WO2020244152A1

WO2020244152A1 - Data pushing method and apparatus, computer device, and storage medium

Info

Publication number: WO2020244152A1
Application number: PCT/CN2019/118034
Authority: WO
Inventors: 乐志能
Original assignee: 平安科技（深圳）有限公司
Priority date: 2019-06-05
Filing date: 2019-11-13
Publication date: 2020-12-10
Also published as: CN110335064A

Abstract

A data pushing method, comprising: receiving a product acquisition request corresponding to a user identifier sent by a terminal; querying, according to the product acquisition request, product feature data corresponding to a preconfigured candidate product identifier and user feature data corresponding to the user identifier; inputting the user feature data and the product feature data into a trained product prediction model, and determining, by means of the product prediction model, the product identifier to be pushed; and pushing the product data corresponding to the product identifier to the terminal for display.

Description

Data pushing method, device, computer equipment and storage medium

This application claims the priority of a Chinese patent application filed with the Chinese Patent Office on June 5, 2019, with the application number 2019104854757 and the application name "Product push method, device, computer equipment and storage medium", the entire content of which is incorporated by reference In this application.

Technical field

This application relates to a data push method, device, computer equipment and storage medium.

Background technique

With the development of Internet technology, due to the advantages of the Internet’s high exposure rate and diverse display forms, more and more products are pushed to users through the Internet, which improves the efficiency of product delivery to a certain extent. When pushing products with the help of the Internet, a designated product display area is usually set on a browser page or a client interface to display one or more products to users through the product display area. The product display area can be set to display multiple products in turn within a specified time period to increase the exposure rate of each product.

However, the inventor realizes that at present, for a product display area that can display multiple products in turn, multiple products to be displayed and corresponding rotation display rules are usually pre-configured, and the products are displayed in the product display area in turn according to the rotation display rules. Multiple products. Rotational display rules are usually based on the total display time, total traffic consumption, total number of impressions, or random display. However, the products displayed in the product display area in turn based on this kind of rotating display rules may not be the products that users are actually interested in, resulting in low product conversion rates.

Summary of the invention

According to various embodiments disclosed in the present application, a data push method, device, computer equipment, and storage medium are provided.

A data push method includes:

Receive the product acquisition request corresponding to the user ID sent by the terminal;

Query the product feature data corresponding to the preconfigured candidate product identifier and the user feature data corresponding to the user identifier according to the product acquisition request;

Input the user characteristic data and the product characteristic data into a trained product prediction model, and determine the pushed product identifier through the product prediction model; and

Push the product data corresponding to the product identifier to the terminal for display.

A data push device includes:

The receiving module is used to receive the product acquisition request corresponding to the user ID sent by the terminal;

The query module is configured to query the product feature data corresponding to the pre-configured candidate product identifier and the user feature data corresponding to the user identifier according to the product acquisition request;

The prediction module is used to input the user feature data and the product feature data into a trained product prediction model, and determine the product identifier to be pushed through the product prediction model; and

The push module is used to push the product data corresponding to the product identifier to the terminal for display.

A computer device includes a memory and one or more processors. The memory stores computer readable instructions. When the computer readable instructions are executed by the one or more processors, the one or more Each processor implements the steps of the data pushing method provided in any embodiment of the present application.

One or more non-volatile computer-readable storage media storing computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors implement any of the present application The steps of the data push method provided in one embodiment.

The details of one or more embodiments of the application are set forth in the following drawings and description. Other features and advantages of this application will become apparent from the description, drawings and claims.

Description of the drawings

In order to more clearly describe the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative work.

Fig. 1 is an application scenario diagram of a data push method according to one or more embodiments.

Fig. 2 is a schematic flowchart of a data push method according to one or more embodiments.

Fig. 3 is a schematic flowchart of a data pushing method in another embodiment.

Fig. 4 is a schematic flowchart of a data pushing method in another embodiment.

Fig. 5 is a block diagram of a data pushing device according to one or more embodiments.

Figure 6 is a block diagram of a computer device according to one or more embodiments.

Detailed ways

In order to make the technical solutions and advantages of the present application clearer, the following further describes the present application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the application, and not used to limit the application.

The data push method provided in this application can be applied to the application environment shown in FIG. 1. The terminal 102 communicates with the server 104 through the network through the network. The server 104 receives the product acquisition request corresponding to the user ID sent by the terminal 102, queries the product feature data corresponding to the pre-configured candidate product ID and the user feature data corresponding to the user ID according to the product acquisition request, and compares the acquired user features The data and product feature data are input into the trained product prediction model to determine the pushed product identification through the product prediction model, and then push the product data corresponding to the product identification to the terminal 102 to display the product data to the user through the terminal 102. The terminal 102 may be, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers, and portable wearable devices. The server 104 may be implemented by an independent server or a server cluster composed of multiple servers.

In one of the embodiments, as shown in FIG. 2, a data push method is provided. Taking the method applied to the server in FIG. 1 as an example for description, the method includes the following steps:

S202: Receive a product acquisition request corresponding to the user ID sent by the terminal.

The product acquisition request is a request used to trigger a product data acquisition operation, and is used to instruct the server to acquire and feed back the product data pushed to the user corresponding to the user identification. The user ID is used to uniquely identify the user, and specifically can be the user's ID card, mobile phone number, or a user name that can be used to uniquely identify the user.

