WO2021218634A1

WO2021218634A1 - Content pushing

Info

Publication number: WO2021218634A1
Application number: PCT/CN2021/087230
Authority: WO
Inventors: 张钦; 王延夺; 杨一帆
Original assignee: 北京三快在线科技有限公司
Priority date: 2020-04-29
Filing date: 2021-04-14
Publication date: 2021-11-04
Also published as: CN111563201A

Abstract

Disclosed is a form of content pushing, comprising: acquiring a candidate push list on the basis of a search request of a terminal, the candidate push list comprising at least one item of candidate push content; acquiring a historical behavior list of a user, the historical behavior list comprising negative influence content; determining, on the basis of each item of negative influence content in the historical behavior list and each item of candidate push content in the candidate push list, the click rate for each item of candidate push content in the candidate push list; determining push content according to the click rate for each item of candidate push content in the candidate push list; and sending the push content to the terminal.

Description

Content push

This application claims the priority of the Chinese patent application with the application number 202010354131.5 and the application name "content pushing method, device, server and storage medium" filed on April 29, 2020, the entire content of which is incorporated into this application by reference .

Technical field

The embodiments of the present application relate to the field of Internet technology, and in particular to a content push.

Background technique

With the continuous development of network technology, various applications can be installed and run in terminals such as smart phones, tablet computers, and other portable devices. These applications can improve people's work, life, and entertainment.

Summary of the invention

The embodiment of the application provides a content push, and the technical solution is as follows:

On the one hand, an embodiment of the present application provides a content pushing method, which includes:

Obtaining a push candidate list based on the search request of the terminal, where the push candidate list includes at least one push content candidate;

Obtain a user's historical behavior list, the historical behavior list including negative influence content;

Determine the click-through rate of each candidate push content in the candidate push list based on each negative influence content in the historical behavior list and each candidate push content in the candidate push list;

Determine the push content according to the click-through rate of each candidate push content in the candidate push list;

Send the push content to the terminal.

In a possible implementation manner, the determination of the click rate of each candidate push content in the candidate push list based on each negative influence content in the historical behavior list and each candidate push content in the candidate push list includes:

Determine the vector corresponding to each negative influence content in the historical behavior list to obtain at least one first vector; determine the vector corresponding to each candidate push content in the candidate push list to obtain at least one second vector; based on the at least one first vector The vector and the at least one second vector determine the click rate of each candidate push content in the candidate push list.

In a possible implementation manner, the determining the click-through rate of each candidate push content in the candidate push list based on the at least one first vector and the at least one second vector includes:

Based on the at least one first vector and the at least one second vector, the negative interest weight corresponding to each candidate push content in the candidate push list is obtained; based on the negative interest weight corresponding to each candidate push content in the candidate push list And the at least one first vector to determine the click rate of each candidate push content in the candidate push list.

In a possible implementation manner, the click rate of each candidate push content in the candidate push list is determined based on the negative interest weight corresponding to each candidate push content in the candidate push list and the at least one first vector, include:

Calculate the weighted average between the negative interest weight corresponding to each candidate push content in the candidate push list and the at least one first vector, and use the weighted average as the negative interest corresponding to each candidate push content in the candidate push list. Interest vector; according to the negative interest vector corresponding to each candidate push content in the candidate push list, the click-through rate of the corresponding candidate push content is determined.

In a possible implementation manner, determining the push content according to the click-through rate of each candidate push content in the candidate push list includes:

Sort the candidate push content in the candidate push list according to the click rate of each candidate push content in the candidate push list;

According to the sorting result, a reference number of candidate push content is selected from the candidate push list as the push content.

In a possible implementation manner, the negative impact content includes at least one of short-term stay content and first-place exposure unclicked content;

The short stay content is content whose browsing time is less than the target browsing time;

The first exposed unclicked content is the first unclicked content in the historical push content.

In one possible implementation manner, the negative impact content includes at least one of short-term stay content and unclicked content with the first exposure;

The first exposed unclicked content includes the content that is displayed first in the page in the historical push content and is not clicked when the user has swiped the page.

In another aspect, a content pushing device is provided, which includes:

The first obtaining module is configured to obtain a push candidate list based on the search request of the terminal, and the push candidate list includes at least one push candidate content;

The second acquisition module is used to acquire a user's historical behavior list, and the historical behavior list includes negative influence content;

The first determining module is configured to determine the click rate of each candidate push content in the candidate push list based on each negative influence content in the historical behavior list and each candidate push content in the candidate push list;

The second determining module is configured to determine the push content according to the click rate of each candidate push content in the candidate push list;

The sending module is used to send the push content to the terminal.

In a possible implementation manner, the first determining module is configured to determine the vector corresponding to each negative influence content in the historical behavior list to obtain at least one first vector; determine each candidate push content in the candidate push list Corresponding vectors, at least one second vector is obtained; based on the at least one first vector and the at least one second vector, the click rate of each candidate push content in the candidate push list is determined.

In a possible implementation manner, the first determining module is configured to obtain the negative interest weight corresponding to each candidate push content in the candidate push list based on the at least one first vector and the at least one second vector; Based on the negative interest weight corresponding to each candidate push content in the candidate push list and the at least one first vector, the click rate of each candidate push content in the candidate push list is determined.

In a possible implementation manner, the first determining module is configured to calculate a weighted average value between the negative interest weight corresponding to each candidate push content in the candidate push list and the at least one first vector, and The weighted average is used as the negative interest vector corresponding to each candidate push content in the candidate push list; the click rate of the corresponding candidate push content is determined according to the negative interest vector corresponding to each candidate push content in the candidate push list.

In a possible implementation, the second determining module is configured to sort the candidate push content in the candidate push list according to the click-through rate of each candidate push content in the candidate push list; Select a reference number of candidate push content from the candidate push list as the push content.

In a possible implementation manner, the negatively affected content includes at least one of short-term stay content and first-place exposure unclicked content; the short-term stay content is content whose user browsing time is less than the target browsing time; and the first-place exposure unclicked content The clicked content is the first in the historical push content and the content that has not been clicked.

