CN113822698A

CN113822698A - Content pushing method and device, computer equipment and storage medium

Info

Publication number: CN113822698A
Application number: CN202110740409.7A
Authority: CN
Inventors: 徐华鹏
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2021-06-30
Filing date: 2021-06-30
Publication date: 2021-12-21

Abstract

The application relates to a content pushing method, a content pushing device, a computer device and a storage medium. The method comprises the following steps: acquiring a target feature combination, acquiring a content push record set, determining a first negative content push record of a push object in the content push record set for performing negative operation on push content, and acquiring a first negative record number corresponding to the first negative content push record; determining a second negative-going content push record corresponding to the target feature combination, and acquiring a second negative-going record number corresponding to the second negative-going content push record; and determining a target feature repulsion degree corresponding to the target feature combination according to the first negative direction record quantity and the second negative direction record quantity, and filtering the candidate push content set of the target push object based on the target feature repulsion degree. By adopting the method, the accuracy of content pushing can be improved, and the candidate pushed content set in the application can be obtained by utilizing a neural network model based on artificial intelligence.

Description

Content pushing method and device, computer equipment and storage medium

Technical Field

The present application relates to the field of computer technologies, and in particular, to a content push method and apparatus, a computer device, and a storage medium.

Background

With the development of computer technology and internet technology, there is a need to push content to users in many scenarios, for example, more and more software products are provided with advertisement spots, which are positions for placing advertisements in software, and advertisers can place advertisements through the advertisement spots, so that users of the software products can see the placed advertisements.

In the conventional content push method, content is pushed to a user according to the popularity of the content or the requirements of a content provider, and the like, for example, content with higher popularity can be pushed to a user terminal based on an artificial intelligence content push method. However, the pushed content has low reliability due to different users' exclusion degrees to the content, which causes interference to the users and low content pushing accuracy.

Disclosure of Invention

In view of the above, it is necessary to provide a content push method, apparatus, computer device, and storage medium capable of improving the push accuracy.

A method of content push, the method comprising: acquiring a target feature combination, wherein the target feature combination is obtained by combining object features and content features; acquiring a content push record set, wherein the content push record set comprises a plurality of content push records, and each content push record comprises push content and a push object corresponding to the push content; determining a first negative-going content push record of a push object performing negative-going operation on push content in the content push record set, and acquiring a first negative-going record number corresponding to the first negative-going content push record; determining a second negative-going content push record corresponding to the target feature combination, and acquiring a second negative-going record quantity corresponding to the second negative-going content push record; wherein, the content features of the push content in the second negative-going content push record are consistent with the content features in the target feature combination, and the object features of the push object in the second negative-going content push record are consistent with the object features in the target feature combination; and determining a target feature rejection degree corresponding to the target feature combination according to the first negative direction record quantity and the second negative direction record quantity, wherein the target feature rejection degree represents a rejection degree between the content features and the object features in the target feature combination, and screening candidate contents to be pushed based on the target feature rejection degree corresponding to the target feature combination.

A content pushing device, the device comprising: the target feature combination acquisition module is used for acquiring a target feature combination, wherein the target feature combination is obtained by combining object features and content features; the content push record set acquisition module is used for acquiring a content push record set, wherein the content push record set comprises a plurality of content push records, and each content push record comprises push content and a push object corresponding to the push content; a first negative direction record number obtaining module, configured to determine a first negative direction content push record in which a push object performs a negative direction operation on a push content in the content push record set, and obtain a first negative direction record number corresponding to the first negative direction content push record; a second negative direction record number obtaining module, configured to determine a second negative direction content push record corresponding to the target feature combination, and obtain a second negative direction record number corresponding to the second negative direction content push record; wherein, the content features of the push content in the second negative-going content push record are consistent with the content features in the target feature combination, and the object features of the push object in the second negative-going content push record are consistent with the object features in the target feature combination; a target feature exclusion degree determining module, configured to determine a target feature exclusion degree corresponding to the target feature combination according to the first negative direction record number and the second negative direction record number, where the target feature exclusion degree indicates a degree of exclusion between a content feature and an object feature in the target feature combination, and screen candidate content to be pushed based on the target feature exclusion degree corresponding to the target feature combination.

In some embodiments, the target feature combinations are multiple, the apparatus is further configured to select, according to a target feature rejection degree corresponding to the target feature combination, a target feature combination satisfying a rejection degree screening condition as a rejection feature combination, where the rejection feature combination is used to filter, when content pushing is performed on a push object having object features in the rejection feature combination, push content having content features in the rejection feature combination; the repulsion degree screening condition includes at least one of the target feature repulsion degree being greater than the repulsion degree threshold or the target repulsion degree ranking being before the repulsion degree ranking threshold.

In some embodiments, the apparatus further comprises: a reference push record set obtaining module, configured to obtain a reference push record set, where the reference push record set includes multiple reference push records, and the reference push record is a content push record corresponding to the target feature combination; a content push value acquisition module, configured to acquire a content push value corresponding to the reference push record; a target push value obtaining module, configured to count content push values corresponding to the reference push records in the reference push record set to obtain a target push value corresponding to the target feature combination; and screening candidate contents to be pushed according to the target feature rejection degree corresponding to the target feature combination and the target pushing value.

In some embodiments, the target feature combinations are multiple, and the apparatus is further configured to select, as the rejection feature combination, a target feature combination that satisfies a rejection screening condition and a push value screening condition according to the target feature rejection and the target push value that correspond to the target feature combination; the exclusion characteristic combination is used for filtering the push content with the content characteristics in the exclusion characteristic combination when the push object with the object characteristics in the exclusion characteristic combination is subjected to content push; the push value screening condition comprises at least one of the target push value being greater than the push value threshold or the target push value ranking being before the value ranking threshold; the repulsion degree screening condition includes at least one of the target feature repulsion degree being greater than the repulsion degree threshold or the target repulsion degree ranking being before the repulsion degree ranking threshold.

In some embodiments, the target push value derivation module comprises: the overall content push value obtaining unit is used for summing the content push values corresponding to the reference push records in the reference push record set to obtain an overall content push value; a reference content record number obtaining unit, configured to obtain a number of reference push records in a reference push record set as a reference content record number; and the target push value obtaining unit is used for calculating the ratio of the total content push value to the reference content record number and obtaining the target push value corresponding to the target feature combination based on the calculated ratio.

In some embodiments, the target feature rejections determination module comprises: a content push record number determining unit, configured to determine a content push record number corresponding to the content push record set; a first negative direction record possibility obtaining unit, configured to calculate a ratio of the first negative direction record number to the content push record number, and obtain a first negative direction record possibility according to the calculated ratio; a second negative-direction record possibility obtaining unit configured to calculate a ratio of the second negative-direction record number to the first negative-direction record number, and obtain a second negative-direction record possibility according to the calculated ratio; and a target feature exclusion degree obtaining unit, configured to obtain a target feature exclusion degree corresponding to the target feature combination based on the first negative-direction recording possibility degree and the second negative-direction recording possibility degree.