Specifically, the terminal detects the user's preset trigger operation in real time, and when the preset trigger operation is detected, triggers the generation of a product acquisition request corresponding to the user ID of the user, and sends the generated product acquisition request to the server. The terminal can specifically detect the user's preset trigger operation for the client running on it in real time, and generate a product acquisition request based on the detected preset trigger operation. The preset trigger operation is, for example, a user trigger operation on a preset trigger control, such as a user click operation on an icon of a client installed on a terminal, or a user click or slide operation on a specific user operation interface of the client or browser.

In one of the embodiments, the server receives the product acquisition request sent by the terminal through the client or a designated webpage or browser. The server receives the product acquisition request sent by the terminal in a wired or wireless manner.

S204: Query the product feature data corresponding to the pre-configured candidate product identifier and the user feature data corresponding to the user identifier according to the product acquisition request.

Candidate product identifiers are product identifiers that can be selected as the product identifiers to be pushed. The product identifier is used to uniquely identify the product, and specifically can be a product name, serial number, or other character string that is composed of at least one of numbers, letters, and symbols and can be used to uniquely identify the product, and the product can specifically be an advertisement. Product feature data is data used to characterize the features of the product or the corresponding features, which can specifically include product type and target group data, etc., product types such as wealth management, insurance, loans, and credit cards, etc., target group data Refers to the characteristic data shared by the intended target population of the product, such as age, gender, marital status, or occupation type. User characteristic data is data that is determined based on the user's user behavior data and user attribute data and used to characterize the user's characteristics. User characteristic data such as the user's gender, age, job type, income, marital status, whether to have a house, whether to have a car, behavior preferences and product type preferences, etc.

Specifically, the server is locally pre-configured with multiple product identifiers as candidate product identifiers, and corresponding product configuration data is configured in the database for each of the pre-configured candidate product identifiers. When receiving the product acquisition request, the server selects candidate product identifiers whose product configuration data meets the preset screening conditions from the database, and queries corresponding product feature data based on the selected candidate product identifiers. Product configuration data includes the total number of remaining impressions, the total remaining duration of the impression, or the total amount of remaining impression traffic. Preset filtering conditions such as the total number of remaining impressions, the total remaining duration of the impression, or the total remaining impression traffic is greater than zero. Further, the server queries the database for the user characteristic data corresponding to the user identification according to the user identification corresponding to the product acquisition request.

In one of the embodiments, the product configuration data also includes a preset display time period of the product data. When a product acquisition request is received, the server acquires the current system time, queries the database for the candidate product identifier corresponding to the preset display segment that matches the current system time, and queries the database according to the query candidate product identifier. Product feature data.

In one of the embodiments, the product acquisition request carries the candidate product identification and the user identification. The server parses the received product acquisition request to obtain the candidate product identification and the user identification, and queries the database for product feature data corresponding to the candidate product identification and user feature data corresponding to the user identification.

S206: Input the user feature data and the product feature data into the trained product prediction model, and determine the pushed product identifier through the product prediction model.

The product prediction model is a prediction model that is obtained by model training based on the pre-acquired training sample set, and can determine the pushed product identification based on user feature data and product feature data. The product prediction model makes predictions based on the user feature data corresponding to the user identification and the product feature data corresponding to each candidate product identification. It can directly predict and output the product identification corresponding to the product data pushed to the user, or predict and output the corresponding product identification for each candidate product. User matching degree, and then screening and pushing product IDs from multiple candidate product IDs according to user matching degree.

Specifically, the server uses the product feature data and user feature data obtained according to the product acquisition request as input features, and inputs them into the trained product prediction model to determine the product data corresponding to the product data pushed to the corresponding user by means of the product prediction model Product ID, as the product ID of the push.

In one of the embodiments, the server uses the acquired user feature data and product feature data as input features, inputs the trained product prediction model to make predictions, and obtains the pushed product identification. Correspondingly, the training step of the product prediction model includes: the server separately obtains the target user characteristic data corresponding to the target user identifier and the target product characteristic data corresponding to the target product identifier, and determines from the target product identifier according to the target user characteristic data and the target product characteristic data The target product identification of the push, the target user characteristic data and the target product characteristic data are used as input characteristics, and the target product identification corresponding to the push is determined as the desired output characteristic, and the initialized product prediction model is model-trained to obtain the trained product Forecast model. It is understandable that in the above model training process, there are multiple target user identifiers, and each target user identifier corresponds to multiple target product identifiers, and one of the multiple target product identifiers serves as the target product identifier to be pushed. Therefore, each target user identifier and multiple target product identifiers corresponding to the target user identifier, and correspondingly determined push target product identifiers constitute a training sample in the training sample set.

In one of the embodiments, step S206 includes: using user feature data and product feature data as input features, inputting the trained product prediction model to make predictions, and obtaining the user matching degree corresponding to each candidate product identifier; selecting from candidate product identifiers Candidate product identifiers whose user matching degree meets the preset screening conditions are output as the pushed product identifiers.

