may be through various data sources (such as a loan website database, QQ, Weibo, WeChat, Snowball, Oriental).
  • the user portraits created include various attribute tags, such as the user's social attribute tags and financial attribute tags, such as age, family status, income, occupation, spending habits, whether there is a loan record Have you ever done a credit card, etc.
  • the attribute tag that is, various attributes in the user data on the application, such as age, gender, preference, income, etc., corresponding to the preset user portraits.
  • the attribute tags are matched, the user data with high matching degree is associated with the user image, and all attribute tags in the user portrait are assigned to the associated user data to form associated user data, so that the application only has the original
  • the user data of the basic attribute also has various attribute tags such as various social attribute tags and financial attribute tags in the user portrait.
  • the loan-related attribute tag preset in the generated associated user data may be marked.
  • the loan-related attribute tag may be set according to the unique attribute feature of the loan user to determine whether there are multiple credit cards, whether Have had loan records, etc.
  • the user data of the user on the APP cannot match the corresponding attribute tags in the preset user portrait, based on the behavior data of the user on the APP (eg, The online time period on the APP, the online time length, the most frequently browsed forum or column category, consumption items and consumption habits, etc., and the cosine similarity calculation to find similar users whose behavior data is similar on the APP.
  • the cosine of the angle between different vectors is closer to 1, it indicates that the closer the angle is to 0 degree, that is, the more similar the two vectors are, that is, the cosine similarity.
  • the behavior data of the user on the APP can be vectorized, and the degree of similarity of the vectors can be judged by the angle between the behavior data vectors of different users. The smaller the angle is, the more similar the representation is.
  • Finding a similar user on the APP that cannot be associated with the user's portrait, and the similar user can be associated with the user's portrait then associating the user's user data with the user's portrait associated with the user's similar user so that it can also Associations result in associated user data with multiple attribute tags.
  • Step S2 extracting, from the associated user data, loan user data that has been converted into a loan user by the application, and training based on the loan-related attribute tag marked in the loan user data to obtain a loan propensity prediction Decision tree model.
  • a user who selects an APP has been converted into a loan user by the APP as a seed user to establish a model, and a loan corresponding to a seed user that has been converted into a loan user by the APP is extracted from the associated user data of all users.
  • the user data is trained based on the loan-related attribute tag marked in the loan user data to obtain a decision tree model.
  • the decision tree is a prediction model, which represents a mapping relationship between object attributes and object values.
  • Each node in the tree represents an object, and each forked path represents a possible attribute value, and each leaf node corresponds to the object represented by the path from the root node to the leaf node. value.
  • each attribute label associated with the loan marked in the loan user data may be used as a positive sample, and each attribute label associated with the loan marked by the ordinary user of the non-loan user in the APP is randomly selected as a negative sample, and a positive sample is to be taken.
  • the negative samples are substituted into the default decision tree model for training, and the decision tree model that can be finally used for loan propensity prediction is obtained.
  • the decision tree model may be a Gradient Boosting Decision Tree (GBDT) model.
  • GBDT Gradient Boosting Decision Tree
  • Step S3 using the decision tree model to predict the loan propensity of the user on the application, and pushing the loan advertisement information according to the predicted result.
  • the associated user data of each user on the APP can be obtained, and the attribute label associated with the loan is obtained from the associated user data of the user, and the loan-related attributes of each user are obtained.
  • the tag is input to the decision tree model, and the decision tree model is used to analyze whether the predicted user is a user with a loan tendency, and if so, only the loan advertisement information is pushed for the user predicted to have a loan tendency to realize the loan advertisement information. Accurate push to avoid unnecessary interference to non-demanding people who do not have a loan propensity.
  • the APP interface is displayed on the terminal device of the user with the loan tendency, the preset loan advertisement is displayed at a preset location such as the bottom, the top, and the like at the key position of the APP interface.
  • the embodiment obtains the associated user data by associating the user data on the application with the preset user image including the attribute tag; and labeling the loan-related attribute tag preset in the associated user data. So that the user data on the application also includes the loan-related attribute tag, and the seed user who has been converted into the loan user through the application is trained and established according to the loan-oriented attribute tag marked by the seed user.
  • the decision tree model for predicting the loan propensity uses the decision tree model to predict the loan propensity of the user on the application, and can accurately locate the user who has the loan tendency in the application, and then has the loan
  • the demanding users can accurately push the loan advertisement information, which can improve the effectiveness of loan advertisement marketing, and can also avoid harassment of non-loan demand people.
  • the process of generating the decision tree model is as follows:
  • the user of the application that has been converted into a loan user through the application is marked as having a loan-prone user, and obtains a loan-related attribute tag marked in the loan user data corresponding to the loan-oriented user;
  • the preferred attribute labels corresponding to each node in the decision tree model are selected by the gain information, and the decision tree model is trained by the classification regression algorithm.
  • the Classification and Regression Tree (CART) algorithm uses a binary recursive segmentation technique to divide the current sample set into two subsample sets, so that each non-leaf node of the generated decision tree has two branches.
  • the gain information may include a Gini GINI index or the like.
  • the Gini GINI index is used to judge the degree of disorder of the decision model. The larger the coefficient, the more confusing it is.
  • the definition is the same as the definition of entropy.
  • the loan-related attribute tag marked in the loan user data corresponding to the loan user is trained to generate a decision tree model, and the leaf classification node of the decision tree model is related to the loan. Property tag.
  • the method for training the decision tree model includes, but is not limited to, a classification regression method, a naive Bayesian NBC algorithm, and the like.
  • the decision tree model When using the generated decision tree model to predict the loan propensity of the user on the application, acquiring the associated user data of the user on the application, and obtaining the attribute label related to the loan from the associated user data of the user; The decision tree model loads the loan-related attribute tag of the user; recursively traverses the decision tree model to find a decision tree leaf classification node corresponding to the loan-related attribute tag of the user, and the leaf node analyzes the user Whether it is a user with a loan preference.
  • FIG. 3 is a schematic flowchart of a method for pushing a loan advertisement in an application in the application, and the method for pushing a loan advertisement in an application includes the following steps:
  • step S10 the user data on the application is associated with the preset user image including the attribute tag to obtain the associated user data; and the loan-related attribute tag preset in the associated user data is marked.
  • the user data on the APP that needs to be pushed by the loan advertisement is associated with the preset user image to obtain the associated user data.
  • the user portrait is a target user model based on a series of real data, and is a tagged user model abstracted based on information such as user social attributes, living habits and consumer behavior.
  • the core task of building user portraits is to label users with "high-profile" features that are analyzed by user information.
  • the preset user portrait in this embodiment may directly call the established user portrait related to the loan personnel, or may be through various data sources (such as a loan website database, QQ, Weibo, WeChat, Snowball, Oriental).