In some embodiments, the target feature exclusion degree obtaining unit is further configured to calculate a product of the first negative-direction record possibility degree and the second negative-direction record possibility degree, so as to obtain a standard negative-direction record possibility degree corresponding to the target feature combination; acquiring the number of content push records corresponding to the target feature combination in the content push record set as the number of combined records; calculating the ratio of the number of the combined records to the number of the content push records, and obtaining the push record possibility corresponding to the target feature combination according to the calculated ratio; and obtaining a target feature rejection degree corresponding to the target feature combination based on the standard negative direction recording possibility degree and the push recording possibility degree.

In some embodiments, the target feature combination acquisition module comprises: a feature combination configuration interface sending unit, configured to send a feature combination configuration interface to a configuration terminal, so that the configuration terminal displays the feature combination configuration interface, and obtains a configured target feature combination in response to a configuration operation for the feature combination configuration interface; and the target characteristic combination receiving unit is used for receiving the target characteristic combination sent by the configuration terminal.

A computer device comprising a memory and a processor, the memory storing a computer program, the processor implementing the following steps when executing the computer program: acquiring a target feature combination, wherein the target feature combination is obtained by combining object features and content features; acquiring a content push record set, wherein the content push record set comprises a plurality of content push records, and each content push record comprises push content and a push object corresponding to the push content; determining a first negative-going content push record of a push object performing negative-going operation on push content in the content push record set, and acquiring a first negative-going record number corresponding to the first negative-going content push record; determining a second negative-going content push record corresponding to the target feature combination, and acquiring a second negative-going record quantity corresponding to the second negative-going content push record; wherein, the content features of the push content in the second negative-going content push record are consistent with the content features in the target feature combination, and the object features of the push object in the second negative-going content push record are consistent with the object features in the target feature combination; and determining a target feature rejection degree corresponding to the target feature combination according to the first negative direction record quantity and the second negative direction record quantity, wherein the target feature rejection degree represents a rejection degree between the content features and the object features in the target feature combination, and screening candidate contents to be pushed based on the target feature rejection degree corresponding to the target feature combination.

A computer-readable storage medium, on which a computer program is stored which, when executed by a processor, carries out the steps of: acquiring a target feature combination, wherein the target feature combination is obtained by combining object features and content features; acquiring a content push record set, wherein the content push record set comprises a plurality of content push records, and each content push record comprises push content and a push object corresponding to the push content; determining a first negative-going content push record of a push object performing negative-going operation on push content in the content push record set, and acquiring a first negative-going record number corresponding to the first negative-going content push record; determining a second negative-going content push record corresponding to the target feature combination, and acquiring a second negative-going record quantity corresponding to the second negative-going content push record; wherein, the content features of the push content in the second negative-going content push record are consistent with the content features in the target feature combination, and the object features of the push object in the second negative-going content push record are consistent with the object features in the target feature combination; and determining a target feature rejection degree corresponding to the target feature combination according to the first negative direction record quantity and the second negative direction record quantity, wherein the target feature rejection degree represents a rejection degree between the content features and the object features in the target feature combination, and screening candidate contents to be pushed based on the target feature rejection degree corresponding to the target feature combination.

In some embodiments, a computer program product or computer program is provided that includes computer instructions stored in a computer-readable storage medium. The computer instructions are read by a processor of a computer device from a computer-readable storage medium, and the computer instructions are executed by the processor to cause the computer device to perform the steps in the above-mentioned method embodiments.

The content push method, the content push device, the computer equipment and the storage medium determine a first negative content push record in which a push object performs negative operation on push content in a content push record set, obtain a first negative record number corresponding to the first negative content push record, so that the first negative record number reflects the number of content push records performing negative operation on the push content, determine a second negative content push record corresponding to a target feature combination, and obtain a second negative record number corresponding to the second negative content push record And determining a target feature repulsion degree corresponding to the target feature combination according to the first negative direction record quantity and the second negative direction record quantity, wherein the target feature repulsion degree represents the repulsion degree between the content features and the object features in the target feature combination, so that the target feature repulsion degree reflects the repulsion degree of the object with the object features in the target feature combination to the content with the content features in the target feature combination, and the larger the target feature repulsion degree is, the more the object is repelled to the content. Therefore, when the candidate contents to be pushed are screened based on the target feature rejection degree corresponding to the target feature combination, the candidate contents with the larger target feature rejection degree can be filtered from each candidate content, that is, the candidate contents corresponding to the content features with the larger rejection degree between the object features of the object of the contents to be pushed are filtered, so that the contents with the larger rejection degree between the object pushing and the object features are reduced, the possibility of pushing the rejected contents to the object is reduced, and the accuracy of the content pushing is improved.

A method of content push, the method comprising: receiving a content push request aiming at a target push object; responding to the content pushing request, and acquiring a candidate pushing content set corresponding to the target pushing object; determining a target rejection characteristic combination based on the target object characteristics corresponding to the target push object, wherein the target rejection characteristic combination is determined based on the target characteristic rejection degree corresponding to the target characteristic combination; and filtering the candidate push contents corresponding to the target rejection characteristic combination from the candidate push content set to obtain the target push contents corresponding to the target push object.

A content pushing device, the device comprising: the content pushing request receiving module is used for receiving a content pushing request aiming at a target pushing object; a candidate push content set acquisition module, configured to respond to the content push request, and acquire a candidate push content set corresponding to the target push object; the target rejection characteristic combination determining module is used for determining a target rejection characteristic combination based on the target object characteristics corresponding to the target push object, wherein the target rejection characteristic combination is determined based on the target characteristic rejection degree corresponding to the target characteristic combination; and the target push content obtaining module is used for filtering the candidate push content corresponding to the target rejection characteristic combination from the candidate push content set to obtain the target push content corresponding to the target push object.