User matching degree refers to the matching degree between the product data corresponding to the candidate product identifier and the product data desired by the user. The user matching degree can be a value within a preset value range. The larger the value, the higher the matching degree. The preset value range is from 0 to 10, the value 0 indicates that the matching degree is 0, that is, no match at all, and the value 10 indicates a complete match. The user matching degree can also be a percentage, such as 60%. The larger the percentage, the higher the matching degree. The preset screening conditions are pre-defined screening conditions, such as screening the candidate product identifiers with the highest matching degree of users as the pushed product identifiers.

Specifically, the server uses the acquired user feature data and product feature data as input features, inputs the trained product prediction model to make predictions, and obtains the user matching degree corresponding to each candidate product identifier. According to the user matching degree corresponding to each candidate product identifier, the server selects the candidate product identifier with the highest user matching degree from the multiple candidate product identifiers as the pushed product identifier.

In one of the embodiments, when the user matching degree corresponding to each candidate product identifier is predicted by the product prediction model, the server prioritizes the multiple candidate product identifiers according to the predicted user matching degree to obtain the product identification sequence. The server selects the first or last candidate product identifier in the ranking position from the product identifier sequence as the pushed product identifier, depending on the sorting method.

In the above embodiment, the trained product prediction model is used to predict according to user feature data and product feature data, and the degree of matching between the product data corresponding to each candidate product identifier and the product data expected by the user is obtained to filter out the degree of matching The highest product data is displayed to users through the terminal, which improves the accuracy of product data push, which can increase the conversion rate of the product.

S208: Push the product data corresponding to the product identifier to the terminal for display.

Specifically, when the product identifier to be pushed is determined with the aid of the trained product prediction model, the server queries the corresponding product data from the database according to the determined product identifier, and pushes the queried product data to the terminal, so as to use the terminal The product data is displayed to the corresponding user.

In one of the embodiments, the server obtains the product data corresponding to the product identifier from other computer devices or online networks according to the product identifier. Other computer equipment, such as a server or terminal for configuring or storing product data corresponding to the product identifier.

In the above data push method, upon receiving the product acquisition request corresponding to the user ID sent by the terminal, according to the product acquisition request, query the product feature data corresponding to the pre-configured candidate product ID and the user feature data corresponding to the user ID, and query The obtained product feature data and user feature data are input into the trained product prediction model, the product identification to be pushed is determined by the product prediction model, and then the product data corresponding to the product identification is pushed to the terminal, and the product data has been displayed to the corresponding terminal through the terminal user. The product data determined based on the user characteristic data and the product characteristic data is the product data that meets the user's expectations, and the pushed product data determined is the product data that meets the user's preferences, which improves the accuracy of product data push, thereby improving product conversion rate. Moreover, the product data to be pushed is determined based on the trained product prediction model according to user characteristic data and product characteristic data, which further improves the accuracy of product data push, thereby further improving the conversion rate of the product.

In one of the embodiments, step S204 includes: determining the product display area identifier according to the product acquisition request, and user historical behavior data and user attribute data corresponding to the user identifier; analyzing the user historical behavior data and user attribute data to obtain the user identifier Corresponding user characteristic data; query the product characteristic data corresponding to the pre-configured candidate product identifier according to the product display area identifier.

The product display area identifier is used to uniquely identify the product display area. The product display area is an area used to display the pushed product data. Taking products as advertisements as an example, the product display area refers to advertising spaces. User historical behavior data refers to data related to user historical behavior. User historical behavior data refers to the data corresponding to the user's operation behaviors such as clicking, browsing, or collecting each product data displayed in the product display area. According to the user's historical behavior data, the user's preference for various products can be determined. User attribute data includes user basic data and user interest preference data. Basic user data such as user's gender, age, job type, marital status, income, whether there is a car or house, etc., user interest preference data such as user preference for financial management or user preference for shopping Wait.

Specifically, when a user’s preset trigger operation is detected, the terminal queries the preconfigured product display area identifier according to the preset trigger operation, and generates a product acquisition request based on the queried product display area identifier and the corresponding user identifier, And send the generated product acquisition request to the server. The server parses the received product acquisition request, and obtains the product display area identifier and the user identifier. The server searches the database for the candidate product identification pre-configured corresponding to the product display area identification according to the parsed product display area identification, and searches the corresponding product feature data according to the searched candidate product identification. Further, the server queries the corresponding user historical behavior data and user attribute data in the database according to the user identifier, and performs statistical analysis on the queried user historical behavior data and user attribute data to obtain user characteristic data corresponding to the user identifier.

In one of the embodiments, the server obtains basic user data and user interest preference data according to user attribute data, and determines user interest feature data, such as the type of product the user prefers, according to the user interest preference data. The server performs statistical analysis on the user's historical behavior data to extract the user's preferred product type and/or product identifier from the user's historical behavior data. The server determines the user characteristic data according to the user's basic data and user interest characteristic data, as well as the product type and/or product identifier extracted based on the user's historical behavior data. For example, the user’s basic data is gender male, age 25, job type is technology research and development, marital status is unmarried, user interest characteristic data is preferred wealth management products, and the user’s preferred product type is determined based on user historical behavior data as fund. It can be determined that the user prefers stock funds and/or investment funds, and the corresponding user characteristic data is gender male, age 25, job type is technology research and development, marital status is unmarried, prefers wealth management products, and the preferred product type is funds. User characteristic data may also include user preferences for stock funds and/or investment funds. In this way, according to the above-mentioned data push method, the product data of fund products can be pushed to the terminal for display.