  • the user portraits created include various attribute tags, such as the user's social attribute tags and financial attribute tags, such as age, family status, income, occupation, spending habits, whether there is a loan record Have you ever done a credit card, etc.
  • the attribute tag that is, various attributes in the user data on the application, such as age, gender, preference, income, etc., corresponding to the preset user portraits.
  • the attribute tags are matched, the user data with high matching degree is associated with the user image, and all attribute tags in the user portrait are assigned to the associated user data to form associated user data, so that the application only has the original
  • the user data of the basic attribute also has various attribute tags such as various social attribute tags and financial attribute tags in the user portrait.
  • the loan-related attribute tag preset in the generated associated user data may be marked.
  • the loan-related attribute tag may be set according to the unique attribute feature of the loan user to determine whether there are multiple credit cards, whether Have had loan records, etc.
  • the user data of the user on the APP cannot match the corresponding attribute tags in the preset user portrait, based on the behavior data of the user on the APP (eg, The online time period on the APP, the online time length, the most frequently browsed forum or column category, consumption items and consumption habits, etc., and the cosine similarity calculation to find similar users whose behavior data is similar on the APP.
  • the cosine of the angle between different vectors is closer to 1, it indicates that the closer the angle is to 0 degree, that is, the more similar the two vectors are, that is, the cosine similarity.
  • the behavior data of the user on the APP can be vectorized, and the degree of similarity of the vectors can be judged by the angle between the behavior data vectors of different users. The smaller the angle is, the more similar the representation is.
  • Finding a similar user on the APP that cannot be associated with the user's portrait, and the similar user can be associated with the user's portrait then associating the user's user data with the user's portrait associated with the user's similar user so that it can also Associations result in associated user data with multiple attribute tags.
  • Step S20 extracting, from the associated user data, loan user data that has been converted into a loan user by the application, and training based on the loan-related attribute tag marked in the loan user data to obtain a loan propensity prediction Decision tree model.
  • a user who selects an APP has been converted into a loan user by the APP as a seed user to establish a model, and a loan corresponding to a seed user that has been converted into a loan user by the APP is extracted from the associated user data of all users.
  • the user data is trained based on the loan-related attribute tag marked in the loan user data to obtain a decision tree model.
  • the decision tree is a prediction model, which represents a mapping relationship between object attributes and object values.
  • Each node in the tree represents an object, and each forked path represents a possible attribute value, and each leaf node corresponds to the object represented by the path from the root node to the leaf node. value.
  • each attribute label associated with the loan marked in the loan user data may be used as a positive sample, and each attribute label associated with the loan marked by the ordinary user of the non-loan user in the APP is randomly selected as a negative sample, and a positive sample is to be taken.
  • the negative samples are substituted into the default decision tree model for training, and the decision tree model that can be finally used for loan propensity prediction is obtained.
  • the decision tree model may be a Gradient Boosting Decision Tree (GBDT) model.
  • GBDT Gradient Boosting Decision Tree
  • Step S30 using the decision tree model to predict the loan propensity of the user on the application, and pushing the loan advertisement information according to the predicted result.
  • the associated user data of each user on the APP can be obtained, and the attribute label associated with the loan is obtained from the associated user data of the user, and the loan-related attributes of each user are obtained.
  • the tag is input to the decision tree model, and the decision tree model is used to analyze whether the predicted user is a user with a loan tendency, and if so, only the loan advertisement information is pushed for the user predicted to have a loan tendency to realize the loan advertisement information. Accurate push to avoid unnecessary interference to non-demanding people who do not have a loan propensity.
  • the APP interface is displayed on the terminal device of the user with the loan tendency, the preset loan advertisement is displayed at a preset location such as the bottom, the top, and the like at the key position of the APP interface.
  • the embodiment obtains the associated user data by associating the user data on the application with the preset user image including the attribute tag; and labeling the loan-related attribute tag preset in the associated user data. So that the user data on the application also includes the loan-related attribute tag, and the seed user who has been converted into the loan user through the application is trained and established according to the loan-oriented attribute tag marked by the seed user.
  • the decision tree model for predicting the loan propensity uses the decision tree model to predict the loan propensity of the user on the application, and can accurately locate the user who has the loan tendency in the application, and then has the loan
  • the demanding users can accurately push the loan advertisement information, which can improve the effectiveness of loan advertisement marketing, and can also avoid harassment of non-loan demand people.
  • the process of generating the decision tree model is as follows:
  • the user of the application that has been converted into a loan user through the application is marked as having a loan-prone user, and obtains a loan-related attribute tag marked in the loan user data corresponding to the loan-oriented user;
  • the preferred attribute labels corresponding to each node in the decision tree model are selected by the gain information, and the decision tree model is trained by the classification regression algorithm.
  • the Classification and Regression Tree (CART) algorithm uses a binary recursive segmentation technique to divide the current sample set into two subsample sets, so that each non-leaf node of the generated decision tree has two branches.
  • the gain information may include a Gini GINI index or the like.
  • the Gini GINI index is used to judge the degree of disorder of the decision model. The larger the coefficient, the more confusing it is.
  • the definition is the same as the definition of entropy.
  • the loan-related attribute tag marked in the loan user data corresponding to the loan user is trained to generate a decision tree model, and the leaf classification node of the decision tree model is related to the loan. Property tag.
  • the method for training the decision tree model includes, but is not limited to, a classification regression method, a naive Bayesian NBC algorithm, and the like.
  • the decision tree model When using the generated decision tree model to predict the loan propensity of the user on the application, acquiring the associated user data of the user on the application, and obtaining the attribute label related to the loan from the associated user data of the user; The decision tree model loads the loan-related attribute tag of the user; recursively traverses the decision tree model to find a decision tree leaf classification node corresponding to the loan-related attribute tag of the user, and the leaf node analyzes the user Whether it is a user with a loan preference.
  • the present application also provides a computer readable storage medium storing a system for pushing a loan advertisement in an application
  • the system for pushing a loan advertisement in the application can be at least one processor Executing, in order for the at least one processor to perform the step of the method of pushing the loan advertisement in the application as in the above embodiment, the step S10, S20, S30, etc. of the method for pushing the loan advertisement in the application As described above, it will not be described here.
  • the foregoing embodiment method can be implemented by means of software plus a necessary general hardware platform, and can also be implemented by hardware, but in many cases, the former is A better implementation.
  • the technical solution of the present application which is essential or contributes to the prior art, may be embodied in the form of a software product stored in a storage medium (such as ROM/RAM, disk,
  • the optical disc includes a number of instructions for causing a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to perform the methods described in various embodiments of the present application.