In some embodiments, the target-exclusive feature combination determination module comprises: a first candidate repulsive feature combination obtaining unit configured to obtain a plurality of candidate repulsive feature combinations; the candidate rejection feature combinations are obtained by screening from each target feature combination based on the target feature rejection degrees corresponding to the target feature combinations; and the target exclusion feature combination obtaining unit is used for comparing the target object features corresponding to the target push object with the object features corresponding to the push objects in the candidate exclusion feature combination, and taking the exclusion feature combination with consistent comparison as the target exclusion feature combination. In some embodiments, the apparatus further comprises a candidate repulsive feature combination derivation module that comprises: a target feature combination obtaining unit configured to obtain a plurality of target feature combinations; the target feature combination is obtained by combining object features and content features; a content push record set obtaining unit, configured to obtain a content push record set, where the content push record set includes multiple content push records, and each content push record includes a push content and a push object corresponding to the push content; a first negative direction record number obtaining unit, configured to determine a first negative direction content push record in which a push object performs a negative direction operation on a push content in the content push record set, and obtain a first negative direction record number corresponding to the first negative direction content push record; a second negative direction record number obtaining unit, configured to determine a second negative direction content push record corresponding to the target feature combination, and obtain a second negative direction record number corresponding to the second negative direction content push record; wherein, the content features of the push content in the second negative-going content push record are consistent with the content features in the target feature combination, and the object features of the push object in the second negative-going content push record are consistent with the object features in the target feature combination; a target feature exclusion degree determining unit, configured to determine, according to the first negative direction record number and the second negative direction record number, a target feature exclusion degree corresponding to the target feature combination, where the target feature exclusion degree indicates a degree of exclusion between a content feature and an object feature in the target feature combination; and the second candidate rejection feature combination obtaining unit is used for screening candidate rejection feature combinations from the target feature combinations based on the target feature rejection degrees corresponding to the target feature combinations.

A computer device comprising a memory and a processor, the memory storing a computer program, the processor implementing the following steps when executing the computer program: receiving a content push request aiming at a target push object; responding to the content pushing request, and acquiring a candidate pushing content set corresponding to the target pushing object; determining a target rejection characteristic combination based on the target object characteristics corresponding to the target push object, wherein the target rejection characteristic combination is determined based on the target characteristic rejection degree corresponding to the target characteristic combination; and filtering the candidate push contents corresponding to the target rejection characteristic combination from the candidate push content set to obtain the target push contents corresponding to the target push object.

A computer-readable storage medium, on which a computer program is stored which, when executed by a processor, carries out the steps of: receiving a content push request aiming at a target push object; responding to the content pushing request, and acquiring a candidate pushing content set corresponding to the target pushing object; determining a target rejection characteristic combination based on the target object characteristics corresponding to the target push object, wherein the target rejection characteristic combination is determined based on the target characteristic rejection degree corresponding to the target characteristic combination; and filtering the candidate push contents corresponding to the target rejection characteristic combination from the candidate push content set to obtain the target push contents corresponding to the target push object.

The content push method, the content push device, the computer equipment and the storage medium receive a content push request aiming at a target push object, respond to the content push request, obtain a candidate push content set corresponding to the target push object, determine a target rejection characteristic combination based on the target object characteristics corresponding to the target push object, because the target rejection characteristic combination is determined based on the target characteristic rejection degree corresponding to the target characteristic combination, and the target characteristic rejection degree reflects the rejection degree of the object with the object characteristics in the target characteristic combination to the content with the content characteristics in the target characteristic combination, the larger the target characteristic rejection degree is, the more the object rejects the content, thereby the target characteristic combination with the larger target characteristic rejection degree can be used as the target rejection characteristic combination, and the candidate push content corresponding to the target rejection characteristic combination is filtered from the candidate push content set, the candidate push contents rejected by the target push object can be filtered, so that the possibility of pushing the rejected contents to the target push object is reduced, and the accuracy of content pushing is improved.

Drawings

FIG. 1 is a diagram of an application environment of a content push method in some embodiments;

FIG. 2 is a flow diagram illustrating a method for pushing content in some embodiments;

FIG. 3 is a schematic diagram of a content push method in some embodiments;

FIG. 4 is a graph of the rejectional score of academic educational advertisements across different academic demographics in some embodiments;

FIG. 5 is a flow diagram illustrating a method for pushing content in some embodiments;

FIG. 6 is a schematic diagram of a content push method in some embodiments;

FIG. 7 is a block diagram of a content push device in some embodiments;

FIG. 8 is a block diagram of a content push device in some embodiments;

FIG. 9 is a diagram of the internal structure of a computer device in some embodiments;

FIG. 10 is a diagram of the internal structure of a computer device in some embodiments.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.

The content pushing method provided by the application can be applied to the application environment shown in fig. 1. Wherein the terminal 102 communicates with the server 104 via a network. The server 104 may be a server for pushing content, the terminal 102 may be installed with a client corresponding to the server 104, the server 104 may push content to the terminal 102, and the terminal 102 may display the pushed content through the client. For example, the terminal 102 may receive an Advertisement (AD) pushed by the server 104 and display the pushed content in the client, where the interface displayed by the client may include a pushed content display area, the pushed content display area is used for displaying the pushed content, and the pushed content display area may be predetermined.

Specifically, the server 104 may obtain a target feature combination, where the target feature combination is obtained by combining an object feature and a content feature, obtain a content push record set, where the content push record set includes a plurality of content push records, where the content push record includes push content and push objects corresponding to the push content, determine a first negative content push record in the content push record set, where the push objects perform negative operations on the push content, obtain a first negative record number corresponding to the first negative content push record, determine a second negative content push record corresponding to the target feature combination, obtain a second negative record number corresponding to the second negative content push record, and the content characteristics of the push content in the second negative-going content push record are consistent with the content characteristics in the target characteristic combination, and the object characteristics of the push object in the second negative-going content push record are consistent with the object characteristics in the target characteristic combination. The server 104 may further determine a target feature exclusion degree corresponding to the target feature combination according to the first negative direction record number and the second negative direction record number, where the target feature exclusion degree indicates a degree of exclusion between the content feature and the object feature in the target feature combination, so as to screen the candidate content to be pushed based on the target feature exclusion degree corresponding to the target feature combination. For example, the server 104 may select a target feature combination satisfying the exclusion degree screening condition according to a target feature exclusion degree corresponding to the target feature combination, as the exclusion feature combination, when a target push object corresponding to the terminal 102 has an object feature in the exclusion feature combination, and when the server 104 receives a content push request for the target push object, a candidate push content set corresponding to the target push object may be obtained in response to the content push request, and the server 104 may filter candidate push contents having a content feature in the exclusion feature combination from the candidate push content set, and push remaining candidate push contents in the candidate push content set to the terminal 102.

Artificial Intelligence (AI) is a theory, method, technique and application system that uses a digital computer or a machine controlled by a digital computer to simulate, extend and expand human Intelligence, perceive the environment, acquire knowledge and use the knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technique of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can react in a manner similar to human intelligence. Artificial intelligence is the research of the design principle and the realization method of various intelligent machines, so that the machines have the functions of perception, reasoning and decision making.