In a possible implementation manner, the negatively affected content includes at least one of short-term stay content and first-place exposure unclicked content; the short-term stay content is content whose user browsing time is less than the target browsing time; and the first-place exposure unclicked content The clicked content includes the content that is displayed in the first position of the page in the historical push content and is not clicked when the user has swiped the page.

On the other hand, a server is provided, including:

One or more processors;

One or more memories for storing the one or more processor-executable instructions;

Wherein, the one or more processors are used for the instruction to implement the content pushing method provided by the foregoing first aspect or any possible implementation manner of the foregoing first aspect.

In another aspect, a computer-readable storage medium is provided. When instructions in the computer-readable storage medium are executed by a processor of a server, the server can execute the first aspect or any one of the first aspects mentioned above. It is possible to implement the content push method provided by the method.

In another aspect, a computer program product is provided, including: the computer program product stores at least one instruction, and the instruction is loaded and executed by a processor to implement the operations performed by the content pushing.

The technical solutions provided by the embodiments of the present application at least bring the following beneficial effects:

In the embodiment provided in this application, according to each negative influence content in the historical behavior list and each candidate push content in the candidate push content list, the click rate of each candidate push content in the candidate push list is determined, so that each candidate push The determination of the click-through rate of the content is more accurate, so that the accuracy of the recommendation of the pushed content can be increased.

It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the embodiments of the present application.

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 that need to be used in the description of 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 a schematic diagram of an implementation environment of a content pushing method provided by an embodiment of the present application;

FIG. 2 is a flowchart of a content pushing method provided by an embodiment of the present application;

Fig. 3 is a schematic diagram of a DIN model provided by an embodiment of the present application;

FIG. 4 is a flowchart of a content pushing method provided by an embodiment of the present application;

FIG. 5 is a schematic structural diagram of a content pushing device provided by an embodiment of the present application;

FIG. 6 is a schematic structural diagram of a server provided by an embodiment of the present application;

FIG. 7 is a schematic structural diagram of a terminal provided by an embodiment of the present application.

Detailed ways

In order to make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the following further describes the embodiments of the present application in detail with reference to the accompanying drawings.

In related technologies, there can be many specific services in the application. For example, there can be a service to push recommended content. The server obtains the user's historical click record, and according to the historical click record, determines the type of content that the user likes, and searches for content related to the type. The relevant content is used as recommended content and pushed to the display interface of the terminal for users to browse.

However, when the user browses the content, he may click on a certain content due to hand error or other reasons. The content may be content that the user does not like, and the server considers the content to be the content that the user likes. The content is similar to the content, which leads to the low matching degree of the pushed content with the user's preferences, which reduces the accuracy of the recommendation to a certain extent.

FIG. 1 is a schematic diagram of an implementation environment provided by an embodiment of the application. Referring to FIG. 1, the implementation environment includes: a terminal 101 and a server 102.

The terminal 101 is connected to the server 102 through a wireless network or a wired network. The aforementioned terminal 101 may be at least one of a smart phone, a game console, a desktop computer, a tablet computer, an MP4 (Moving Picture Experts Group Audio Layer IV) player, and a laptop portable computer. The terminal 101 can install and run an application client that supports content browsing. The client can be any of a social application client and a casual shopping client. Of course, the client can also be other types of customers. On the other hand, the embodiments of this application do not limit the type of the client. The terminal 101 may generate a search request in response to the user's search operation, and the terminal 101 may also send the search request to the server 102.

The server 102 may be one server or a server cluster composed of multiple servers. The server 102 may also be at least one of a cloud computing platform and a virtualization center, which is not limited in the embodiment of the present application. The server 102 may obtain the candidate push list based on the search request of the terminal. The server 102 may also obtain a list of historical behaviors of the user. The server 102 may determine the click rate of each candidate push content in the candidate push list based on the content in the historical behavior list and the content in the candidate push list. The server 102 may also determine the push content according to the click rate of the candidate push content. The server 102 may also send the pushed content to the terminal 101 for display by the terminal 101. Of course, the server 102 may also include other functional servers to provide more comprehensive and diversified services.

The terminal 101 may generally refer to one of multiple terminals, and this embodiment only uses the terminal 101 as an example for illustration. Those skilled in the art may know that the number of the aforementioned terminals 101 may be more or less. For example, there may be only a few terminals 101 mentioned above, or there may be dozens or hundreds of terminals 101 mentioned above, or a larger number. The embodiment of the present application does not limit the number of terminals 101 and the device type.

Based on the foregoing implementation environment, an embodiment of the present application provides a content pushing method. Taking the flowchart of the content pushing method provided in the embodiment of the present application shown in FIG. 2 as an example, the method can be executed by the server 102 in FIG. 1 . As shown in Figure 2, the method includes the following steps:

In step 201, based on the search request of the terminal, a candidate push list is obtained, and the candidate push list includes at least one candidate push content.

In the embodiment of the present application, an application client that supports content browsing is installed and running in the terminal, and the user can input the name of the content to be queried in the client and click the search button. The terminal generates a search request in response to the user's search operation. The search request can carry the user identification of the user. The user identification can be the user's account information or other user information, as long as it can be used to identify the user. The embodiment of the present application does not limit the user identification. The terminal sends the search request to the server. The server obtains a push candidate list corresponding to the search request based on the search request of the terminal, and the push candidate list includes at least one push content candidate.

For example, the client application is a takeaway application. The user enters "ramen" in the client and clicks the search button. The terminal generates a search request in response to the user's click operation. The search request carries the user ID of the user. The user ID is the user's account. The terminal sends the search request to the server, and the server obtains a candidate push list based on the search request, and the candidate push list includes at least one noodle restaurant.

In step 202, a user's historical behavior list is obtained, and the historical behavior list includes negative influence content.

In an exemplary embodiment, the negative influence content includes at least one of short-term stay content and first-place exposure unclicked content. Short stay content is content whose browsing time is less than the target browsing time. The first exposed unclicked content is the first unclicked content in the historical push content. Among them, the determination of the first unclicked content to be exposed requires the user to slide down the terminal page. For example, if the first content is ranked first in the historical push content, the user does not click the first content, and the user does not slide down the page, the first content that is ranked first cannot be the first to be exposed and not clicked content. The second content ranks first in the historical push content, and the user does not click the second content, but the user slides down the page, then the first ranked second content can be used as the first to expose the unclicked content. That is, in an exemplary embodiment, the first exposed unclicked content includes content that is displayed first on the page in the historical push content and is not clicked when the user has swiped the page.