Landscapes

  • Business, Economics & Management (AREA)
  • Strategic Management (AREA)
  • Engineering & Computer Science (AREA)
  • Accounting & Taxation (AREA)
  • Development Economics (AREA)
  • Finance (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Marketing (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)

Abstract

本申请涉及一种在应用程序中推送贷款广告的方法、电子装置及可读存储介质，该方法包括：将应用程序上的用户数据与包含属性标签的预设用户画像进行关联得到关联用户数据；标注所述关联用户数据中预设的与贷款相关的属性标签；从所述关联用户数据中提取出已通过该应用程序转化为贷款用户的贷款用户数据，并基于所述贷款用户数据中标注的与贷款相关的属性标签进行训练得到用于贷款倾向预测的决策树模型；利用所述决策树模型对该应用程序上的用户进行贷款倾向预测，并根据预测结果进行贷款广告信息的推送。本申请能对具有贷款倾向的需求用户进行贷款广告信息的精准推送，且还能避免对非贷款需求人群造成骚扰。

Description

在应用程序中推送贷款广告的方法、电子装置及介质
优先权申明
本申请基于巴黎公约申明享有2017年9月30日递交的申请号为CN 201710916516.4、名称为“在应用程序中推送贷款广告的方法、电子装置及介质”中国专利申请的优先权，该中国专利申请的整体内容以参考的方式结合在本申请中。
技术领域
本申请涉及计算机技术领域，尤其涉及一种在应用程序中推送贷款广告的方法、电子装置及可读存储介质。
背景技术
当前在APP上的贷款营销方式通常是固定位置运营或群发短信，即对所有的APP用户进行贷款营销。缺乏对需求人群的精准定位，浪费了宝贵的广告资源，效果往往很差，同时，也对非目标人群造成了不必要的骚扰，影响用户体验。
发明内容
本申请提供一种在应用程序中推送贷款广告的方法、电子装置及可读存储介质，旨在向APP用户中有贷款需求人群进行贷款广告的精准推送。
为实现上述目的，本申请第一方面提供一种电子装置，所述电子装置包括存储器、处理器，所述存储器上存储有可在所述处理器上运行的在应用程序中推送贷款广告的系统，所述在应用程序中推送贷款广告的系统被所述处理器执行时实现如下步骤：
A、将应用程序上的用户数据与包含属性标签的预设用户画像进行关联得到关联用户数据；标注所述关联用户数据中预设的与贷款相关的属性标签；
B、从所述关联用户数据中提取出已通过该应用程序转化为贷款用户的贷款用户数据，并基于所述贷款用户数据中标注的与贷款相关的属性标签进行训练得到用于贷款倾向预测的决策树模型；
C、利用所述决策树模型对该应用程序上的用户进行贷款倾向预测，并根据预测结果进行贷款广告信息的推送。
此外，为实现上述目的，本申请第二方面还提供一种在应用程序中推送贷款广告的方法，所述在应用程序中推送贷款广告的方法包括：
步骤一、将应用程序上的用户数据与包含属性标签的预设用户画 像进行关联得到关联用户数据；标注所述关联用户数据中预设的与贷款相关的属性标签；
步骤二、从所述关联用户数据中提取出已通过该应用程序转化为贷款用户的贷款用户数据，并基于所述贷款用户数据中标注的与贷款相关的属性标签进行训练得到用于贷款倾向预测的决策树模型；
步骤三、利用所述决策树模型对该应用程序上的用户进行贷款倾向预测，并根据预测结果进行贷款广告信息的推送。
进一步地，为实现上述目的，本申请第三方面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存储有在应用程序中推送贷款广告的系统，所述在应用程序中推送贷款广告的系统可被至少一个处理器执行，以使所述至少一个处理器执行如上述的在应用程序中推送贷款广告的方法的步骤。
本申请提出的在应用程序中推送贷款广告的方法、系统及可读存储介质，通过将应用程序上的用户数据与包含属性标签的预设用户画像进行关联得到关联用户数据；标注所述关联用户数据中预设的与贷款相关的属性标签，使得应用程序上的用户数据中也包含与贷款相关的属性标签，并将已通过该应用程序转化为贷款用户的作为种子用户，根据种子用户标注的具有贷款倾向的属性标签来训练、建立用于贷款倾向预测的决策树模型，利用所述决策树模型对该应用程序上的用户进行贷款倾向预测，可对该应用程序上的用户中具有贷款倾向的需求用户进行精准定位，进而对具有贷款倾向的需求用户进行贷款广告信息的精准推送，能提升贷款广告营销的有效性，且还能避免对非贷款需求人群造成骚扰。
附图说明
图1为本申请各个实施例一可选的应用环境示意图；
图2是图1中电子装置一实施例的硬件架构的示意图；
图3为本申请在应用程序中推送贷款广告的方法一实施例的流程示意图。
具体实施方式
为了使本申请的目的、技术方案及优点更加清楚明白，以下结合附图及实施例，对本申请进行进一步详细说明。应当理解，此处所描述的具体实施例仅用以解释本申请，并不用于限定本申请。基于本申请中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。
需要说明的是，在本申请中涉及“第一”、“第二”等的描述仅 用于描述目的，而不能理解为指示或暗示其相对重要性或者隐含指明所指示的技术特征的数量。由此，限定有“第一”、“第二”的特征可以明示或者隐含地包括至少一个该特征。另外，各个实施例之间的技术方案可以相互结合，但是必须是以本领域普通技术人员能够实现为基础，当技术方案的结合出现相互矛盾或无法实现时应当认为这种技术方案的结合不存在，也不在本申请要求的保护范围之内。
参阅图1，是本申请各个实施例一可选的应用环境示意图。
在本实施例中，本申请可应用于包括，但不仅限于，电子装置1、终端设备2、网络3的应用环境中。其中，电子装置1是一种能够按照事先设定或者存储的指令，自动进行数值计算和/或信息处理的设备。电子装置1可以是计算机、也可以是单个网络服务器、多个网络服务器组成的服务器组或者基于云计算的由大量主机或者网络服务器构成的云，其中云计算是分布式计算的一种，由一群松散耦合的计算机集组成的一个超级虚拟计算机。