In the foregoing embodiment, the product feature data is determined according to the product display area identifier to ensure that the pushed product data is product data matching the product display area, which improves the accuracy of product data push. Further, the user historical behavior data and user attribute data corresponding to the user identification are characterized to obtain user characteristic data corresponding to the user identification, so that the user characteristic data can better reflect the characteristics and preferences of the user. In this way, the product data pushed to the user determined according to the user characteristic data and the product characteristic data has a higher degree of matching with the product data expected by the user, which improves the accuracy of product data push, thereby increasing the conversion rate of the product.

In one of the embodiments, step S208 includes: querying the product data and product configuration data corresponding to the product identification; generating corresponding product display strategy parameters according to the product configuration data; pushing the product data to the terminal for display, and obtaining the product data in real time When the display statistical parameters are consistent with the product display strategy parameters, the product identifiers to be pushed again are filtered from the candidate product identifiers according to the user matching degree; the product data corresponding to the product identifiers pushed again is pushed to the terminal for display.

The product display strategy parameter is a quantitative parameter in the product display strategy, which is used to characterize the strategy of displaying product data, such as the total display time of product data, or the total traffic consumption when displaying product data. Display statistics parameters refer to display parameters that are counted in real time during the display of product data, such as total display duration or total traffic consumption.

Specifically, when determining the product identifier to be pushed, the server queries the database for corresponding product data and product configuration parameters according to the determined product identifier, generates corresponding product display strategy parameters based on the queried product configuration parameters, and generates The product display strategy parameters are cached locally. The server pushes the queried product data to the terminal for display, counts the display statistical parameters corresponding to the product data in real time, and compares the statistical display statistical parameters with the cached product display strategy parameters. When the display statistical parameters are consistent with the product display strategy parameters, the server selects the candidate product identifier with the highest user matching degree from the multiple candidate product identifiers that have not yet pushed the corresponding product data according to the user matching degree corresponding to each candidate product identifier, as Product ID that is pushed again. The server queries the corresponding product data according to the product identification pushed again, and pushes the queried product data to the terminal for display.

In one of the embodiments, for the product identification that is pushed again, the server generates corresponding product display strategy parameters in the above-mentioned manner, counts the corresponding display statistical parameters in real time, and continues when the product display strategy parameters are consistent with the corresponding display statistical parameters Perform the step of screening out product identifiers to be pushed again from candidate product identifiers according to the user matching degree, and stop the iteration until the preset stopping conditions are met. The preset stop condition, for example, the product data corresponding to each candidate product identifier is pushed sequentially in the above-mentioned manner, or the server obtains a product data push termination instruction, etc.

In one of the embodiments, the server dynamically updates the product configuration data corresponding to the product identifier in the database according to the display statistical parameters calculated in real time. When the display statistical parameters are consistent with the product display strategy parameters, the server obtains the product configuration data corresponding to the currently pushed product identifier from the database. When the currently acquired product configuration data meets the preset filtering conditions, the server determines the currently pushed product identifier as the product identifier to be pushed again.

In the above embodiment, the product data pushed to the terminal for display is dynamically updated according to the product configuration data, which ensures the diversity of product data push, improves the push efficiency of product data, and improves the overall conversion rate of the product. At the same time, it is pushed sequentially according to the user matching degree of each candidate product identifier, so that the user can quickly locate the product data of interest among the pushed product data, and further improve the efficiency of product data push.

In one of the embodiments, the candidate product identifiers whose user matching degree meets the preset filtering conditions are filtered from the candidate product identifiers, and the product identifiers to be pushed include: according to the user matching degree corresponding to each candidate product identifier, compare each candidate product identifier The identifications are prioritized to obtain the product identification sequence; multiple candidate product identifications that meet the preset screening conditions are selected from the product identification sequence as the pushed product identification.

The product identification sequence is a sequence composed of multiple candidate product identifications arranged in a specific priority order. The preset screening conditions include, for example, screening out a preset number of candidate product identifiers that are ranked higher in the product identification sequence, or screening out candidate product identifiers whose user matching degree reaches a preset matching degree threshold.

Specifically, there are multiple product identifiers pushed. When determining the user matching degree corresponding to each candidate product identifier, the server prioritizes the candidate product identifiers according to the user matching degree corresponding to each candidate product identifier, and obtains a product identification sequence composed of the multiple candidate product identifiers. The higher the user matching degree, the higher the priority of the corresponding candidate product identification, that is, the higher the ranking position of the candidate product identification in the product identification sequence. The server screens out the preset number of candidate product identifiers with the highest ranking position from the product identifier sequence as the pushed product identifiers. The preset number can be customized, such as 3.

In one of the embodiments, the server selects a preset number of candidate product identifiers from the multiple candidate product identifiers according to the order of the user matching degree from high to low, as the pushed product identifiers. The server may filter out candidate product identifiers with a user matching degree greater than or equal to a preset matching degree threshold from the multiple candidate product identifiers, as the pushed product identifiers. The preset matching degree threshold is, for example, 7 or 70%.

In the above embodiment, the product identifiers to be pushed are screened among the pre-configured multiple candidate product identifiers according to the user matching degree. Under the premise of ensuring the diversity of the product data pushed, the accuracy of product data push is improved, and the product can be improved. The overall conversion rate.