The artificial intelligence technology is a comprehensive subject and relates to the field of extensive technology, namely the technology of a hardware level and the technology of a software level. The artificial intelligence infrastructure generally includes technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation/interaction systems, mechatronics, and the like. The artificial intelligence software technology mainly comprises a computer vision technology, a voice processing technology, a natural language processing technology, machine learning/deep learning, automatic driving, intelligent traffic and the like.

Machine Learning (ML) is a multi-domain cross discipline, and relates to a plurality of disciplines such as probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and the like. The special research on how a computer simulates or realizes the learning behavior of human beings so as to acquire new knowledge or skills and reorganize the existing knowledge structure to continuously improve the performance of the computer. Machine learning is the core of artificial intelligence, is the fundamental approach for computers to have intelligence, and is applied to all fields of artificial intelligence. Machine learning and deep learning generally include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and formal education learning.

With the research and progress of artificial intelligence technology, the artificial intelligence technology is developed and researched in a plurality of fields, such as common smart homes, smart wearable devices, virtual assistants, smart speakers, smart marketing, unmanned driving, automatic driving, unmanned aerial vehicles, robots, smart medical services, smart customer service, internet of vehicles, automatic driving, smart traffic and the like.

The scheme provided by the embodiment of the application relates to the technologies such as machine learning of artificial intelligence and the like, and is specifically explained by the following embodiment: the server 104 may obtain a candidate push content set corresponding to the target push object by using a content push model. Wherein, the content push model may be a trained neural network model. The content push model can determine candidate object content corresponding to the push object according to the object information of the push object and the content information of the push content. For example, the server may obtain pre-stored push objects, determine, by using a content push model, respective candidate push contents corresponding to a target push object from the pre-stored push objects, and compose a candidate push content set.

The terminal 102 may be, but not limited to, various personal computers, notebook computers, smart phones, tablet computers, and portable wearable devices, and the server 104 may be implemented by an independent server or a server cluster formed by a plurality of servers.

It is to be understood that the above application scenario is only an example, and does not constitute a limitation to the content push method provided in the embodiment of the present application, and the method provided in the embodiment of the present application may also be applied in other application scenarios, for example, the content push method provided in the present application may be executed by the terminal 102, the terminal 102 may upload the obtained target feature exclusion degree corresponding to the target feature combination to the server 104, the server 104 may store the target feature exclusion degree corresponding to the target feature combination, and may also forward the target feature exclusion degree corresponding to the target feature combination to other devices.

In some embodiments, as shown in fig. 2, a content pushing method is provided, where the method may be executed by a terminal or a server, or may be executed by both the terminal and the server, and in this embodiment, the method is described as applied to the server 104 in fig. 1, and includes the following steps:

s202, acquiring a target feature combination, wherein the target feature combination is obtained by combining object features and content features.

The object refers to a thing to which content can be pushed, for example, the object may be a user, the object may also be referred to as a pushing object, the object characteristic refers to a characteristic possessed by an attribute of the pushing object, and the pushing object may have a plurality of attributes, for example, when the pushing object is a user, the attribute of the pushing object may include at least one attribute of age, gender, love and marriage status, occupation, academic calendar, and the like. Each object attribute may correspond to an object feature. Object properties refer to properties of the pushed object. The object features may be user features, for example. The user characteristics may include at least one of age characteristics, gender characteristics, or occupation characteristics, among others.

Content refers to information that may be pushed, content may be, for example, advertisements, which may be, for example, online advertisements. Online advertisements may also be referred to as internet advertisements, which refers to advertisements placed on ad spots on internet platforms. The content may also be referred to as push content. The advertiser can realize the advertisement putting through the advertisement trading platform. An AD Exchange (ADX) is an entity that connects a media host and an advertiser, and the AD Exchange places the advertiser's AD on an AD spot provided by the media host. The advertisement trading platform can collect information of users, create user figures, and deliver advertisements to the users according to data of interests or geographic positions of the users. The content pushing method provided by the application can be deployed in an advertisement trading platform, so that the accuracy of advertisement putting is improved.

The content feature refers to a feature possessed by an attribute Of the content, the content may have multiple attributes, for example, when the content is an advertisement, at least one attribute Of an industry, a commodity, an optimization target, a sensitive word or a targeted POI (Point Of Interest, information Point location/Interest Point), and the like may be included, and each content attribute may correspond to a content feature. Content attributes refer to attributes of the pushed content. The content feature may be, for example, an advertisement feature. The advertisement characteristics may include at least one of industry characteristics, sensitive word characteristics. The directional POI means POI positioning, the POI positioning means information point positioning, and the POI can be a house, a shop, a mailbox, a bus station or the like. The optimization target (bid _ objective) refers to the behavior that the advertiser expects the user to generate the advertisement, for example, the advertiser in the game industry can use the "download" behavior as the optimization target of the advertisement when the advertiser puts the APP game advertisement.

The feature combination is formed by combining object features and content features, and each feature combination comprises an object feature and a content feature. At least one of the object feature and the content feature included in the different feature combinations is not the same. The object features included in different feature combinations may be the same or different. The content features included in different feature combinations may be the same or different. For example, the target feature set 1 is (Ua1, Ub1), the target feature set 2 is (Ua1, Ub2), among which Ua1 is an object feature, and Ub1 and Ub2 are content features. The target feature combination may be any feature combination, or may be a feature combination selected from a plurality of feature combinations. There may be one or more combinations of target features, a plurality referring to at least two.

Specifically, the server may determine an object attribute set corresponding to the pushed object, where the object attribute set includes a plurality of object attributes, and for example, the object attribute set may include object attributes such as age, gender, academic calendar, occupation, and the like. The server can obtain an attribute value corresponding to an object attribute of the pushed object, and determine an object feature corresponding to the object attribute based on the attribute value corresponding to the object attribute. For example, the server may perform normalization processing on the attribute values, and take the result of the normalization processing as the object feature, or the server may determine a range to which the attribute values belong, and determine the object feature based on the range to which the attribute values belong.

In some embodiments, the server may obtain an attribute value corresponding to an object attribute of the push object, and determine the object characteristic based on the attribute value, for example, when the object attribute is age, if the age of the push object is 30 years, the attribute value is 30. For example, when the object attribute is age, the age of the object 1 is 20, and the age of the object 2 is 50, then for age, the attribute value obtained by the object 1 is 20, and the attribute value obtained by the object 2 is 50. The server can use the object characteristics corresponding to the object attributes to represent the push object. For example, a push object may be represented by (a1, a2, … ai …, an), where ai represents an object feature corresponding to the ith object attribute of the push object. i is a positive integer of 1 or more and n or less.