终端设备2包括，但不限于，任何一种可与用户通过键盘、鼠标、遥控器、触摸板或者声控设备等方式进行人机交互的电子产品，例如，个人计算机、平板电脑、智能手机、个人数字助理(Personal Digital Assistant，PDA)，游戏机、交互式网络电视(Internet Protocol Television,IPTV)、智能式穿戴式设备等。
所述网络3可以是企业内部网(Intranet)、互联网(Internet)、全球移动通讯系统(Global S在应用程序中推送贷款广告的stem of Mobile communication，GSM)、宽带码分多址(Wideband Code Division Multiple Access，WCDMA)、4G网络、5G网络、蓝牙(Bluetooth)、Wi-Fi等无线或有线网络。其中，所述电子装置1通过所述网络3分别与一个或多个所述终端设备2通信连接。
参阅图2，是图1中电子装置1一可选的硬件架构的示意图，本实施例中，电子装置1可包括，但不仅限于，可通过系统总线相互通信连接的存储器11、处理器12、网络接口13。需要指出的是，图2仅示出了具有组件11-13的电子装置1，但是应理解的是，并不要求实施所有示出的组件，可以替代的实施更多或者更少的组件。
其中，所述存储器11至少包括一种类型的可读存储介质，所述可读存储介质包括闪存、硬盘、多媒体卡、卡型存储器(例如，SD或DX存储器等)、随机访问存储器(RAM)、静态随机访问存储器(SRAM)、只读存储器(ROM)、电可擦除可编程只读存储器(EEPROM)、可编程只读存储器(PROM)、磁性存储器、磁盘、光盘等。在一些实施例中，所述存储器11可以是所述电子装置1的内部存储单元，例如该电子装置1的硬盘或内存。在另一些实施例中，所述存储器11也可以是所述电子装置1的外部存储设备，例如该电 子装置1上配备的插接式硬盘，智能存储卡(Smart Media Card,SMC)，安全数字(Secure Digital,SD)卡，闪存卡(Flash Card)等。当然，所述存储器11还可以既包括所述电子装置1的内部存储单元也包括其外部存储设备。本实施例中，所述存储器11通常用于存储安装于所述电子装置1的操作系统和各类应用软件，例如所述在应用程序中推送贷款广告的系统10的程序代码等。此外，所述存储器11还可以用于暂时地存储已经输出或者将要输出的各类数据。
所述处理器12在一些实施例中可以是中央处理器(Central Processing Unit，CPU)、控制器、微控制器、微处理器、或其他数据处理芯片。该处理器12通常用于控制所述电子装置1的总体操作，例如执行与所述终端设备2进行数据交互或者通信相关的控制和处理等。本实施例中，所述处理器12用于运行所述存储器11中存储的程序代码或者处理数据，例如运行所述在应用程序中推送贷款广告的系统10等。
所述网络接口13可包括无线网络接口或有线网络接口，该网络接口13通常用于在所述电子装置1与其他电子设备之间建立通信连接。本实施例中，所述网络接口13主要用于通过所述网络3将所述电子装置1与一个或多个所述终端设备2相连，在所述电子装置1与一个或多个所述终端设备2之间建立数据传输通道和通信连接。
在应用程序中推送贷款广告的系统10包括至少一个存储在所述存储器11中的计算机可读指令，该至少一个计算机可读指令可被所述处理器12执行，以实现本申请各实施例。
其中，上述在应用程序中推送贷款广告的系统10被所述处理器12执行时实现如下步骤：
步骤S1，将应用程序上的用户数据与包含属性标签的预设用户画像进行关联得到关联用户数据；标注所述关联用户数据中预设的与贷款相关的属性标签。
本实施例中，将需进行贷款广告推送的应用程序即APP上的用户数据关联上预设用户画像，以得到关联用户数据。其中，用户画像是建立在一系列真实数据之上的目标用户模型，是根据用户社会属性、生活习惯和消费行为等信息而抽象出的一个标签化的用户模型。构建用户画像的核心工作即是给用户贴“标签”，而标签是通过对用户信息分析而来的高度精炼的特征标识。本实施例中的预设用户画像可以是直接调用已建立好的与贷款人员相关的用户画像，也可以是通过各种数据源(如贷款网站数据库、QQ、微博、微信、雪球、东方财富社交软件等)来建立用户画像，建立的用户画像中包括各种属性标签，如用户的社会属性标签和金融属性标签，如年龄、家庭状况、收入、职业、消费习惯、是否有过贷款记录、是否办过信用卡等等。
将应用程序上的用户数据与包含属性标签的预设用户画像进行关联，即将应用程序上的用户数据中的各项属性如年龄、性别、喜好、收入等与预设用户画像中对应的各种属性标签进行匹配，将匹配度较高的用户数据与用户画像进行关联，将用户画像中的所有属性标签赋予与之相关联的用户数据，形成关联用户数据，这样，使得应用程序上本来仅有基本属性的用户数据也具有了用户画像中的各种属性标签如各种社会属性标签和金融属性标签等。进一步地，还可标注生成的关联用户数据中预设的与贷款相关的属性标签，例如，可预先根据贷款用户的特有属性特征，设定与贷款相关的属性标签为是否具有多张信用卡、是否有过贷款记录等。
此外，若APP上有用户无法关联到用户画像，如该用户在APP上的用户数据不能与预设用户画像中对应的各种属性标签相匹配，则基于该用户在APP上的行为数据(如在APP上的在线时间段、在线时长、最常浏览的论坛或栏目类别、消费项目及消费习惯等等)，并通过余弦相似度计算找出在APP上该用户的行为数据相似的相似用户。具体地，由于不同向量之间夹角的余弦值越接近1，就表明夹角越接近0度，也就是两个向量越相似，即余弦相似性。因此，可将用户在APP上的行为数据向量化，通过不同用户的行为数据向量之间夹角的大小，来判断向量的相似程度，夹角越小，就代表越相似。找到无法关联到用户画像的用户在APP上的相似用户，而且该相似用户能关联到用户画像，则将该用户的用户数据关联到该用户的相似用户所关联的用户画像，以使其也能关联得到具有多种属性标签的关联用户数据。
步骤S2，从所述关联用户数据中提取出已通过该应用程序转化为贷款用户的贷款用户数据，并基于所述贷款用户数据中标注的与贷款相关的属性标签进行训练得到用于贷款倾向预测的决策树模型。
本实施例中，选取APP的用户中已经通过该APP转化为贷款用户的作为种子用户来建立模型，从所有用户的关联用户数据中提取出已通过该APP转化为贷款用户的种子用户对应的贷款用户数据，基于所述贷款用户数据中标注的与贷款相关的属性标签进行训练得到决策树模型。