In one of the embodiments, step S208 includes: respectively querying product data and product configuration data corresponding to each product identifier; generating product display strategy parameters corresponding to the product data according to each product configuration data; displaying the product data and the corresponding product The strategy parameters are pushed to the terminal to instruct the terminal to display product data in turn according to the product display strategy parameters.

Product display strategy parameters such as the number of exposures of product data, the duration of a single exposure, or the total amount of traffic consumed by a single exposure.

Specifically, when multiple product identifiers that are pushed from candidate product identifiers are filtered out according to the user’s matching degree, the server queries the corresponding product data and product configuration data according to the selected product identifiers, and identifies the corresponding product according to each product identifier. The configuration data respectively generates corresponding product display strategy parameters. The server pushes the product data and product display strategy parameters corresponding to each product identifier to the terminal. The terminal displays the corresponding product data in the corresponding product display area in turn according to the received product display strategy parameters.

Taking products as an example of advertising, the product display area is a carousel advertising space. When the terminal receives multiple advertisement data pushed by the server and the advertising display strategy parameter corresponding to each advertising data, it will display the carousel advertising space in accordance with each advertising display strategy parameter. Rotate the corresponding advertising data.

In the above embodiment, the product display strategy parameters for displaying the corresponding product data in turns are determined according to the product configuration data of each product identification, so that the terminal displays the product data to users in turns in the product display area according to the product display strategy parameters, which improves the display in turn. Conversion rate of products.

In one of the embodiments, in the above data push method, the training step of the product prediction model includes: obtaining target user characteristic data corresponding to the target user identification and target product characteristic data corresponding to the target product identification; according to the target user characteristic data and the target The product feature data is used to obtain the target user matching degree corresponding to each target product identifier; the target user feature data and target product feature data are used as input features, and the corresponding target user matching degree is used as the desired output feature, and the initial product prediction model is performed Model training to obtain the trained product prediction model.

Specifically, the server obtains target user characteristic data corresponding to each of the multiple target user identifiers, and respectively obtains target product characteristic data corresponding to each target product identifier for the multiple target product identifiers corresponding to each target user identifier. For each target user identifier, the server identifies each target product in the multiple target product identifiers according to the target user characteristic data corresponding to the target user identifier and the target product characteristic data corresponding to the corresponding multiple target product identifiers. Marking is performed to obtain the target user matching degree corresponding to each target product identifier. The server takes the target user feature data corresponding to each target user identifier and the corresponding multiple target product feature data as input features, takes the corresponding multiple target user matching degrees as the desired output features, and performs model training on the initialized product prediction model. Get the trained product prediction model.

In one of the embodiments, the machine learning algorithm involved in the above model training process may be a logistic regression algorithm, decision tree, random forest, neural network, support vector machine, and so on.

For example, taking the logistic regression algorithm as an example, the corresponding logistic regression function is:

x is the input feature vector, α is the weight parameter, and h(x) is the output feature vector. Through continuous training, the cost function of logistic regression is minimized, and the optimal value of the weight parameter is determined correspondingly, so as to obtain the logistic regression model after training.

In the foregoing embodiment, the initialized product prediction model is trained in advance to obtain the trained product prediction model, so that when data is pushed, the trained product prediction model is used to determine the user matching degree corresponding to each candidate product identifier, and then Determining the pushed product identification according to the user matching degree improves the accuracy of product data push, thereby increasing the conversion rate of the product.

As shown in Figure 3, in one of the embodiments, a data push method is provided, and the method specifically includes the following steps:

S302: Receive a product acquisition request corresponding to the user ID sent by the terminal.

S304: Determine a product display area identifier, and user historical behavior data and user attribute data corresponding to the user identifier according to the product acquisition request.

S306: Analyze user historical behavior data and user attribute data to obtain user characteristic data corresponding to the user identifier.

S308: Query the product feature data corresponding to the pre-configured candidate product identifier according to the product display area identifier.

S310: Using user feature data and product feature data as input features, input the trained product prediction model to make predictions, and obtain the user matching degree corresponding to each candidate product identifier.

S312: Filter candidate product identifiers whose user matching degree meets preset screening conditions from candidate product identifiers, and use them as pushed product identifiers.

S314: Query the product data and product configuration data corresponding to the product identification.

S316: Generate corresponding product display strategy parameters according to the product configuration data.

S318: Push the product data to the terminal for display, and obtain display statistical parameters of the product data in real time.

S320: When the display statistical parameters are consistent with the product display strategy parameters, filter the product identifiers to be pushed again from the candidate product identifiers according to the user matching degree.

S322: Push the product data corresponding to the product identifier pushed again to the terminal for display.

In the above embodiment, with the aid of the trained product prediction model, the user matching degree corresponding to each candidate product identifier is predicted based on the user characteristic data and the product characteristic data, and then the pushed product identifiers are filtered according to the user matching degree and the product data is pushed. , Improve the accuracy of product data push, which can increase the product conversion rate. Further, the displayed product data is dynamically updated according to the display statistical parameters and product display strategy parameters corresponding to the currently displayed product data, so that the user can choose the product data that he or she is interested in in the updated display, so as to improve the overall conversion rate of the product.