It should be noted that the above-mentioned target browsing time can be set based on experience, can also be adjusted according to application scenarios, or can be manually set by the user, and this embodiment of the application does not limit the value of the target browsing time.

In a possible implementation manner, the server may allocate a first storage space for each user. The first storage space is used to store the user identification and historical behavior list of the user. The historical behavior list includes the corresponding Negatively affect content. When the server receives the search request sent by the terminal, it parses the search request to obtain the user ID of the user. A search is performed in the storage space based on the user identification, so that the first storage space corresponding to the user identification can be obtained. Extracting the historical behavior list corresponding to the user identifier in the first storage space means acquiring the historical behavior list of the user, so that the negative influence content corresponding to the user can be acquired.

In a possible implementation, the server's determination of the negative impact content can be as follows: the server can count the user’s negative interest characteristics based on the RFM (Recency, Frequency, Monetary, last transaction, transaction frequency, transaction amount) model, In order to determine the negative impact content. Among them, R represents the time interval from the user's last transaction to the current time. The larger the R, the longer the user has not had a transaction. F represents the number of transactions of the user in the most recent period of time. The larger the F, the more frequent the transactions of the user. M represents the user's transaction amount each time, which can be the most recent transaction amount or the past average transaction amount. The RFM model can better record the user's recent transaction records. The time length of the recent period of time can be set according to the transaction type, or it can be manually set by the user. The embodiment of the present application does not limit the time length of the recent period of time. The server can also count the user's historical behavior types, historical behavior prices, and the time interval from the historical behavior to the current time based on the RFM model offline, and calculate the RFM value of the historically pushed content. Based on the RFM value of the historically pushed content, the user's long-term stable negative interest point can be more reliably reflected, and the content corresponding to the negative interest point is also negatively influencing content. Of course, the foregoing method of determining the negative influence content based on the RFM model is only an example, and this embodiment may also adopt other methods to determine the negative influence content.

In step 203, based on each negative influence content in the historical behavior list and each candidate push content in the candidate push list, the click rate of each candidate push content in the candidate push list is determined.

In the embodiment of the present application, based on each negative influence content in the historical behavior list and each candidate push content in the candidate push list, determining the click rate of each candidate push content in the candidate push list may include the following steps:

Step 2031: Determine the vector corresponding to each negative influence content in the historical behavior list, and obtain at least one first vector.

In a possible implementation manner, the vector corresponding to the negative influence content is stored in the server. Based on the user's historical behavior list, the server obtains a vector corresponding to each negative influence content in the historical behavior list from its storage space, and obtains at least one first vector.

For example, if the user's historical behavior list includes 2 negative influence contents, the first vectors corresponding to the 2 negative influence contents may be extracted to obtain 2 first vectors.

Step 2032: Determine the vector corresponding to each candidate push content in the candidate push list, and obtain at least one second vector;

In a possible implementation manner, the vector corresponding to each candidate push content is stored in the server. Based on the candidate push list of the user, the server obtains a vector corresponding to each candidate push content in the candidate push list from its storage space, and obtains at least one second vector.

For example, if the user's candidate push list includes 5 candidate push contents, the second vectors corresponding to the 5 candidate push contents may be extracted to obtain 5 second vectors.

Step 2033: Based on at least one first vector and at least one second vector, determine the click rate of each candidate push content in the candidate push list.

In a possible implementation manner, based on at least one first vector and at least one second vector, determining the click-through rate of each candidate push content in the candidate push list may include the following steps:

Step 1. Based on at least one first vector and at least one second vector, obtain a negative interest weight corresponding to each candidate push content in the candidate push list;

In a possible implementation manner, based on the at least one first vector obtained in step 2031 and the at least one second vector obtained in step 2032, the negative interest weight corresponding to each candidate push content in the candidate push list is calculated .

Exemplarily, the historical behavior list includes two negative influence contents, namely the first negative influence content and the second negative influence content. Determine the first first vector corresponding to the first negative influence content, and determine the second first vector corresponding to the second negative influence content. The candidate push list includes 5 candidate push content, which are the first candidate push content, the second candidate push content, the third candidate push content, the fourth candidate push content, and the fifth candidate push content. Determine the first second vector corresponding to the first candidate push content, determine the second second vector corresponding to the second candidate push content, determine the third second vector corresponding to the third candidate push content, and determine the fourth candidate push The fourth second vector corresponding to the content determines the fifth second vector corresponding to the fifth candidate push content.

Take the calculation process of the negative interest weight corresponding to the first candidate pushing content in the candidate pushing list as an example for description. Calculate the inner product between the first second vector corresponding to the first candidate push content and the first first vector corresponding to the first negative influence content in the historical behavior list to obtain the first negative interest weight. Calculate the inner product between the first second vector corresponding to the first candidate push content and the second first vector corresponding to the second negative influence content in the historical behavior list to obtain the second negative interest weight. The two negative interest weights are both the negative interest weights corresponding to the first candidate push content.

In a possible implementation manner, the server may also construct the first sequence according to the user's point of interest, location, user's historical behavior, and other dimensions. The constructed first sequence includes the user's points of interest, the region and the user's historical behavior list. A second sequence is constructed according to the candidate push list, and the constructed second sequence includes the candidate push content searched according to the user's point of interest, region, and category of historical behavior. The elements in the first sequence and the second sequence can be vectorized based on the deep network, and the negative interest weight of each candidate push content can be obtained based on the candidate push content in the second sequence and the content in the first sequence.

It should be noted that the calculation process of the negative interest weight of each candidate push content in the candidate push list is consistent with the calculation process of the negative interest weight of the first candidate push content, and will not be repeated here.

Step 2: Determine the click rate of each candidate push content in the candidate push list based on the negative interest weight corresponding to each candidate push content in the candidate push list and at least one first vector.

In a possible implementation manner, based on the negative interest weight corresponding to each candidate push content in the candidate push list and at least one first vector, the process of determining the click rate of each candidate push content in the candidate push list includes the following step:

The first step is to calculate the weighted average between the negative interest weight corresponding to each candidate push content in the candidate push list and at least one first vector, and use the weighted average as the negative interest corresponding to each candidate push content in the candidate push list. To the interest vector.