其中，决策树是一个预测模型，它代表的是对象属性与对象值之间的一种映射关系。树中每个节点表示某个对象，而每个分叉路径则代表的某个可能的属性值，而每个叶结点则对应从根节点到该叶节点所经历的路径所表示的对象的值。本实施例中，可将贷款用户数据中标注的与贷款相关的各个属性标签作为正样本，随机抽取APP中非贷款用户的普通用户标注的与贷款相关的各个属性标签作为负样本，将正样本、负样本代入预设的决策树模型进行训练，得到最终可用于 贷款倾向预测的决策树模型。可选的，该决策树模型可以是梯度提升决策树(Gradient Boosting Decision Tree，GBDT)模型。
步骤S3，利用所述决策树模型对该应用程序上的用户进行贷款倾向预测，并根据预测结果进行贷款广告信息的推送。
本实施例中，针对该APP上的用户，可获取该APP上各个用户的关联用户数据，并从该用户的关联用户数据中得到与贷款相关的属性标签，将各个用户的与贷款相关的属性标签输入至所述决策树模型，利用所述决策树模型来分析预测用户是否为具有贷款倾向的用户，若是，则仅针对预测为具有贷款倾向的用户推送贷款广告信息，以实现贷款广告信息的精准推送，避免对不具有贷款倾向的非需求人群造成不必要的干扰。
在对具有贷款倾向的用户进行贷款广告信息推送时，可通过短信、站内信、消息推送push等方式使用不同文案定期推送贷款广告，以保持贷款营销热度。同时还可根据设备号在具有贷款倾向的用户的终端设备显示该APP界面时在预设位置如底部、顶端等APP界面关键位置展示预设的贷款广告。
与现有技术相比，本实施例通过将应用程序上的用户数据与包含属性标签的预设用户画像进行关联得到关联用户数据；标注所述关联用户数据中预设的与贷款相关的属性标签，使得应用程序上的用户数据中也包含与贷款相关的属性标签，并将已通过该应用程序转化为贷款用户的作为种子用户，根据种子用户标注的具有贷款倾向的属性标签来训练、建立用于贷款倾向预测的决策树模型，利用所述决策树模型对该应用程序上的用户进行贷款倾向预测，可对该应用程序上的用户中具有贷款倾向的需求用户进行精准定位，进而对具有贷款倾向的需求用户进行贷款广告信息的精准推送，能提升贷款广告营销的有效性，且还能避免对非贷款需求人群造成骚扰。
在一可选的实施例中，在上述图2的实施例的基础上，所述决策树模型的生成过程如下：
将应用程序的用户中已通过该应用程序转化为贷款用户的标注为具有贷款倾向用户，并获取具有贷款倾向用户对应的贷款用户数据中标注的与贷款相关的属性标签；
根据标注的具有贷款倾向用户及其对应的与贷款相关的属性标签，采用递归的方法将与贷款相关的属性标签划分为多个较小的子集；
通过增益信息选择决策树模型中各个节点所对应的较佳属性标签，采用分类回归算法训练得到决策树模型，当然还可以采用其它数 据挖掘算法。所述分类回归CART(Classification And Regression Tree)算法采用一种二分递归分割的技术，将当前的样本集分为两个子样本集，使得生成的决策树的每个非叶子节点都有两个分支。所述增益信息，可以包括基尼GINI指数等。其中基尼GINI指数用来判断决策模型的类别的杂乱程度，系数越大表示越混乱，其定义与熵的定义相同。通过对所述指数的增益信息进行比较，选择较优的属性标签构成决策树模型。
本实施例中通过提取出的贷款用户，对贷款用户对应的贷款用户数据中标注的与贷款相关的属性标签进行训练，生成决策树模型，所述决策树模型的叶子分类节点为与贷款相关的属性标签。通过对已标注为具有贷款倾向用户对应的与贷款相关的属性标签进行训练，将训练的结果不断完善，可得到较优的决策树模型。其中，所述训练生成决策树模型的方法，包括但不限于分类回归法、朴素贝叶斯NBC算法等。
在利用生成的决策树模型对该应用程序上的用户进行贷款倾向预测时，获取该应用程序上用户的关联用户数据，从该用户的关联用户数据中得到与贷款相关的属性标签；根据训练得到的决策树模型加载该用户的与贷款相关的属性标签；递归遍历所述决策树模型，查找该用户的与贷款相关的属性标签对应的决策树叶子分类节点，由所述叶子节点分析出该用户是否为具有贷款倾向的用户。
如图3所示，图3为本申请在应用程序中推送贷款广告的方法一实施例的流程示意图，该在应用程序中推送贷款广告的方法包括以下步骤：
步骤S10，将应用程序上的用户数据与包含属性标签的预设用户画像进行关联得到关联用户数据；标注所述关联用户数据中预设的与贷款相关的属性标签。
本实施例中，将需进行贷款广告推送的应用程序即APP上的用户数据关联上预设用户画像，以得到关联用户数据。其中，用户画像是建立在一系列真实数据之上的目标用户模型，是根据用户社会属性、生活习惯和消费行为等信息而抽象出的一个标签化的用户模型。构建用户画像的核心工作即是给用户贴“标签”，而标签是通过对用户信息分析而来的高度精炼的特征标识。本实施例中的预设用户画像可以是直接调用已建立好的与贷款人员相关的用户画像，也可以是通过各种数据源(如贷款网站数据库、QQ、微博、微信、雪球、东方财富社交软件等)来建立用户画像，建立的用户画像中包括各种属性标签，如用户的社会属性标签和金融属性标签，如年龄、家庭状况、收入、职业、消费习惯、是否有过贷款记录、是否办过信用卡等等。
将应用程序上的用户数据与包含属性标签的预设用户画像进行关联，即将应用程序上的用户数据中的各项属性如年龄、性别、喜好、收入等与预设用户画像中对应的各种属性标签进行匹配，将匹配度较高的用户数据与用户画像进行关联，将用户画像中的所有属性标签赋予与之相关联的用户数据，形成关联用户数据，这样，使得应用程序上本来仅有基本属性的用户数据也具有了用户画像中的各种属性标签如各种社会属性标签和金融属性标签等。进一步地，还可标注生成的关联用户数据中预设的与贷款相关的属性标签，例如，可预先根据贷款用户的特有属性特征，设定与贷款相关的属性标签为是否具有多张信用卡、是否有过贷款记录等。
此外，若APP上有用户无法关联到用户画像，如该用户在APP上的用户数据不能与预设用户画像中对应的各种属性标签相匹配，则基于该用户在APP上的行为数据(如在APP上的在线时间段、在线时长、最常浏览的论坛或栏目类别、消费项目及消费习惯等等)，并通过余弦相似度计算找出在APP上该用户的行为数据相似的相似用户。具体地，由于不同向量之间夹角的余弦值越接近1，就表明夹角越接近0度，也就是两个向量越相似，即余弦相似性。因此，可将用户在APP上的行为数据向量化，通过不同用户的行为数据向量之间夹角的大小，来判断向量的相似程度，夹角越小，就代表越相似。找到无法关联到用户画像的用户在APP上的相似用户，而且该相似用户能关联到用户画像，则将该用户的用户数据关联到该用户的相似用户所关联的用户画像，以使其也能关联得到具有多种属性标签的关联用户数据。