In some embodiments, the server may obtain a content attribute set corresponding to the pushed content, where the content attribute set includes a plurality of content attributes, for example, the content attribute set may include content attributes such as industry, goods, or sensitive words. The server may obtain an attribute value corresponding to a content attribute of the pushed content, and determine a content feature corresponding to the content attribute based on the attribute value of the content attribute. For example, the attribute values of the content attributes may be normalized to obtain the content features, and certainly, the content features may also be obtained in other manners, which is not limited herein. For example, when the content attribute is a content type, the content type of the push content a is a love category, and the content type of the push content B is a non-love category, then for the content type, the attribute value obtained by the push content a is a love category, and the attribute value obtained by the push content B is a non-love category. The server can represent the push content by using the content characteristics corresponding to the content attributes respectively. For example, one push content may be represented by (b1, b2, … bj …, bm), where bj represents a content attribute of the push content. j is a positive integer of 1 or more and m or less.

In some embodiments, the server may obtain a plurality of attribute combinations based on the content attribute set and the object attribute set, where an attribute combination is a combination of a content attribute and an object attribute, and each attribute combination includes a content attribute and an object attribute. For example, the server may obtain a content attribute from the set of content attributes, obtain an object attribute from the set of object attributes, and combine the obtained content attribute and the obtained object attribute into an attribute combination. The server may determine an object feature corresponding to an object attribute in the attribute combination, determine a content feature corresponding to a content attribute in the attribute combination, and combine the determined object feature and the determined content feature into a candidate feature combination. Because the object characteristics obtained by different pushed objects may be different in the same object attribute, there may be a plurality of object characteristics corresponding to the same object attribute, and similarly, there may be a plurality of content characteristics corresponding to the same content attribute, so that a plurality of candidate characteristic combinations may be obtained by the same attribute combination. For example, the attribute combination is (gender, content type), since "gender" may include "male" and "female" and "content type" may include "marriage" and "non-marriage", assuming that "male" corresponds to a feature of 1, and "female" corresponds to a feature of 0, and "content type" corresponds to a feature of 2, and "non-marriage" corresponds to a feature of 4, the candidate feature combination may include at least one of (1, 2), (1, 4), (0, 2), or (0, 4).

In some embodiments, the target feature combination may be filtered from a set of candidate feature combinations. The set of candidate feature combinations may include a plurality of candidate feature combinations. The server may obtain the reference feature exclusion degrees corresponding to the candidate feature combinations, where the reference feature exclusion degrees may be preset, for example, preset according to expert knowledge. The reference feature exclusion degree is used for predicting the exclusion degree of the object with the object features in the candidate feature combination to the content with the content features in the candidate feature combination, and the reference feature exclusion degree is a predicted value and may be different from the real feature exclusion degree. The server may filter the target feature combination from the candidate feature combination set based on the reference feature exclusion degree, for example, may obtain a candidate feature combination with the reference feature exclusion degree greater than a reference threshold from the candidate feature combination as the target feature combination. Or the server may obtain, from the candidate feature combination set, a candidate feature combination with a reference rejection degree ranked before the reference rejection degree threshold as a target feature combination, where the reference threshold and the reference rejection degree threshold may be preset or may be set as needed. The reference repulsion degree sequence is a sequence obtained by arranging the reference feature repulsion degrees corresponding to each candidate feature combination according to the sequence from large to small, and the larger the reference feature sequence degree is, the higher the sequence in the reference repulsion degree sequence is. The degree of reference feature exclusion may also be determined according to the knowledge in the relevant fields of operation or products, for example, obtained by analysis on the basis of general knowledge: the degree of rejection of the people with the high school calendar to the educational advertisement of the school calendar is high, and the degree of rejection of the reference feature of the candidate feature combination (the features of the doctor and the features corresponding to the educational education of the school calendar) is set to be larger if the people with the high school calendar are, for example, doctor people.

S204, a content push record set is obtained, the content push record set comprises a plurality of content push records, and the content push records comprise push contents and push objects corresponding to the push contents.

The push content refers to push content, and may be, for example, a push advertisement. The push object corresponding to the push content is the push object to which the push content is pushed, for example, the push content is S1, and when the push content S1 is pushed to the object 1, the object 1 is the push object corresponding to the push content S1. Each content push record may include a push content and a push object, and the content push record may be, for example, "push content S1 is pushed to push object 1".

Specifically, the content push record may be obtained according to a log of the push content, and when the server performs content push, the server may record the push content and a push object corresponding to the push content in the log, and may record an object behavior of the push object on the push content in the log. The object behavior refers to the behavior of the push object with respect to the push content. The object behavior may also be referred to as user behavior, and the user behavior may include at least one of exposure, negative feedback behavior, or positive feedback behavior, but may also include other behaviors, such as conversion behavior. The conversion behavior refers to a behavior that occurs on the advertisement trading platform by the user, for example, for APP (Application) type advertisements, and may include at least one of downloading, activating, paying, and the like.

In some embodiments, the content push records in the set of content push records may be derived based on a targeted combination of features. Specifically, the server may store a plurality of push objects, and the server may obtain, as the first push object, a push object having an object feature in the target feature combination from the plurality of push objects; the server can also store a plurality of pieces of push content, and the server can acquire a plurality of pieces of push content with the content characteristics in the target characteristic combination from the plurality of pieces of push content to serve as first push content; the method comprises the steps of pushing a first push content to a first push object, obtaining an operation type of the first push object aiming at the first push content, taking the first push content as a push content in a content push record, taking the first push object as a push object in the content push record, and taking the operation type as an operation type in the content push record to generate the content push record. One push may correspond to one content push record. The target feature combination may be plural. The content push record set may include content push records corresponding to the respective target feature combinations.

In some embodiments, the set of content push records may be a set of a plurality of content push records resulting from content pushes within a historical period of time.

S206, determining a first negative-direction content push record of a push object in the content push record set for performing negative-direction operation on the push content, and acquiring the number of the first negative-direction records corresponding to the first negative-direction content push record.

The operations for pushing the content may be various, and the operations for pushing the content may be divided into various operation types, for example, the operations for pushing the content may include at least one operation type of a positive operation or a negative operation. The negative operation is used for reflecting that the object is exclusive to the pushed content, and the negative operation can be set as required or preset, for example, a displayed advertisement can have a control of 'no interest', and the negative operation can be a trigger operation of the control of 'no interest'. Negative operation is a concept relative to positive operation. The operations for pushing content may be considered positive operations other than negative operations, e.g., positive operations may include at least one of a click or a like. Positive operation may also be referred to as positive feedback behavior and negative operation may also be referred to as negative feedback behavior.

The content push record may further include an operation type of the push object on the push content, where the operation type may be any one of a negative operation and a positive operation. The first negative-going content push record is a content push record whose operation type is a negative-going operation. The first negative-going number of records refers to the number of first negative-going content push records.