In a possible implementation manner, there may be any one of the following implementation manners to determine the negative interest vector corresponding to each candidate push content in the candidate push list:

Implementation method 1: Based on the negative interest weight of each candidate push content in the candidate push list obtained in step 1, calculate the negative interest weight corresponding to each candidate push content and the at least one first vector obtained in step 2031. The weighted average of the time, the weighted average is used as the negative interest vector corresponding to the candidate push content.

Take the calculation process of the negative interest vector corresponding to the first candidate pushing content in the candidate pushing list as an example for description. The first candidate push content corresponds to a first negative interest weight and a second negative interest weight. Based on the first negative interest weight, the second negative interest weight, the first first vector and the second first vector, the negative interest vector corresponding to the first candidate push content is obtained.

For example, multiply the first negative interest weight and the first first vector to obtain a vector, and multiply the second negative interest weight and the second first vector to obtain a vector. The two obtained vectors are correspondingly added to serve as the negative interest vector corresponding to the first candidate push content.

It should be noted that the calculation process of the negative interest vector of each candidate push content in the candidate push list is consistent with the calculation process of the negative interest vector of the first candidate push content described above, and will not be repeated here.

Implementation manner 2: The server may also use the DIN (Deep Interest Network) model to calculate the negative interest vector of each candidate push content in the candidate push list. The following Figure 3 shows a schematic diagram of a DIN model shown in an embodiment of the application. In FIG. 3, the negative interest vector of the candidate push content is calculated according to the vector of the candidate push content and the vector of the negative influence content in the historical behavior list. The DIN model can truly reveal the user's desire to click on the candidate push content in the candidate push list based on each negative content in the user's historical behavior list. The behavioral data in the DIN model includes two structures: content diversity and local activation. The content diversity of behavioral data reflects the different points of interest of users. Users often click on a certain content because the content touches part of the user's points of interest. The Attention Unit (activation unit) is added to the input layer of the DIN model. Attention Unit uses the Attention mechanism to achieve negative impact on each candidate push content in the candidate push list based on the negative influence content in the historical behavior list. To determine the interest vector. The Attention Unit includes FCs and Concat, where FCs are fieldbus control, and Concat is used to connect the negative influence content in the historical behavior list and the candidate push content in the candidate push list.

In a possible implementation, the negative interest weight of each candidate push content in the candidate push list and each negative content vector in the historical behavior list are calculated to calculate the negative direction of each candidate push content in the candidate push list. Interest vector, that is, through the weighted pooling layer, based on the negative interest weight of each candidate push content in the candidate push list and each negative interest vector in the historical behavior list to obtain the negative interest vector of each candidate push content . Among them, the expression formula of the negative interest vector of each candidate push content in the candidate push list obtained through the weighted pooling layer is as follows:

In the above formula, N is the number of negative effects content, V _i represents the historical behavior list the negative effects of contents of the vector of i, V _a candidate push list of a candidate push content negative interest vector, g V _i, V _a ) represents the inner product of the vector of negatively affecting content i and the candidate push content in the historical behavior list, that is, the negative interest weight of the candidate push content, which also corresponds to W _i in the above formula.

It should be noted that the server can select any of the foregoing implementation manners to determine the negative interest vector of each candidate push content in the candidate push list, which is not limited in the embodiment of the present application.

The second step is to determine the click rate of the corresponding candidate push content according to the negative interest vector corresponding to each candidate push content in the candidate push list.

In a possible implementation manner, the server calculates the click rate of the candidate push content based on the negative interest vector of the candidate push content obtained in the first step.

For example, the negative interest vector of the first candidate push content is (2, 1, 2), calculate the modulus length L of the negative interest vector,

The modulus length of the negative interest vector is taken as the click rate of the candidate push content, that is, the click rate of the first candidate push content is 3.

In a possible implementation manner, in order to make the calculation of the click-through rate of the candidate push content more accurate, the click-through rate of the candidate push content may also be calculated according to the negative interest vector and the positive interest vector of the candidate push content. Wherein, the calculation process of the negative interest vector of the candidate push content is as the process of the first step described above. The calculation process of the forward interest vector of the candidate push content is as follows:

In a possible implementation manner, the user's historical behavior list may also include positive influence content. The positive influence content may be the user's historical click content, the vector corresponding to the positive influence content is determined, and at least one third vector is obtained. Based on the at least one third vector and the at least one second vector, a forward interest weight corresponding to each candidate pushing content in the candidate pushing list is obtained. Calculate the weighted average between the forward interest weight corresponding to each candidate push content in the candidate push list and at least one third vector, and use the weighted average as the forward interest vector corresponding to each candidate push content in the candidate push list .

Wherein, based on the at least one third vector and at least one second vector, the process of obtaining the positive interest weight corresponding to each candidate push content in the candidate push list is the same as the negative interest weight corresponding to the candidate push content in step 1 above. The calculation process is the same, so I won't repeat it here. Calculate the weighted average between the forward interest weight corresponding to each candidate push content in the candidate push list and at least one third vector, and use the weighted average as the forward interest vector corresponding to each candidate push content in the candidate push list The process of is consistent with the calculation process of the negative interest vector of the candidate push content in the first step above, and will not be repeated here.

In a possible implementation manner, a weighted average calculation is performed on the negative interest vector and the positive interest vector of the candidate push content to obtain the interest vector of the candidate push content, and the candidate push is determined based on the interest vector of the candidate push content The click-through rate of the content.

For example, the negative interest vector of the first candidate push content is (2, 1, 2), and the positive interest vector is (1, 2, 1). Based on the negative interest vector and the positive interest vector, the first candidate is obtained Push the interest vector (2, 2, 2) of the content, calculate the modulus L of the interest vector,

That is, the click-through rate of the first candidate push content

It should be noted that if the click-through rate of the first candidate push content in the candidate push list is calculated based on the negative interest vector, then the click-through rate calculation of the other candidate push content in the candidate push list is also based on the negative The interest vector is calculated. If the click-through rate of the first candidate push content in the candidate push list is calculated based on the negative interest vector and the positive interest vector, then the click-through rate calculation of the other candidate push content in the candidate push list is also based on the negative interest The vector and the forward interest vector are calculated.