As shown in Figure 4, in one of the embodiments, a data push method is provided, and the method specifically includes the following steps:

S402: Receive a product acquisition request corresponding to the user ID sent by the terminal.

S404: Determine a product display area identifier, and user historical behavior data and user attribute data corresponding to the user identifier according to the product acquisition request.

S406: Analyze user historical behavior data and user attribute data to obtain user characteristic data corresponding to the user identifier.

S408: Query the product feature data corresponding to the pre-configured candidate product identifier according to the product display area identifier.

S410: Using user feature data and product feature data as input features, input the trained product prediction model for prediction, and obtain the user matching degree corresponding to each candidate product identifier.

S412: Prioritize each candidate product identifier according to the user matching degree corresponding to each candidate product identifier to obtain a product identifier sequence.

S414: Filter out multiple candidate product identifiers that meet the preset screening conditions from the product identifier sequence as the pushed product identifiers.

S416, respectively query the product data and product configuration data corresponding to each product identifier.

S418: Generate product display strategy parameters corresponding to the product data according to each product configuration data.

S420: Push the product data and corresponding product display strategy parameters to the terminal to instruct the terminal to display the product data in turn according to the product display strategy parameters.

In the above embodiment, the product prediction model is used to predict the user matching degree of each candidate product identifier according to the user characteristic data and the product characteristic data, and then multiple pushed product identifiers are screened according to the user matching degree, and the multiple product identifiers Product data is pushed to the terminal for display in turn, so that users can quickly filter their interested product data from multiple product data, which improves the overall product conversion rate.

It should be understood that although the various steps in the flowcharts of FIGS. 2-4 are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless specifically stated in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least some of the steps in Figures 2-4 may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. These sub-steps or stages The execution order of is not necessarily performed sequentially, but may be performed alternately or alternately with at least a part of other steps or sub-steps or stages of other steps.

In one of the embodiments, as shown in FIG. 5, a data pushing device 500 is provided, including: a receiving module 502, a query module 504, a prediction module 506, and a pushing module 508, wherein:

The receiving module 502 is configured to receive the product acquisition request corresponding to the user ID sent by the terminal;

The query module 504 is configured to query the product feature data corresponding to the pre-configured candidate product identifier and the user feature data corresponding to the user identifier according to the product acquisition request;

The prediction module 506 is used to input user feature data and product feature data into the trained product prediction model, and determine the pushed product identification through the product prediction model;

The pushing module 508 is used to push the product data corresponding to the product identifier to the terminal for display.

In one of the embodiments, the query module 504 is further configured to determine the product display area identifier and the user historical behavior data and user attribute data corresponding to the user identifier according to the product acquisition request; analyze the user historical behavior data and user attribute data, Obtain the user feature data corresponding to the user ID; query the product feature data corresponding to the pre-configured candidate product ID according to the product display area ID.

In one of the embodiments, the prediction module 506 is also used to use user feature data and product feature data as input features, input the trained product prediction model to predict, and obtain the user matching degree corresponding to each candidate product identifier; Candidate product identifiers whose user matching degree meets the preset filtering conditions are filtered out from the identifiers, and used as the pushed product identifiers.

In one of the embodiments, the push module 508 is also used to query product data and product configuration data corresponding to the product identification; generate corresponding product display strategy parameters according to the product configuration data; push the product data to the terminal for display, and real-time Obtain the display statistical parameters of the product data; when the display statistical parameters are consistent with the product display strategy parameters, filter out the product identifiers to be pushed again from the candidate product identifiers according to the user matching degree; push the product data corresponding to the product identifiers pushed again to the terminal To show.

In one of the embodiments, the prediction module 506 is also used to prioritize each candidate product identifier according to the user matching degree corresponding to each candidate product identifier to obtain the product identifier sequence; and filter the product identifier sequence to match the preset The multiple candidate product identifiers of the screening conditions are used as the pushed product identifiers.

In one of the embodiments, the push module 508 is also used to query product data and product configuration data corresponding to each product identifier; generate product display strategy parameters corresponding to the product data according to each product configuration data; combine the product data and corresponding product data The product display strategy parameters of the product are pushed to the terminal to instruct the terminal to display product data in turn according to the product display strategy parameters.

In one of the embodiments, the prediction module 506 is also used to perform the training step of the product prediction model, including: obtaining target user characteristic data corresponding to the target user identification and target product characteristic data corresponding to the target product identification; according to the target user characteristic data With the target product feature data, the target user matching degree corresponding to each target product identifier is obtained; the target user feature data and target product feature data are used as input features, and the corresponding target user matching degree is used as the desired output feature to predict the initial product The model performs model training to obtain a trained product prediction model.

For the specific limitation of the data pushing device, please refer to the above limitation on the data pushing method, which will not be repeated here. Each module in the above-mentioned data pushing device can be implemented in whole or in part by software, hardware, and a combination thereof. The foregoing modules may be embedded in the form of hardware or independent of the processor in the computer device, or may be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the foregoing modules.

In one of the embodiments, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. 6. The computer equipment includes a processor, a memory, a network interface, and a database connected through a system bus. The processor of the computer device is used to provide calculation and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer readable instructions, and a database. The internal memory provides an environment for the operation of the operating system and computer-readable instructions in the non-volatile storage medium. The database of the computer device is used to store the product feature data and product data corresponding to the pre-configured candidate product identification, as well as the user feature data corresponding to the user identification and the trained product prediction model. The network interface of the computer device is used to communicate with an external terminal through a network connection. The computer-readable instructions are executed by the processor to realize a data push method.