Specifically, the server may obtain operation types corresponding to each content push record in the content push record set, use a content push record whose operation type is a negative operation as a first negative content push record, count the number of each first negative content push record, and use the counted number as the number of the first negative records.

S208, determining a second negative-going content push record corresponding to the target feature combination, and acquiring a second negative-going record number corresponding to the second negative-going content push record; and the content characteristics of the pushed content in the second negative-going content pushing record are consistent with the content characteristics in the target characteristic combination, and the object characteristics of the pushed object in the second negative-going content pushing record are consistent with the object characteristics in the target characteristic combination.

The content features of the push content in the second negative-going content push record are the same as the content features in the target feature combination, and the object features of the push object in the second negative-going content push record are the same as the object features in the target feature combination.

Specifically, the second negative-going number of records refers to a number of second negative-going content push records. The server may determine, from the content push record set, a second negative-going content push record corresponding to the target feature combination, or may determine, from each of the first negative-going content push records, a second negative-going content push record corresponding to the target feature combination.

In some embodiments, the server may determine an object attribute corresponding to an object feature in the target feature combination, as a target object attribute, determine a content attribute corresponding to a content feature in the target feature combination, as a target content attribute, obtain an object feature of a push object in the first negative-direction content push record on the target object attribute, as a comparison object feature, obtain a content feature of the push content in the first negative-direction content push record on the target content attribute, as a comparison content feature, and filter, based on the comparison object feature and the comparison content feature, a second negative-direction content push record from each of the first negative-direction content push records. For example, the server may compare the object feature in the target feature combination with the comparison object feature, compare the content feature in the target feature combination with the comparison content feature when the comparison is consistent, and use the first negative-going content push record as the second negative-going content push record when the comparison is consistent. Or the server may compare the content features in the target feature combination with the comparison content features, compare the object features in the target feature combination with the comparison object features when the comparison is consistent, and use the first negative-going content push record as the second negative-going content push record when the comparison is consistent.

In some embodiments, when the object feature in the target feature combination is not consistent with the comparison object feature in comparison, or when the content feature in the target feature combination is not consistent with the comparison content feature in comparison, it is determined that the first negative-going content push record is not the second negative-going content push record corresponding to the target feature combination.

S210, determining a target feature rejection degree corresponding to the target feature combination according to the first negative direction record quantity and the second negative direction record quantity, wherein the target feature rejection degree represents the rejection degree between the content features and the object features in the target feature combination, and screening candidate contents to be pushed based on the target feature rejection degree corresponding to the target feature combination.

Wherein the target feature rejections are positively correlated with the second negative-going number of records. The target feature exclusion degree is used for reflecting the exclusion degree of the push object with the object feature to the push content with the content feature. The rejection degree and the rejection degree of the target feature are in a positive correlation relationship, and the greater the rejection degree of the target feature is, the greater the rejection degree is. The candidate content refers to content to be pushed. The candidate contents corresponding to different objects may be the same or different. The candidate content to be pushed refers to the pushed content corresponding to the pushing object needing content pushing, and the candidate content to be pushed corresponding to different pushing objects may be the same or different.

Wherein, the positive correlation refers to: under the condition that other conditions are not changed, the changing directions of the two variables are the same, and when one variable changes from large to small, the other variable also changes from large to small. It is understood that a positive correlation herein means that the direction of change is consistent, but does not require that when one variable changes at all, another variable must also change. For example, it may be set that the variable b is 100 when the variable a is 10 to 20, and the variable b is 120 when the variable a is 20 to 30. Thus, the change directions of a and b are both such that when a is larger, b is also larger. But b may be unchanged in the range of 10 to 20 a. The negative correlation relationship refers to: under the condition that other conditions are not changed, the changing directions of the two variables are opposite, and when one variable is changed from large to small, the other variable is changed from small to large. It is understood that the negative correlation herein means that the direction of change is reversed, but it is not required that when one variable changes at all, the other variable must also change.

Specifically, when the server obtains a content push request for a target push object, candidate content to be pushed corresponding to the target push object may be obtained, for example, a candidate push content set corresponding to the target push object may be obtained, candidate push content in the candidate push content set of the target push object is filtered based on a target feature rejection degree corresponding to a target feature combination, and remaining candidate push content in the candidate push content set after filtering is pushed to the target push object. The target push object may be any push object.

In some embodiments, the server may obtain feature exclusion degrees corresponding to each candidate push content in the candidate push content set, and arrange the candidate push contents according to a sequence from small feature exclusion degrees to large feature exclusion degrees to obtain a candidate push content sequence, where the smaller the feature exclusion degree is, the earlier the candidate push contents are ranked in the candidate push content sequence, and when content is pushed to a target push object, the content with the earlier rank in the candidate push content sequence is preferentially pushed, so that a possibility that the content with a larger exclusion degree is pushed to the target push object is reduced.

In some embodiments, the server may count the number of content push records included in the content push record set, use the counted number as the number of content push records, and determine the target feature rejections corresponding to the target feature combinations based on the first negative direction record number, the second negative direction record number, and the number of content push records. For example, the server may calculate a ratio of the first negative direction record number to the content push record number, obtain a first negative direction record possibility according to the calculated ratio, calculate a ratio of the second negative direction record number to the first negative direction record number, obtain a second negative direction record possibility according to the calculated ratio, and calculate a target feature rejection corresponding to the target feature combination based on the first negative direction record possibility and the second negative direction record possibility.

In some embodiments, the target feature combinations are multiple, the server may screen the rejection feature combinations from the respective target feature combinations based on the target feature rejection degrees, and screen the candidate content to be pushed based on the rejection feature combinations.

In some embodiments, the server may obtain a target push value corresponding to the target feature combination, obtain a rejection feature combination by screening from each target feature combination based on the target feature rejection degree and the target push value, and screen the candidate content to be pushed based on the rejection feature combination. Wherein the target push value is used for reflecting the benefit that pushing the push content with the content characteristics in the target characteristic combination to the push object with the object characteristics in the target characteristic combination can bring to the owner of the push content.