In step 204, the push content is determined according to the click rate of each candidate push content in the candidate push list.

In the embodiment of the present application, since the number of push contents displayed on the interface of the client installed and running in the terminal is limited, it is necessary to determine the target number of push contents from the candidate push contents. The method for determining the content to be pushed can be any of the following:

Implementation method 1: Determine the push content based on the click-through rate of each candidate push content in the candidate push list.

Sort the candidate push content in the candidate push list according to the click rate of each candidate push content in the candidate push list; according to the sorting result, select a reference number of candidate push content from the candidate push list as the push content.

For example, there are 5 candidate push content in the candidate push list, the first candidate push content has a click rate of 0.83, the second candidate push content has a click rate of 0.95, the third candidate push content has a click rate of 0.80, and the fourth candidate push has a click rate of 0.80. The click-through rate of the content is 0.90, and the click-through rate of the fifth candidate push content is 0.88, sorted according to the click-through rate of the candidate push content, and the ranking results obtained are the second candidate push content, the fourth candidate push content, and the fifth candidate push content. , The first candidate push content, and the third candidate push content. In the ranking result, the top three push content candidates are determined as push content, that is, the second push content candidate, the fourth push content candidate, and the fifth push content candidate are determined to be push content.

It should be noted that the reference quantity can be set based on experience, can also be adjusted according to different clients, or can be manually set by the user, and the embodiment of the application does not limit the value of the reference quantity.

Implementation method 2: Determine the candidate push content whose click rate exceeds the target click rate in the candidate push list as the push content.

The server can set a target click rate, filter the candidate push content with a click rate higher than the target click rate in the candidate push list, and determine the filtered candidate push content as the push content.

For example, there are 5 candidate push content in the candidate push list, the first candidate push content has a click rate of 0.83, the second candidate push content has a click rate of 0.95, the third candidate push content has a click rate of 0.80, and the fourth candidate push has a click rate of 0.80. The click-through rate of the content is 0.90, and the click-through rate of the fifth candidate push content is 0.88. The target click rate is 0.85. The candidate push content whose click rate is higher than the target click rate is determined as the push content, that is, the second candidate push content, the fourth candidate push content, and the fifth candidate push content are determined as the push content.

It should be noted that the target click-through rate can be set based on experience, or adjusted according to different clients, or manually set by the user. The embodiment of the application does not limit the value of the target click-through rate.

In step 205, the push content is sent to the terminal.

In the embodiment of the present application, the server sends the push content to the terminal corresponding to the user identification according to the user identification of the user, and the client installed in the terminal displays the push content.

Referring to FIG. 4, FIG. 4 is a flowchart of a content pushing method provided by an embodiment of the present application. Taking the flowchart of a content pushing method provided by an embodiment of the present application shown in FIG. 4 as an example, the method can be described in FIG. 1 The interaction between the terminal 101 and the server 102 in FIG. As shown in Figure 4, the method includes the following steps:

In step 401, the terminal sends a search request to the server.

In the embodiment of this application, an application client that supports content browsing is installed and running in the terminal. The client can be a social application client or a casual shopping client. Make a limit. The user can browse the content on the client installed on the terminal device. The user can enter the name of the content that the user wants to search in the search box of the client, and click the search button. The terminal responds to the user's search operation and generates a search request. The search request carries the user ID of the user. The user ID can be The user's account information may also be other user information, as long as it can be used to identify the user, and the embodiment of the present application does not limit the user identification.

In a possible implementation manner, after the terminal obtains the search request, it can directly send the search request to the server.

In step 402, the server receives the search request sent by the terminal, and obtains a push candidate list based on the search request, and the push candidate list includes at least one push candidate content.

In this step 402, the method for obtaining the candidate push list is the same as the method in the above step 201, and will not be repeated here.

In step 403, the server obtains the user's historical behavior list, and the historical behavior list includes negative influence content.

In this step 403, the method of obtaining the user's historical behavior list is the same as the method in the above step 202, and will not be repeated here.

In step 404, the server determines the click rate of each candidate push content in the candidate push list based on each negative influence content in the historical behavior list and each candidate push content in the candidate push list.

In this step 404, the method for determining the click rate of each candidate push content in the candidate push list is the same as the method in the above step 203, and will not be repeated here.

In step 405, the server determines the push content according to the click rate of each candidate push content in the candidate push list.

In this step 405, the method for determining the content to be pushed is the same as the method in the above step 204, and will not be repeated here.

In step 406, the server sends the pushed content to the terminal.

In this step 406, the process of the server sending the pushed content to the terminal is the same as the process in the foregoing step 205, and will not be repeated here.

In step 407, the terminal receives the push content sent by the server, and displays the push content.

In this step 407, after the terminal receives the push content sent by the server, the push content is displayed on the interface of the client installed on the terminal, so that the user can browse and view the push content.

FIG. 5 is a schematic structural diagram of a content pushing device provided by an embodiment of the application. As shown in FIG. 5, the device includes:

The first obtaining module 501 is configured to obtain a push candidate list based on a search request of the terminal, and the push candidate list includes at least one push candidate content;

The second obtaining module 502 is configured to obtain a user's historical behavior list, and the historical behavior list includes negative influence content;

The first determining module 503 is configured to determine the click-through rate of each candidate push content in the candidate push list based on each negative influence content in the historical behavior list and each candidate push content in the candidate push list;

The second determining module 504 is configured to determine the push content according to the click rate of each candidate push content in the candidate push list;

The sending module 505 is configured to send the pushed content to the terminal.

In a possible implementation manner, the first determining module 503 is configured to determine the vector corresponding to each negative influence content in the historical behavior list to obtain at least one first vector; determine each candidate push in the candidate push list The vector corresponding to the content obtains at least one second vector; based on the at least one first vector and the at least one second vector, the click rate of each candidate push content in the candidate push list is determined.

In a possible implementation manner, the first determining module 503 is configured to obtain the negative interest weight corresponding to each candidate push content in the candidate push list based on the at least one first vector and the at least one second vector ; Based on the negative interest weight corresponding to each candidate push content in the candidate push list and the at least one first vector, the click rate of each candidate push content in the candidate push list is determined.