Those skilled in the art can understand that the structure shown in FIG. 6 is only a block diagram of part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may Including more or fewer parts than shown in the figure, or combining some parts, or having a different arrangement of parts.

A computer device including a memory and one or more processors. The memory stores computer readable instructions. When the computer readable instructions are executed by one or more processors, the one or more processors can realize any one of the present application. The steps of the data push method provided in the embodiment.

One or more non-volatile computer-readable storage media storing computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors implement any one of the embodiments of the present application. Provide the steps of the data push method.

Persons of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by computer-readable instructions to instruct relevant hardware. The computer-readable instructions can be stored in a non-volatile computer readable. In the storage medium, when the computer-readable instructions are executed, they may include the procedures of the above-mentioned method embodiments. Any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and/or volatile memory. Non-volatile memory may include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not a limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous chain Channel (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction between the combinations of these technical features, they should It is considered as the range described in this specification.

The above-mentioned embodiments only express several implementation manners of the present application, and the description is relatively specific and detailed, but it should not be understood as a limitation on the scope of the invention patent. It should be pointed out that for those of ordinary skill in the art, without departing from the concept of this application, several modifications and improvements can be made, and these all fall within the protection scope of this application. Therefore, the scope of protection of the patent of this application shall be subject to the appended claims.

Claims

A data push method, including:

Receive the product acquisition request corresponding to the user ID sent by the terminal;

Query the product feature data corresponding to the preconfigured candidate product identifier and the user feature data corresponding to the user identifier according to the product acquisition request;

Input the user characteristic data and the product characteristic data into a trained product prediction model, and determine the pushed product identifier through the product prediction model; and

Push the product data corresponding to the product identifier to the terminal for display.
The method according to claim 1, wherein the querying the product feature data corresponding to the preconfigured candidate product identifier and the user feature data corresponding to the user identifier according to the product acquisition request comprises:

Determining a product display area identifier, and user historical behavior data and user attribute data corresponding to the user identifier according to the product acquisition request;

Analyze the user historical behavior data and the user attribute data to obtain user characteristic data corresponding to the user identifier; and

Query the product feature data corresponding to the pre-configured candidate product identifier according to the product display area identifier.
The method according to claim 1, wherein the inputting the user characteristic data and the product characteristic data into a trained product prediction model, and determining the pushed product identifier through the product prediction model, comprises:

Using the user feature data and the product feature data as input features, input the trained product prediction model to make predictions, and obtain the user matching degree corresponding to each of the candidate product identifiers; and

From the candidate product identifiers, the candidate product identifiers whose user matching degree meets the preset screening conditions are selected as the pushed product identifiers.
The method according to claim 3, wherein the pushing the product data corresponding to the product identifier to the terminal for display comprises:

Query product data and product configuration data corresponding to the product identifier;

Generate corresponding product display strategy parameters according to the product configuration data;

Pushing the product data to the terminal for display, and obtaining display statistical parameters of the product data in real time;

When the display statistical parameters are consistent with the product display strategy parameters, filter out product identifiers to be pushed again from the candidate product identifiers according to the user matching degree; and

Push the product data corresponding to the product identifier pushed again to the terminal for display.
The method according to claim 3, wherein the screening of candidate product identifications whose user matching degree meets a preset screening condition from the candidate product identifications, as the pushed product identification, comprises:

Prioritize each candidate product identifier according to the user matching degree corresponding to each candidate product identifier to obtain a product identifier sequence; and

A plurality of candidate product identifiers that meet the preset screening conditions are screened out from the product identifier sequence as the pushed product identifiers.
The method according to claim 5, wherein the pushing the product data corresponding to the product identifier to the terminal for display comprises:

Respectively query the product data and product configuration data corresponding to each of the product identifiers;

Generating product display strategy parameters corresponding to the product data according to each of the product configuration data; and

Push the product data and the corresponding product display strategy parameters to the terminal to instruct the terminal to display the product data in turn according to the product display strategy parameters.
The method according to any one of claims 1 to 6, wherein the training step of the product prediction model comprises:

Acquiring target user characteristic data corresponding to the target user identifier and target product characteristic data corresponding to the target product identifier;

According to the target user characteristic data and the target product characteristic data, obtain the target user matching degree corresponding to each target product identifier; and

Use the target user feature data and the target product feature data as input features, and use the corresponding target user matching degree as the desired output feature, and perform model training on the initialized product prediction model to obtain a trained product prediction model .
A data push device includes:

The receiving module is used to receive the product acquisition request corresponding to the user ID sent by the terminal;

The query module is configured to query the product feature data corresponding to the pre-configured candidate product identifier and the user feature data corresponding to the user identifier according to the product acquisition request;

The prediction module is used to input the user feature data and the product feature data into a trained product prediction model, and determine the product identifier to be pushed through the product prediction model; and

The push module is used to push the product data corresponding to the product identifier to the terminal for display.
A computer device includes a memory and one or more processors. The memory stores computer readable instructions. When the computer readable instructions are executed by the one or more processors, the one or more Each processor performs the following steps:

Receive the product acquisition request corresponding to the user ID sent by the terminal;

Query the product feature data corresponding to the preconfigured candidate product identifier and the user feature data corresponding to the user identifier according to the product acquisition request;

Input the user characteristic data and the product characteristic data into a trained product prediction model, and determine the pushed product identifier through the product prediction model; and

Push the product data corresponding to the product identifier to the terminal for display.
The computer device according to claim 9, wherein the processor further executes the following steps when executing the computer-readable instruction:

Determining a product display area identifier, and user historical behavior data and user attribute data corresponding to the user identifier according to the product acquisition request;

Analyze the user historical behavior data and the user attribute data to obtain user characteristic data corresponding to the user identifier; and

Query the product feature data corresponding to the pre-configured candidate product identifier according to the product display area identifier.
The computer device according to claim 9, wherein the processor further executes the following steps when executing the computer-readable instruction:

Using the user feature data and the product feature data as input features, input the trained product prediction model to make predictions, and obtain the user matching degree corresponding to each of the candidate product identifiers; and

From the candidate product identifiers, the candidate product identifiers whose user matching degree meets the preset screening conditions are selected as the pushed product identifiers.
The computer device according to claim 11, wherein the processor further executes the following steps when executing the computer-readable instruction:

Query product data and product configuration data corresponding to the product identifier;

Generate corresponding product display strategy parameters according to the product configuration data;

Push the product data to the terminal for display, and obtain display statistical parameters of the product data in real time;

When the display statistical parameters are consistent with the product display strategy parameters, filter out product identifiers to be pushed again from the candidate product identifiers according to the user matching degree; and

Push the product data corresponding to the product identifier pushed again to the terminal for display.
The computer device according to claim 11, wherein the processor further executes the following steps when executing the computer-readable instruction:

Prioritize each candidate product identifier according to the user matching degree corresponding to each candidate product identifier to obtain a product identifier sequence; and

A plurality of candidate product identifiers that meet the preset screening conditions are selected from the product identifier sequence as the pushed product identifiers.
The computer device according to claim 13, wherein the processor further executes the following steps when executing the computer-readable instruction:

Respectively query the product data and product configuration data corresponding to each of the product identifiers;

Generating product display strategy parameters corresponding to the product data according to each of the product configuration data; and

Push the product data and the corresponding product display strategy parameter to the terminal to instruct the terminal to display the product data in turn according to the product display strategy parameter.
The computer device according to any one of claims 9 to 14, wherein the processor further executes the following steps when executing the computer-readable instruction:

Acquiring target user characteristic data corresponding to the target user identifier and target product characteristic data corresponding to the target product identifier;

According to the target user characteristic data and the target product characteristic data, obtain the target user matching degree corresponding to each target product identifier; and

Use the target user feature data and the target product feature data as input features, and use the corresponding target user matching degree as the desired output feature, and perform model training on the initialized product prediction model to obtain a trained product prediction model .
One or more non-volatile computer-readable storage media storing computer-readable instructions, which when executed by one or more processors, cause the one or more processors to perform the following steps:

Receive the product acquisition request corresponding to the user ID sent by the terminal;

Query the product feature data corresponding to the preconfigured candidate product identifier and the user feature data corresponding to the user identifier according to the product acquisition request;

Input the user characteristic data and the product characteristic data into a trained product prediction model, and determine the pushed product identifier through the product prediction model; and

Push the product data corresponding to the product identifier to the terminal for display.
The storage medium according to claim 16, wherein the following steps are further executed when the computer-readable instructions are executed by the processor:

Determining a product display area identifier, and user historical behavior data and user attribute data corresponding to the user identifier according to the product acquisition request;

Analyze the user historical behavior data and the user attribute data to obtain user characteristic data corresponding to the user identifier; and

Query the product feature data corresponding to the pre-configured candidate product identifier according to the product display area identifier.
The storage medium according to claim 16, wherein the following steps are further executed when the computer-readable instructions are executed by the processor:

Using the user feature data and the product feature data as input features, input the trained product prediction model to make predictions, and obtain the user matching degree corresponding to each of the candidate product identifiers; and

From the candidate product identifiers, the candidate product identifiers whose user matching degree meets the preset screening conditions are selected as the pushed product identifiers.
18. The storage medium of claim 18, wherein the following steps are further performed when the computer-readable instructions are executed by the processor:

Query product data and product configuration data corresponding to the product identifier;

Generate corresponding product display strategy parameters according to the product configuration data;

Pushing the product data to the terminal for display, and obtaining display statistical parameters of the product data in real time;

When the display statistical parameters are consistent with the product display strategy parameters, filter out product identifiers to be pushed again from the candidate product identifiers according to the user matching degree; and

Push the product data corresponding to the product identifier pushed again to the terminal for display.
The storage medium according to any one of claims 16 to 19, wherein the following steps are further executed when the computer-readable instructions are executed by the processor:

Acquiring target user characteristic data corresponding to the target user identifier and target product characteristic data corresponding to the target product identifier;

According to the target user characteristic data and the target product characteristic data, obtain the target user matching degree corresponding to each target product identifier; and

Use the target user feature data and the target product feature data as input features, and use the corresponding target user matching degree as the desired output feature, and perform model training on the initialized product prediction model to obtain a trained product prediction model .