步骤S20，从所述关联用户数据中提取出已通过该应用程序转化为贷款用户的贷款用户数据，并基于所述贷款用户数据中标注的与贷款相关的属性标签进行训练得到用于贷款倾向预测的决策树模型。
本实施例中，选取APP的用户中已经通过该APP转化为贷款用户的作为种子用户来建立模型，从所有用户的关联用户数据中提取出已通过该APP转化为贷款用户的种子用户对应的贷款用户数据，基于所述贷款用户数据中标注的与贷款相关的属性标签进行训练得到决策树模型。
其中，决策树是一个预测模型，它代表的是对象属性与对象值之间的一种映射关系。树中每个节点表示某个对象，而每个分叉路径则代表的某个可能的属性值，而每个叶结点则对应从根节点到该叶节点所经历的路径所表示的对象的值。本实施例中，可将贷款用户数据中标注的与贷款相关的各个属性标签作为正样本，随机抽取APP中非贷款用户的普通用户标注的与贷款相关的各个属性标签作为负样本，将正样本、负样本代入预设的决策树模型进行训练，得到最终可用于 贷款倾向预测的决策树模型。可选的，该决策树模型可以是梯度提升决策树(Gradient Boosting Decision Tree，GBDT)模型。
步骤S30，利用所述决策树模型对该应用程序上的用户进行贷款倾向预测，并根据预测结果进行贷款广告信息的推送。
本实施例中，针对该APP上的用户，可获取该APP上各个用户的关联用户数据，并从该用户的关联用户数据中得到与贷款相关的属性标签，将各个用户的与贷款相关的属性标签输入至所述决策树模型，利用所述决策树模型来分析预测用户是否为具有贷款倾向的用户，若是，则仅针对预测为具有贷款倾向的用户推送贷款广告信息，以实现贷款广告信息的精准推送，避免对不具有贷款倾向的非需求人群造成不必要的干扰。
在对具有贷款倾向的用户进行贷款广告信息推送时，可通过短信、站内信、消息推送push等方式使用不同文案定期推送贷款广告，以保持贷款营销热度。同时还可根据设备号在具有贷款倾向的用户的终端设备显示该APP界面时在预设位置如底部、顶端等APP界面关键位置展示预设的贷款广告。
与现有技术相比，本实施例通过将应用程序上的用户数据与包含属性标签的预设用户画像进行关联得到关联用户数据；标注所述关联用户数据中预设的与贷款相关的属性标签，使得应用程序上的用户数据中也包含与贷款相关的属性标签，并将已通过该应用程序转化为贷款用户的作为种子用户，根据种子用户标注的具有贷款倾向的属性标签来训练、建立用于贷款倾向预测的决策树模型，利用所述决策树模型对该应用程序上的用户进行贷款倾向预测，可对该应用程序上的用户中具有贷款倾向的需求用户进行精准定位，进而对具有贷款倾向的需求用户进行贷款广告信息的精准推送，能提升贷款广告营销的有效性，且还能避免对非贷款需求人群造成骚扰。
在一可选的实施例中，在上述实施例的基础上，所述决策树模型的生成过程如下：
将应用程序的用户中已通过该应用程序转化为贷款用户的标注为具有贷款倾向用户，并获取具有贷款倾向用户对应的贷款用户数据中标注的与贷款相关的属性标签；
根据标注的具有贷款倾向用户及其对应的与贷款相关的属性标签，采用递归的方法将与贷款相关的属性标签划分为多个较小的子集；
通过增益信息选择决策树模型中各个节点所对应的较佳属性标签，采用分类回归算法训练得到决策树模型，当然还可以采用其它数 据挖掘算法。所述分类回归CART(Classification And Regression Tree)算法采用一种二分递归分割的技术，将当前的样本集分为两个子样本集，使得生成的决策树的每个非叶子节点都有两个分支。所述增益信息，可以包括基尼GINI指数等。其中基尼GINI指数用来判断决策模型的类别的杂乱程度，系数越大表示越混乱，其定义与熵的定义相同。通过对所述指数的增益信息进行比较，选择较优的属性标签构成决策树模型。
本实施例中通过提取出的贷款用户，对贷款用户对应的贷款用户数据中标注的与贷款相关的属性标签进行训练，生成决策树模型，所述决策树模型的叶子分类节点为与贷款相关的属性标签。通过对已标注为具有贷款倾向用户对应的与贷款相关的属性标签进行训练，将训练的结果不断完善，可得到较优的决策树模型。其中，所述训练生成决策树模型的方法，包括但不限于分类回归法、朴素贝叶斯NBC算法等。
在利用生成的决策树模型对该应用程序上的用户进行贷款倾向预测时，获取该应用程序上用户的关联用户数据，从该用户的关联用户数据中得到与贷款相关的属性标签；根据训练得到的决策树模型加载该用户的与贷款相关的属性标签；递归遍历所述决策树模型，查找该用户的与贷款相关的属性标签对应的决策树叶子分类节点，由所述叶子节点分析出该用户是否为具有贷款倾向的用户。
此外，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存储有在应用程序中推送贷款广告的系统，所述在应用程序中推送贷款广告的系统可被至少一个处理器执行，以使所述至少一个处理器执行如上述实施例中的在应用程序中推送贷款广告的方法的步骤，该在应用程序中推送贷款广告的方法的步骤S10、S20、S30等具体实施过程如上文所述，在此不再赘述。
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。
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来实现，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。
以上参照附图说明了本申请的优选实施例，并非因此局限本申请的权利范围。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。另外，虽然在流程图中示出了逻辑顺序，但是在某些情况下，可以以不同于此处的顺序执行所示出或描述的步骤。
本领域技术人员不脱离本申请的范围和实质，可以有多种变型方案实现本申请，比如作为一个实施例的特征可用于另一实施例而得到又一实施例。凡在运用本申请的技术构思之内所作的任何修改、等同替换和改进，均应在本申请的权利范围之内。