In a possible implementation manner, the first determining module 503 is configured to calculate the weighted average value between the negative interest weight corresponding to each candidate push content in the candidate push list and the at least one first vector, and The weighted average is used as the negative interest vector corresponding to each candidate push content in the candidate push list; the click rate of the corresponding candidate push content is determined according to the negative interest vector corresponding to each candidate push content in the candidate push list.

In a possible implementation, the second determining module 504 is configured to sort the candidate push content in the candidate push list according to the click-through rate of each candidate push content in the candidate push list; A reference number of candidate push content is selected from the candidate push list as the push content.

The above device determines the click rate of each candidate push content in the candidate push list according to each negatively affected content in the historical behavior list and each candidate push content in the candidate push content list, so that the click rate of each candidate push content is more determined. Accurate, which can increase the recommendation accuracy of the pushed content.

It should be noted that when the content pushing device provided in the above embodiment pushes content, only the division of the above functional modules is used as an example for illustration. In actual applications, the above functions can be allocated by different functional modules as needed. That is, the internal structure of the content pushing device is divided into different functional modules to complete all or part of the functions described above. In addition, the content pushing device provided in the foregoing embodiment and the content pushing method embodiment belong to the same concept, and the implementation process is detailed in the method embodiment, which will not be repeated here.

6 is a schematic structural diagram of a server 600 provided by an embodiment of the present application. The server 600 may have relatively large differences due to different configurations or performance, and may include one or more central processing units (CPU) 601 and One or more memories 602, wherein at least one instruction is stored in the one or more memories 602, and the at least one instruction is loaded and executed by the one or more processors 601 to implement the content pushing method provided by the above method embodiment . Of course, the server 600 may also have components such as a wired or wireless network interface, a keyboard, and an input/output interface for input and output. The server 600 may also include other components for implementing device functions, which will not be repeated here.

FIG. 7 is a schematic structural diagram of a terminal provided by an embodiment of the present application. The terminal 700 may be a smart phone, a tablet computer, an MP4 (Moving Picture Experts Group Audio Layer IV, moving picture expert compression standard audio layer 4) player, a notebook computer, or a desktop computer. The terminal 700 may also be called user equipment, portable terminal, laptop terminal, desktop terminal and other names.

Generally, the terminal 700 includes: one or more processors 701 and one or more memories 702.

The processor 701 may include one or more processing cores, such as a 4-core processor, an 8-core processor, and so on. The processor 701 can adopt at least one hardware form among DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array, Programmable Logic Array). accomplish. The processor 701 may also include a main processor and a co-processor. The main processor is a processor used to process data in the awake state, also called a CPU (Central Processing Unit, central processing unit); the co-processor is A low-power processor used to process data in the standby state. In some embodiments, the processor 701 may be integrated with a GPU (Graphics Processing Unit, image processor), and the GPU is used for rendering and drawing content that needs to be displayed on the display screen. In some embodiments, the processor 701 may further include an AI (Artificial Intelligence) processor, and the AI processor is used to process computing operations related to machine learning.

The memory 702 may include one or more computer-readable storage media, which may be non-transitory. The memory 702 may also include high-speed random access memory and non-volatile memory, such as one or more magnetic disk storage devices and flash memory storage devices. In some embodiments, the non-transitory computer-readable storage medium in the memory 702 is used to store at least one instruction, and the at least one instruction is used to be executed by the processor 701 to implement the method provided in the embodiment of the present application. Content push method.

In some embodiments, the terminal 700 may optionally further include: a peripheral device interface 703 and at least one peripheral device. The processor 701, the memory 702, and the peripheral device interface 703 may be connected by a bus or a signal line. Each peripheral device can be connected to the peripheral device interface 703 through a bus, a signal line, or a circuit board. Exemplarily, the peripheral device includes: at least one of a radio frequency circuit 704, a display screen 705, a camera component 706, an audio circuit 707, a positioning component 708, and a power supply 709.

The peripheral device interface 703 may be used to connect at least one peripheral device related to I/O (Input/Output) to the processor 701 and the memory 702. In some embodiments, the processor 701, the memory 702, and the peripheral device interface 703 are integrated on the same chip or circuit board; in some other embodiments, any one of the processor 701, the memory 702, and the peripheral device interface 703 or The two can be implemented on a separate chip or circuit board, which is not limited in this embodiment.

The radio frequency circuit 704 is used for receiving and transmitting RF (Radio Frequency, radio frequency) signals, also called electromagnetic signals. The radio frequency circuit 704 communicates with a communication network and other communication devices through electromagnetic signals. The radio frequency circuit 704 converts electrical signals into electromagnetic signals for transmission, or converts received electromagnetic signals into electrical signals. Optionally, the radio frequency circuit 704 includes: an antenna system, an RF transceiver, one or more amplifiers, a tuner, an oscillator, a digital signal processor, a codec chipset, a user identity module card, and so on. The radio frequency circuit 704 can communicate with other terminals through at least one wireless communication protocol. The wireless communication protocol includes, but is not limited to: metropolitan area networks, various generations of mobile communication networks (2G, 3G, 4G, and 5G), wireless local area networks, and/or WiFi (Wireless Fidelity, wireless fidelity) networks. In some embodiments, the radio frequency circuit 704 may also include a circuit related to NFC (Near Field Communication), which is not limited in the embodiment of the present application.

The display screen 705 is used to display a UI (User Interface, user interface). The UI can include graphics, text, icons, videos, and any combination thereof. When the display screen 705 is a touch display screen, the display screen 705 also has the ability to collect touch signals on or above the surface of the display screen 705. The touch signal can be input to the processor 701 as a control signal for processing. At this time, the display screen 705 may also be used to provide virtual buttons and/or virtual keyboards, also called soft buttons and/or soft keyboards. In some embodiments, there may be one display screen 705, which is provided with the front panel of the terminal 700; in other embodiments, there may be at least two display screens 705, which are respectively arranged on different surfaces of the terminal 700 or in a folding design; In some embodiments, the display screen 705 may be a flexible display screen, which is arranged on a curved surface or a folding surface of the terminal 700. Furthermore, the display screen 705 can also be set as a non-rectangular irregular pattern, that is, a special-shaped screen. The display screen 705 may be made of materials such as LCD (Liquid Crystal Display), OLED (Organic Light-Emitting Diode, organic light emitting diode).