In the content push method, a first negative-going content push record in which a push object performs negative-going operations on a push content in a content push record set is determined, a first negative-going record number corresponding to the first negative-going content push record is obtained, so that the first negative-going record number reflects the number of content push records performing negative-going operations on the push content, a second negative-going content push record corresponding to a target feature combination is determined, a second negative-going record number corresponding to the second negative-going content push record is obtained, because content features of the push content in the second negative-going content push record are consistent with content features in the target feature combination, object features of the push object in the second negative-going content push record are consistent with object features in the target feature combination, so that the second negative-going record number reflects the number of content push records corresponding to the target feature combination in each content push record performing negative-going operations, and determining a target feature repulsion degree corresponding to the target feature combination according to the first negative direction record quantity and the second negative direction record quantity, wherein the target feature repulsion degree represents the repulsion degree between the content features and the object features in the target feature combination, so that the target feature repulsion degree reflects the repulsion degree of the object with the object features in the target feature combination to the content with the content features in the target feature combination, and the larger the target feature repulsion degree is, the more the object is repelled to the content. Therefore, when the candidate contents to be pushed are screened based on the target feature rejection degree corresponding to the target feature combination, the candidate contents with the larger target feature rejection degree can be filtered from each candidate content, that is, the candidate contents corresponding to the content features with the larger rejection degree between the object features of the object of the contents to be pushed are filtered, so that the contents with the larger rejection degree between the object pushing and the object features are reduced, the possibility of pushing the rejected contents to the object is reduced, and the accuracy of the content pushing is improved.

In some embodiments, the target feature is combined into a plurality, the method further comprising: selecting a target feature combination meeting rejection degree screening conditions according to the rejection degree of the target features corresponding to the target feature combination to serve as a rejection feature combination, wherein the rejection feature combination is used for filtering push contents with content features in the rejection feature combination when the push objects with the object features in the rejection feature combination are subjected to content push; the repulsion degree screening condition includes at least one of the target feature repulsion degree being greater than the repulsion degree threshold or the target repulsion degree ranking being before the repulsion degree ranking threshold.

The rejection feature combination refers to a target feature combination satisfying the rejection degree screening condition among the target feature combinations. The combination of the exclusion characteristics may be one or more, and a plurality means at least two. The magnitude of the rejection degree threshold may be predetermined, the rejection degree threshold may be a fixed value such as 0.9 or 0.8, and the rejection degree threshold may also be determined according to the target feature rejection degrees respectively corresponding to each target feature combination, for example, the server may perform a mean value operation on the rejection degrees of the target features respectively corresponding to each target feature combination to obtain a rejection degree mean value, obtain a preset rejection degree coefficient, perform a multiplication operation on the rejection degree mean value and the rejection degree coefficient, and use the result of the multiplication operation as the rejection degree threshold. Wherein the repulsion coefficient may be a positive number greater than 1.

The target rejection degree sequence refers to the sequence of the target feature rejection degrees in the rejection degree sequence, and the rejection degree sequence is a sequence obtained by arranging the target feature rejection degrees respectively corresponding to each target feature combination according to the sequence from large to small. The greater the degree of repulsion of the target features, the more advanced the ordering in the sequence of degrees of repulsion.

The magnitude of the rejection degree sorting threshold may be preset, for example, the rejection degree sorting threshold may be a fixed value such as 10 or 15, or the rejection degree sorting threshold may also be determined according to the number of the target feature combinations, for example, the server may obtain a first sorting coefficient, the first sorting coefficient may be preset, the first sorting coefficient may be a positive number smaller than 1, and may be 0.5 or 0.6, for example, the first sorting coefficient and the number of the target feature combinations are multiplied, and a result of the multiplication is used as the rejection degree sorting threshold.

Specifically, the server may arrange the target feature rejections of each target feature combination in descending order, use the arranged sequence as a rejections sequence, compare the ranking of the target feature rejections of the target feature combinations in the rejections sequence with a rejections ranking threshold, i.e., compare the target rejection ranking corresponding to the target feature rejections with the rejections ranking threshold, and use the target feature combination corresponding to the target feature rejections as a rejections feature combination when the target rejection ranking corresponding to the target feature rejections is before the rejections ranking threshold. As shown in fig. 3, the server obtains a target feature combination based on expert knowledge, obtains a content push record set based on a log of push content, determines a target feature rejection corresponding to the target feature combination based on the content push record set, and screens each target feature combination by using a rejection screening condition to obtain a rejection feature combination.

In some embodiments, the server may compare the target feature rejection degree with a rejection degree threshold, and when the target feature rejection degree is greater than the rejection degree threshold, take a target feature combination corresponding to the target feature rejection degree as the rejection feature combination.

In some embodiments, the server may treat as the exclusion feature combination, each target feature combination for which the target feature repulsion is greater than the repulsion threshold and for which the target repulsion is ranked before the repulsion ranking threshold.

In some embodiments, when the server obtains a content push request for a target push object, the server may compare an object feature of the target push object with an object feature in an exclusion feature combination, when the comparison is consistent, obtain a content feature in the exclusion feature combination with which the comparison is consistent, obtain a candidate push content set of the target push object, compare a content feature of each candidate push content in the candidate push content set with a content feature in the exclusion feature combination with which the comparison is consistent, filter the candidate push content with which the comparison is consistent from the candidate push content set, use remaining candidate push contents in the candidate push content set as target push contents to be pushed to the target push object, and the server may push each target push content to the target push object.

In this embodiment, a target feature combination meeting the rejection degree screening condition is selected according to the rejection degree of the target feature combination, and is used as the rejection feature combination, because the rejection degree screening condition includes that the rejection degree of the target feature is greater than the rejection degree threshold or the rejection degree of the target is at least one of the rejection degree before the rejection degree sorting threshold, the target feature combination with the greater rejection degree of the target feature can be used as the rejection feature combination, and because the rejection feature combination is used for filtering the push content with the content feature in the rejection feature combination when pushing the content of the push object with the object feature in the rejection feature combination, the content corresponding to the content feature with the greater rejection degree of the object feature can be filtered out, the content with the greater rejection degree of the target feature rejection degree is pushed to the push object is reduced, and the situation of pushing the rejected content to the push object is reduced, the accuracy of content push is improved.

In some embodiments, the method further comprises: acquiring a reference push record set, wherein the reference push record set comprises a plurality of reference push records, and the reference push records are content push records corresponding to the target feature combination; acquiring a content push value corresponding to the reference push record; counting content push values corresponding to the reference push records in the reference push record set to obtain target push values corresponding to the target feature combinations; and screening candidate contents to be pushed according to the target feature rejection degree and the target pushing value corresponding to the target feature combination.

Wherein, the reference push record is the content push record corresponding to the target feature combination. And the push object in the reference push record has the object characteristics in the target characteristic combination, and the push content in the reference push record has the content characteristics in the target characteristic combination. For example, the target feature combination 1 is (Ua1, Ub1), the reference push record corresponding to the target feature combination 1 is "push content S1 to push object 1", the push object 1 has the object feature of Ua1, and the push content S1 has the content feature of Ub 1. The content push record in the reference push record set may be the same as or different from the content push record in the content push record set.