Claims (20)

  1. 一种电子装置，其特征在于，所述电子装置包括存储器、处理器，所述存储器上存储有可在所述处理器上运行的在应用程序中推送贷款广告的系统，所述在应用程序中推送贷款广告的系统被所述处理器执行时实现如下步骤：
    A、将应用程序上的用户数据与包含属性标签的预设用户画像进行关联得到关联用户数据；标注所述关联用户数据中预设的与贷款相关的属性标签；
    B、从所述关联用户数据中提取出已通过该应用程序转化为贷款用户的贷款用户数据，并基于所述贷款用户数据中标注的与贷款相关的属性标签进行训练得到用于贷款倾向预测的决策树模型；
    C、利用所述决策树模型对该应用程序上的用户进行贷款倾向预测，并根据预测结果进行贷款广告信息的推送。
  2. 如权利要求1所述的电子装置，其特征在于，所述决策树模型的生成过程如下：
    将应用程序的用户中已通过该应用程序转化为贷款用户的标注为具有贷款倾向用户，并获取具有贷款倾向用户对应的贷款用户数据中标注的与贷款相关的属性标签；
    根据标注的具有贷款倾向用户及其对应的与贷款相关的属性标签，采用递归的方法将与贷款相关的属性标签划分为多个子集；
    通过增益信息选择决策树模型中各个节点所对应的最佳属性标签，采用分类回归算法训练得到决策树模型。
  3. 如权利要求1所述的电子装置，其特征在于，所述步骤C包括：
    获取该应用程序上用户的关联用户数据，从该用户的关联用户数据中得到与贷款相关的属性标签；
    根据训练得到的决策树模型加载该用户的与贷款相关的属性标签；
    递归遍历所述决策树模型，查找该用户的与贷款相关的属性标签对应的决策树叶子分类节点，由所述叶子节点分析出该用户是否为具有贷款倾向的用户；
    针对由所述决策树模型分析出的具有贷款倾向的用户，通过短信、站内信和/或消息推送push的方式推送预设的贷款广告信息，或者，在该用户应用程序界面的预设位置展示预设的贷款广告信息。
  4. 如权利要求2所述的电子装置，其特征在于，所述步骤C包括：
    获取该应用程序上用户的关联用户数据，从该用户的关联用户数据中得到与贷款相关的属性标签；
    根据训练得到的决策树模型加载该用户的与贷款相关的属性标签；
    递归遍历所述决策树模型，查找该用户的与贷款相关的属性标签对应的决策树叶子分类节点，由所述叶子节点分析出该用户是否为具有贷款倾向的用户；
    针对由所述决策树模型分析出的具有贷款倾向的用户，通过短信、站内信和/或消息推送push的方式推送预设的贷款广告信息，或 者，在该用户应用程序界面的预设位置展示预设的贷款广告信息。
  5. 如权利要求1所述的电子装置，其特征在于，所述步骤A包括：
    若有应用程序上的用户数据无法关联到用户画像，则基于该用户在应用程序上的行为数据，并通过余弦相似度计算找出应用程序上与该用户的相似用户，并将该用户的用户数据关联到该用户的相似用户所关联的用户画像；其中，该相似用户为应用程序上能关联到用户画像的用户。
  6. 如权利要求2所述的电子装置，其特征在于，所述步骤A包括：
    若有应用程序上的用户数据无法关联到用户画像，则基于该用户在应用程序上的行为数据，并通过余弦相似度计算找出应用程序上与该用户的相似用户，并将该用户的用户数据关联到该用户的相似用户所关联的用户画像；其中，该相似用户为应用程序上能关联到用户画像的用户。
  7. 如权利要求1所述的电子装置，其特征在于，所述在应用程序中推送贷款广告的系统被所述处理器执行时，还包括：
    通过预设数据源建立预设用户画像，所述预设用户画像包括用户的社会属性标签和金融属性标签。
  8. 如权利要求2所述的电子装置，其特征在于，所述在应用程 序中推送贷款广告的系统被所述处理器执行时，还包括：
    通过预设数据源建立预设用户画像，所述预设用户画像包括用户的社会属性标签和金融属性标签。
  9. 一种在应用程序中推送贷款广告的方法，其特征在于，所述在应用程序中推送贷款广告的方法包括：
    步骤一、将应用程序上的用户数据与包含属性标签的预设用户画像进行关联得到关联用户数据；标注所述关联用户数据中预设的与贷款相关的属性标签；
    步骤二、从所述关联用户数据中提取出已通过该应用程序转化为贷款用户的贷款用户数据，并基于所述贷款用户数据中标注的与贷款相关的属性标签进行训练得到用于贷款倾向预测的决策树模型；
    步骤三、利用所述决策树模型对该应用程序上的用户进行贷款倾向预测，并根据预测结果进行贷款广告信息的推送。
  10. 如权利要求9所述的在应用程序中推送贷款广告的方法，其特征在于，所述决策树模型的生成过程如下：
    将应用程序的用户中已通过该应用程序转化为贷款用户的标注为具有贷款倾向用户，并获取具有贷款倾向用户对应的贷款用户数据中标注的与贷款相关的属性标签；
    根据标注的具有贷款倾向用户及其对应的与贷款相关的属性标签，采用递归的方法将与贷款相关的属性标签划分为多个子集；
    通过增益信息选择决策树模型中各个节点所对应的最佳属性标签，采用分类回归算法训练得到决策树模型。
  11. 如权利要求9所述的在应用程序中推送贷款广告的方法，其特征在于，所述步骤三包括：
    获取该应用程序上用户的关联用户数据，从该用户的关联用户数据中得到与贷款相关的属性标签；
    根据训练得到的决策树模型加载该用户的与贷款相关的属性标签；
    递归遍历所述决策树模型，查找该用户的与贷款相关的属性标签对应的决策树叶子分类节点，由所述叶子节点分析出该用户是否为具有贷款倾向的用户；
    针对由所述决策树模型分析出的具有贷款倾向的用户，通过短信、站内信和/或消息推送push的方式推送预设的贷款广告信息，或者，在该用户应用程序界面的预设位置展示预设的贷款广告信息。
  12. 如权利要求10所述的在应用程序中推送贷款广告的方法，其特征在于，所述步骤三包括：
    获取该应用程序上用户的关联用户数据，从该用户的关联用户数据中得到与贷款相关的属性标签；
    根据训练得到的决策树模型加载该用户的与贷款相关的属性标签；
    递归遍历所述决策树模型，查找该用户的与贷款相关的属性标签对应的决策树叶子分类节点，由所述叶子节点分析出该用户是否为具有贷款倾向的用户；
    针对由所述决策树模型分析出的具有贷款倾向的用户，通过短 信、站内信和/或消息推送push的方式推送预设的贷款广告信息，或者，在该用户应用程序界面的预设位置展示预设的贷款广告信息。
  13. 如权利要求9所述的在应用程序中推送贷款广告的方法，其特征在于，还包括：
    若有应用程序上的用户数据无法关联到用户画像，则基于该用户在应用程序上的行为数据，并通过余弦相似度计算找出应用程序上与该用户的相似用户，并将该用户的用户数据关联到该用户的相似用户所关联的用户画像；其中，该相似用户为应用程序上能关联到用户画像的用户。
  14. 如权利要求10所述的在应用程序中推送贷款广告的方法，其特征在于，还包括：
    若有应用程序上的用户数据无法关联到用户画像，则基于该用户在应用程序上的行为数据，并通过余弦相似度计算找出应用程序上与该用户的相似用户，并将该用户的用户数据关联到该用户的相似用户所关联的用户画像；其中，该相似用户为应用程序上能关联到用户画像的用户。
  15. 如权利要求9所述的在应用程序中推送贷款广告的方法，其特征在于，还包括：
    通过预设数据源建立预设用户画像，所述预设用户画像包括用户的社会属性标签和金融属性标签。
  16. 如权利要求10所述的在应用程序中推送贷款广告的方法，其特征在于，还包括：
    通过预设数据源建立预设用户画像，所述预设用户画像包括用户的社会属性标签和金融属性标签。
  17. 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在应用程序中推送贷款广告的系统，所述在应用程序中推送贷款广告的系统被处理器执行时实现如下步骤：
    步骤一、将应用程序上的用户数据与包含属性标签的预设用户画像进行关联得到关联用户数据；标注所述关联用户数据中预设的与贷款相关的属性标签；
    步骤二、从所述关联用户数据中提取出已通过该应用程序转化为贷款用户的贷款用户数据，并基于所述贷款用户数据中标注的与贷款相关的属性标签进行训练得到用于贷款倾向预测的决策树模型；
    步骤三、利用所述决策树模型对该应用程序上的用户进行贷款倾向预测，并根据预测结果进行贷款广告信息的推送。
  18. 如权利要求17所述的计算机可读存储介质，其特征在于，所述决策树模型的生成过程如下：
    将应用程序的用户中已通过该应用程序转化为贷款用户的标注为具有贷款倾向用户，并获取具有贷款倾向用户对应的贷款用户数据中标注的与贷款相关的属性标签；
    根据标注的具有贷款倾向用户及其对应的与贷款相关的属性标签，采用递归的方法将与贷款相关的属性标签划分为多个子集；
    通过增益信息选择决策树模型中各个节点所对应的最佳属性标签，采用分类回归算法训练得到决策树模型。
  19. 如权利要求17所述的计算机可读存储介质，其特征在于，所述步骤三包括：
    获取该应用程序上用户的关联用户数据，从该用户的关联用户数据中得到与贷款相关的属性标签；
    根据训练得到的决策树模型加载该用户的与贷款相关的属性标签；
    递归遍历所述决策树模型，查找该用户的与贷款相关的属性标签对应的决策树叶子分类节点，由所述叶子节点分析出该用户是否为具有贷款倾向的用户；
    针对由所述决策树模型分析出的具有贷款倾向的用户，通过短信、站内信和/或消息推送push的方式推送预设的贷款广告信息，或者，在该用户应用程序界面的预设位置展示预设的贷款广告信息。
  20. 如权利要求18所述的计算机可读存储介质，其特征在于，所述步骤三包括：
    获取该应用程序上用户的关联用户数据，从该用户的关联用户数据中得到与贷款相关的属性标签；
    根据训练得到的决策树模型加载该用户的与贷款相关的属性标签；
    递归遍历所述决策树模型，查找该用户的与贷款相关的属性标签对应的决策树叶子分类节点，由所述叶子节点分析出该用户是否为具 有贷款倾向的用户；
    针对由所述决策树模型分析出的具有贷款倾向的用户，通过短信、站内信和/或消息推送push的方式推送预设的贷款广告信息，或者，在该用户应用程序界面的预设位置展示预设的贷款广告信息。
PCT/CN2018/077677 2017-09-30 2018-02-28 在应用程序中推送贷款广告的方法、电子装置及介质 WO2019062021A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710916516.4A CN107895277A (zh) 2017-09-30 2017-09-30 在应用程序中推送贷款广告的方法、电子装置及介质
CN201710916516.4 2017-09-30