The camera assembly 706 is used to capture images or videos. Optionally, the camera assembly 706 includes a front camera and a rear camera. Generally, the front camera is set on the front panel of the terminal, and the rear camera is set on the back of the terminal. In some embodiments, there are at least two rear cameras, each of which is a main camera, a depth-of-field camera, a wide-angle camera, and a telephoto camera, so as to realize the fusion of the main camera and the depth-of-field camera to realize the background blur function, the main camera Integrate with the wide-angle camera to realize panoramic shooting and VR (Virtual Reality) shooting functions or other fusion shooting functions. In some embodiments, the camera assembly 706 may also include a flash. The flash can be a single-color flash or a dual-color flash. Dual color temperature flash refers to a combination of warm light flash and cold light flash, which can be used for light compensation under different color temperatures.

The audio circuit 707 may include a microphone and a speaker. The microphone is used to collect sound waves from the user and the environment, convert the sound waves into electrical signals and input them to the processor 701 for processing, or input to the radio frequency circuit 704 to implement voice communication. For the purpose of stereo collection or noise reduction, there may be multiple microphones, which are respectively set in different parts of the terminal 700. The microphone can also be an array microphone or an omnidirectional collection microphone. The speaker is used to convert the electrical signal from the processor 701 or the radio frequency circuit 704 into sound waves. The speaker can be a traditional thin-film speaker or a piezoelectric ceramic speaker. When the speaker is a piezoelectric ceramic speaker, not only can the electrical signal be converted into human audible sound waves, but also the electrical signal can be converted into human inaudible sound waves for purposes such as distance measurement. In some embodiments, the audio circuit 707 may also include a headphone jack.

The positioning component 708 is used to locate the current geographic location of the terminal 700 to implement navigation or LBS (Location Based Service, location-based service). The positioning component 708 may be a positioning component based on the GPS (Global Positioning System, Global Positioning System) of the United States, the Beidou system of China, the Grenas system of Russia, or the Galileo system of the European Union.

The power supply 709 is used to supply power to various components in the terminal 700. The power source 709 may be alternating current, direct current, disposable batteries, or rechargeable batteries. When the power source 709 includes a rechargeable battery, the rechargeable battery may support wired charging or wireless charging. The rechargeable battery can also be used to support fast charging technology.

In some embodiments, the terminal 700 further includes one or more sensors 170. The one or more sensors 170 include, but are not limited to: an acceleration sensor 711, a gyroscope sensor 712, a pressure sensor 711, a fingerprint sensor 714, an optical sensor 715, and a proximity sensor 716.

The acceleration sensor 711 can detect the magnitude of acceleration on the three coordinate axes of the coordinate system established by the terminal 700. For example, the acceleration sensor 711 may be used to detect the components of gravitational acceleration on three coordinate axes. The processor 701 may control the display screen 705 to display the user interface in a horizontal view or a vertical view according to the gravity acceleration signal collected by the acceleration sensor 711. The acceleration sensor 711 may also be used for the collection of game or user motion data.

The gyroscope sensor 712 can detect the body direction and the rotation angle of the terminal 700, and the gyroscope sensor 712 can cooperate with the acceleration sensor 711 to collect the user's 3D actions on the terminal 700. The processor 701 can implement the following functions according to the data collected by the gyroscope sensor 712: motion sensing (for example, changing the UI according to the user's tilt operation), image stabilization during shooting, game control, and inertial navigation.

The pressure sensor 711 may be arranged on the side frame of the terminal 700 and/or the lower layer of the display screen 705. When the pressure sensor 711 is arranged on the side frame of the terminal 700, the user's holding signal of the terminal 700 can be detected, and the processor 701 performs left and right hand recognition or quick operation according to the holding signal collected by the pressure sensor 711. When the pressure sensor 711 is arranged on the lower layer of the display screen 705, the processor 701 controls the operability controls on the UI interface according to the user's pressure operation on the display screen 705. The operability control includes at least one of a button control, a scroll bar control, an icon control, and a menu control.

The fingerprint sensor 714 is used to collect the fingerprint of the user, and the processor 701 recognizes the user's identity according to the fingerprint collected by the fingerprint sensor 714, or the fingerprint sensor 714 recognizes the user's identity according to the collected fingerprint. When it is recognized that the user's identity is a trusted identity, the processor 701 authorizes the user to perform related sensitive operations, including unlocking the screen, viewing encrypted information, downloading software, paying, and changing settings. The fingerprint sensor 714 may be provided on the front, back, or side of the terminal 700. When a physical button or a manufacturer logo is provided on the terminal 700, the fingerprint sensor 714 can be integrated with the physical button or the manufacturer logo.

The optical sensor 715 is used to collect the ambient light intensity. In an embodiment, the processor 701 may control the display brightness of the display screen 705 according to the ambient light intensity collected by the optical sensor 715. Exemplarily, when the ambient light intensity is high, the display brightness of the display screen 705 is increased; when the ambient light intensity is low, the display brightness of the display screen 705 is decreased. In another embodiment, the processor 701 may also dynamically adjust the shooting parameters of the camera assembly 706 according to the ambient light intensity collected by the optical sensor 715.

The proximity sensor 716, also called a distance sensor, is usually arranged on the front panel of the terminal 700. The proximity sensor 716 is used to collect the distance between the user and the front of the terminal 700. In one embodiment, when the proximity sensor 716 detects that the distance between the user and the front of the terminal 700 gradually decreases, the processor 701 controls the display screen 705 to switch from the on-screen state to the off-screen state; when the proximity sensor 716 detects When the distance between the user and the front of the terminal 700 gradually increases, the processor 701 controls the display screen 705 to switch from the rest screen state to the bright screen state.

Those skilled in the art can understand that the structure shown in FIG. 7 does not constitute a limitation on the terminal 700, and may include more or fewer components than shown in the figure, or combine certain components, or adopt different component arrangements.

In an exemplary embodiment, a computer-readable storage medium is also provided, the storage medium stores at least one piece of program code, and the at least one piece of program code is loaded and executed by a processor of a computer device to implement any of the foregoing Content push method.