Publications (1)

Publication Number Publication Date
WO2019062021A1 true WO2019062021A1 (zh) 2019-04-04

Family

ID=61802771

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/077677 WO2019062021A1 (zh) 2017-09-30 2018-02-28 在应用程序中推送贷款广告的方法、电子装置及介质

Country Status (2)

Country Link
CN (1) CN107895277A (zh)
WO (1) WO2019062021A1 (zh)

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108711103A (zh) * 2018-06-04 2018-10-26 中国平安人寿保险股份有限公司 个人贷款偿还风险预测方法、装置、计算机设备以及介质
CN108898429B (zh) * 2018-06-19 2023-04-18 平安科技（深圳）有限公司 电子装置、偏好倾向预测方法和计算机可读存储介质
CN109447126A (zh) * 2018-09-27 2019-03-08 长威信息科技发展股份有限公司 一种实体和实体属性动态聚合构造人物画像的方法及设备
CN109543442A (zh) * 2018-10-12 2019-03-29 平安科技（深圳）有限公司 数据安全处理方法、装置、计算机设备及存储介质
CN109272408A (zh) * 2018-10-31 2019-01-25 平安科技（深圳）有限公司 车贷款金融产品智能推荐方法、装置、设备及介质
CN109740620B (zh) * 2018-11-12 2023-09-26 平安科技（深圳）有限公司 人群画像分类模型的建立方法、装置、设备和存储介质
CN110246012A (zh) * 2019-06-14 2019-09-17 哈尔滨哈银消费金融有限责任公司 基于社交数据的消费金融产品推荐方法、装置和设备
CN110324418B (zh) * 2019-07-01 2022-09-20 创新先进技术有限公司 基于用户关系推送业务的方法和装置
CN111062422B (zh) * 2019-11-29 2023-07-14 上海观安信息技术股份有限公司 一种套路贷体系化识别方法及装置
CN111476657B (zh) * 2020-04-07 2023-07-18 度小满科技(北京)有限公司 信息推送方法、装置及系统
CN111583016B (zh) * 2020-04-09 2021-06-01 上海淇毓信息科技有限公司 基于gbst的用户推荐方法、装置及电子设备
CN111583015A (zh) * 2020-04-09 2020-08-25 上海淇毓信息科技有限公司 一种信贷申请分类方法、装置及电子设备
CN112328869A (zh) * 2020-09-28 2021-02-05 苏宁金融科技(南京)有限公司 一种用户贷款意愿的预测方法、装置及计算机系统
CN112819529A (zh) * 2021-01-30 2021-05-18 北京奇保信安科技有限公司 基于广告dsp的资源保障产品信息推送方法、装置和电子设备

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050004860A1 (en) * 2003-07-01 2005-01-06 E-Loan, Inc. Method and system for determining optimal loan options
CN105608604A (zh) * 2015-12-30 2016-05-25 合一网络技术(北京)有限公司 一种品牌广告效果优化的连续计算方法
CN106780012A (zh) * 2016-12-29 2017-05-31 深圳微众税银信息服务有限公司 一种互联网信贷方法及系统

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102521770A (zh) * 2011-12-26 2012-06-27 Tcl集团股份有限公司 在终端设备的应用程序界面内置广告的方法及系统
CN103458042B (zh) * 2013-09-10 2016-06-01 上海交通大学 一种微博广告用户检测方法
CN104966215A (zh) * 2015-07-01 2015-10-07 小米科技有限责任公司 广告信息推送方法及装置
CN106779810A (zh) * 2016-11-25 2017-05-31 北京金山安全软件有限公司 广告展现方法、装置和终端设备
CN106682952A (zh) * 2017-01-17 2017-05-17 北京薇途科技有限公司 一种基于场景热点的精准广告推荐系统及方法
CN107103514A (zh) * 2017-04-25 2017-08-29 北京京东尚科信息技术有限公司 商品性别标签确定方法和装置

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050004860A1 (en) * 2003-07-01 2005-01-06 E-Loan, Inc. Method and system for determining optimal loan options
CN105608604A (zh) * 2015-12-30 2016-05-25 合一网络技术(北京)有限公司 一种品牌广告效果优化的连续计算方法
CN106780012A (zh) * 2016-12-29 2017-05-31 深圳微众税银信息服务有限公司 一种互联网信贷方法及系统

Also Published As

Publication number Publication date
CN107895277A (zh) 2018-04-10

Similar Documents

Publication Publication Date Title
WO2019062021A1 (zh) 在应用程序中推送贷款广告的方法、电子装置及介质
WO2018188576A1 (zh) 资源推送方法及装置
WO2019061989A1 (zh) 贷款风险控制方法、电子装置及可读存储介质
WO2019061994A1 (zh) 电子装置、保险产品推荐方法、系统及计算机可读存储介质
WO2019169756A1 (zh) 产品推荐方法、装置及存储介质
US9430776B2 (en) Customized E-books
JP6573418B2 (ja) データソースに基づく業務カスタマイズ装置、方法、システム及び記憶媒体
US20220101161A1 (en) Probabilistic methods and systems for resolving anonymous user identities based on artificial intelligence
US9715486B2 (en) Annotation probability distribution based on a factor graph
JP2017533531A (ja) フォーカスト・センチメント分類
WO2019061664A1 (zh) 电子装置、基于用户上网数据的产品推荐方法及存储介质
CN111125566B (zh) 信息获取方法和装置、电子设备和存储介质
US10147020B1 (en) System and method for computational disambiguation and prediction of dynamic hierarchical data structures
Chamberlain et al. Probabilistic inference of twitter users’ age based on what they follow
US11640420B2 (en) System and method for automatic summarization of content with event based analysis
CN116821475B (zh) 基于客户数据的视频推荐方法、装置及计算机设备
CN110880006A (zh) 用户分类方法、装置、计算机设备和存储介质
US20220019689A1 (en) Privacy Preserving Server-Side Personalized Content Selection
WO2021081914A1 (zh) 推送对象确定方法、装置、终端设备及存储介质
CN109685091B (zh) 使用贝叶斯方式的数字体验目标确定
US10529002B2 (en) Classification of visitor intent and modification of website features based upon classified intent
US20230266966A1 (en) User support content generation
CN114154052A (zh) 信息推荐方法、装置、计算机设备及存储介质
CN111127057B (zh) 一种多维用户画像恢复方法
Singh et al. Email Personalization and User Profiling Using RANSAC Multi Model Response Regression Based Optimized Pruning Extreme Learning Machines and Gradient Boosting Trees

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: 18863166

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205 DATED 24/09/2020)

122 Ep: pct application non-entry in european phase

Ref document number: 18863166

Country of ref document: EP

Kind code of ref document: A1