Optionally, the foregoing computer-readable storage medium may be Read-Only Memory (ROM), Random Access Memory (RAM), Compact Disc Read-Only Memory, CD-ROM ), magnetic tapes, floppy disks and optical data storage devices.

It should be understood that the "plurality" mentioned herein refers to two or more. "And/or" describes the association relationship of the associated objects, indicating that there can be three types of relationships, for example, A and/or B, which can mean: A alone exists, A and B exist at the same time, and B exists alone. The character "/" generally indicates that the associated objects before and after are in an "or" relationship.

The above are only exemplary embodiments of the application, and are not intended to limit the embodiments of the application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the embodiments of the application shall be included in the application Within the protection scope of the embodiment.

Claims

A content pushing method, wherein the method includes:

Obtaining a push candidate list based on the search request of the terminal, where the push candidate list includes at least one push candidate content;

Acquiring a list of historical behaviors of the user, the list of historical behaviors including negative influence content;

Determine the click rate of each candidate push content in the candidate push list based on each negative influence content in the historical behavior list and each candidate push content in the candidate push list;

Determine the push content according to the click rate of each candidate push content in the candidate push list;

Sending the pushed content to the terminal.
The method according to claim 1, wherein said determining each candidate push content in said candidate push list based on each negative influence content in said historical behavior list and each candidate push content in said candidate push list The click-through rate, including:

Determine the vector corresponding to each negative influence content in the historical behavior list to obtain at least one first vector;

Determining a vector corresponding to each candidate pushing content in the candidate pushing list to obtain at least one second vector;

Based on the at least one first vector and the at least one second vector, the click rate of each candidate push content in the candidate push list is determined.
The method according to claim 2, wherein the determining the click-through rate of each candidate push content in the candidate push list based on the at least one first vector and the at least one second vector comprises:

Obtaining, based on the at least one first vector and the at least one second vector, a negative interest weight corresponding to each candidate pushing content in the candidate pushing list;

Based on the negative interest weight corresponding to each candidate push content in the candidate push list and the at least one first vector, the click rate of each candidate push content in the candidate push list is determined.
The method according to claim 3, wherein said determining each candidate in the candidate pushing list based on the negative interest weight corresponding to each candidate pushing content in the candidate pushing list and the at least one first vector The click-through rate of the pushed content, including:

Calculate the weighted average between the negative interest weight corresponding to each candidate push content in the candidate push list and the at least one first vector, and use the weighted average as each candidate push in the candidate push list The negative interest vector corresponding to the content;

The click rate of the corresponding candidate push content is determined according to the negative interest vector corresponding to each candidate push content in the candidate push list.
The method according to claim 1, wherein the determining the push content according to the click-through rate of each candidate push content in the candidate push list comprises:

Sorting the candidate push content in the candidate push list according to the click rate of each candidate push content in the candidate push list;

According to the sorting result, a reference number of candidate push content is selected from the candidate push list as the push content.
The method according to any one of claims 1 to 5, wherein the negative influence content includes at least one of short-term stay content and first-place exposure unclicked content;

The short stay content is content for which the user's browsing time is less than the target browsing time;

The first exposed unclicked content is the first unclicked content in the historical push content.
The method according to any one of claims 1 to 5, wherein the negatively influencing content includes at least one of short-term stay content and first-place exposure unclicked content;

The short stay content is content for which the user's browsing time is less than the target browsing time;

The first exposed unclicked content includes content that is displayed first in the page in the historical push content and is not clicked when the user has swiped the page.
A content pushing device, wherein the device includes:

The first obtaining module is configured to obtain a push candidate list based on a search request of the terminal, where the push candidate list includes at least one push candidate content;

The second acquisition module is configured to acquire a user's historical behavior list, where the historical behavior list includes negative influence content;

The first determining module is configured to determine the click rate of each candidate push content in the candidate push list based on each negative influence content in the historical behavior list and each candidate push content in the candidate push list;

The second determining module is configured to determine the push content according to the click rate of each candidate push content in the candidate push list;

The sending module is used to send the pushed content to the terminal.
8. The device according to claim 8, wherein the first determining module is configured to determine a vector corresponding to each negative influence content in the historical behavior list to obtain at least one first vector; and determine the candidate push list The vector corresponding to each candidate push content obtains at least one second vector; based on the at least one first vector and the at least one second vector, the click rate of each candidate push content in the candidate push list is determined.
The device according to claim 9, wherein the first determining module is configured to obtain the corresponding content of each candidate push in the candidate push list based on the at least one first vector and the at least one second vector Based on the negative interest weight corresponding to each candidate push content in the candidate push list and the at least one first vector, the click rate of each candidate push content in the candidate push list is determined.
The apparatus according to claim 10, wherein the first determining module calculates a weighted average value between the negative interest weight corresponding to each candidate push content in the candidate push list and the at least one first vector, The weighted average is used as the negative interest vector corresponding to each candidate push content in the candidate push list; the click of the corresponding candidate push content is determined according to the negative interest vector corresponding to each candidate push content in the candidate push list Rate.
8. The device according to claim 8, wherein the second determining module sorts the candidate push content in the candidate push list according to the click-through rate of each candidate push content in the candidate push list; according to the sorting result A reference number of candidate push content is selected from the candidate push list as the push content.
The device according to any one of claims 8-12, wherein the negatively influencing content includes at least one of short-term stay content and top-exposure unclicked content; the short-term stay content is that the user browsing time is less than the target browsing time Time content; the first exposed unclicked content is the first unclicked content in the historical push content.
The device according to any one of claims 8-12, wherein the negatively influencing content includes at least one of short-term stay content and top-exposure unclicked content; the short-term stay content is that the user browsing time is less than the target browsing time Time content; the first exposed unclicked content includes content that is displayed first in the page in the historical push content and is not clicked when the user has swiped the page.
A server, wherein the server includes a processor and a memory, and at least one piece of program code is stored in the memory, and the at least one piece of program code is loaded and executed by the processor to implement any of claims 1 to 7 1. The content push method described.
A computer-readable storage medium, wherein at least one piece of program code is stored in the computer-readable storage medium, and the at least one piece of program code is loaded and executed by a processor, so as to implement any one of claims 1 to 7 Content push method.