The content push value is used for reflecting the income brought by the delivered advertisements to content owners, the content owners refer to the content owners, for example, when the content is the advertisements, the content owners can be advertisers, and the advertisers refer to entities displaying the advertisements through advertisement positions of the internet platform. For example, when the push content is a commodity advertisement, the content push value may be an amount of money spent by the user to purchase the commodity through the commodity advertisement, each reference push record may correspond to the content push value, and the content push values corresponding to different reference push records may be the same or different. The target push value is a value obtained by counting the content push values respectively corresponding to the reference push records.

Specifically, the server may obtain content push values corresponding to respective reference push records in the reference push record set, and perform statistical operation on the respective content push values, where the statistical operation may include at least one of summation operation, mean operation, median operation, or mode operation, for example, the server may perform summation operation on the respective content push values, and use a result of the summation operation as a target push value corresponding to the target feature combination.

In some embodiments, the reference set of push records may be content push records within the target time period. For example, the server may perform content push to the push object having the object feature in the target feature combination within the target time period, and form a reference push record set by content push records corresponding to the push object having the object feature in the target feature combination in the target time period.

In this embodiment, candidate contents to be pushed are screened based on the target feature rejection and the target pushing value corresponding to the target feature combination, so that the contents to be pushed can be screened by integrating the feature rejection and the pushing value, the content screening accuracy is improved, and the content pushing accuracy is improved.

In some embodiments, the target feature is combined into a plurality, the method further comprising: selecting a target feature combination meeting rejection degree screening conditions and pushing value screening conditions as a rejection feature combination according to the target feature rejection degree and the target pushing value corresponding to the target feature combination; the exclusion feature combination is used for filtering the push content with the content features in the exclusion feature combination when the push object with the object features in the exclusion feature combination is subjected to content push; the push value screening condition comprises that the target push value is larger than the push value threshold value or the target push value is sequenced before the value sequencing threshold value; the repulsion degree screening condition includes at least one of the target feature repulsion degree being greater than the repulsion degree threshold or the target repulsion degree ranking being before the repulsion degree ranking threshold.

The rejection feature combination may be a target feature combination that satisfies both the rejection degree screening condition and the push value screening condition.

The size of the push value threshold may be predetermined or determined according to the target push value corresponding to each target feature combination, for example, the server may perform a mean operation on the target push values corresponding to each target feature combination to obtain a push value mean, obtain a preset push value coefficient, perform a multiplication operation on the push value coefficient and the push value mean, and use the result of the multiplication operation as the push value threshold, where the push value coefficient may be a positive number smaller than 1.

The target push value sequence refers to the sequence of the target push values in a push value sequence, and the push value sequence is a sequence obtained by arranging the target push values respectively corresponding to each target feature combination from small to large. The smaller the target push value, the more advanced the ordering in the sequence of push values.

The value sorting threshold may be set in advance, for example, may be a fixed value such as 5 or 10, or may be determined according to the number of target feature combinations, for example, the server may acquire the second sorting coefficient, multiply the second sorting coefficient by the number of target feature combinations, and use the result of the multiplication as the value sorting threshold. The second ordering coefficient may be the same as the first ordering coefficient, or the second ordering coefficient may be different from the first ordering coefficient. The second ordering coefficient may be a positive number less than 1, for example, may be 0.3 or 0.5, etc.

Specifically, the server may obtain a target feature combination satisfying the rejection degree screening condition from each target feature combination to form a first feature combination set, obtain a target feature combination satisfying the push value screening condition from each target feature combination to form a second feature combination set, perform intersection operation on the first feature combination set and the second feature combination set, use a set obtained by the intersection operation as a third feature combination set, and use each target feature combination in the third feature combination set as a rejection feature combination.

As shown in fig. 4, GPM effect and distribution of repulsion of academic education advertisements on different academic people are shown, wherein the academic education is the type of advertisement. GPM (Cost Per mill, thousand Cost) is the revenue received from a thousand exposures, which may also be referred to as "exposure charge by thousand", i.e., the advertisement is charged by exposure, with each thousand exposures as a settlement unit. The solid rectangle box represents GPM and the curve represents the degree of repulsion, as can be seen from the figure, the population of the doctor's scholars has a high degree of repulsion to the educational advertisements of the scholars, and the population of the doctor's scholars has a low GPM to the educational advertisements of the scholars. Therefore, when the advertisement is pushed to the crowd of the doctor's academic calendar, the educational advertisement of the doctor's academic calendar can be filtered from the candidate content to be pushed corresponding to the crowd of the doctor's academic calendar, and the pushing accuracy is improved.

In some embodiments, the GPM may be calculated using equation (1). Where GMV (ai, bj) represents the total amount of deals made by users with ai user characteristics to an advertisement with bj advertisement characteristics over a period of time, and expose _ num is the number of exposures. GPM (ai, bj) represents the revenue that an ad with ad feature bj would have been exposed a thousand times to a user with user feature ai.

In this embodiment, according to the rejection degree of the target feature and the target push value corresponding to the target feature combination, the target feature combination satisfying the rejection degree screening condition and the push value screening condition is selected as the rejection feature combination, since the push value screening condition includes at least one of the target push value being greater than the push value threshold or the target push value ranking being before the value ranking threshold, and the repulsion screening condition includes at least one of the target feature repulsion being greater than the repulsion threshold or the target repulsion ranking being before the repulsion ranking threshold, therefore, a combination of target features with a greater degree of rejection of the target features and a smaller target push value may be used as a combination of rejection features, therefore, when the content is pushed, the rejection degree is high, the pushed content with low value is filtered, and the accuracy of content pushing is improved.

In some embodiments, the counting content push values corresponding to reference push records in the reference push record set to obtain a target push value corresponding to a target feature combination includes: summing and calculating content push values corresponding to the reference push records in the reference push record set to obtain a total content push value; acquiring the number of reference push records in a reference push record set as the number of reference content records; and calculating the ratio of the total content push value to the number of the reference content records, and obtaining a target push value corresponding to the target feature combination based on the calculated ratio.

Wherein, the reference content record number refers to the number of content push records included in the reference push record set.

Specifically, the server may perform a summation operation on the content push values corresponding to the respective reference push records, and use a result of the summation operation as the overall content push value.

In some embodiments, the server may calculate a ratio of the total content push value to the number of reference content records as a unit content push value corresponding to the target feature combination, where the target push value corresponding to the target feature combination has a positive correlation with the unit content push value. For example, the server may obtain a preset number of content records, perform a product operation on the preset number of content records and the unit content push value, and use a result of the operation as the target push value, where the preset number of content records may be a preset fixed number, such as 1000.

In this embodiment, the ratio of the total content push value to the number of reference content records is calculated, and the target push value corresponding to the target feature combination is obtained based on the calculated ratio, so that the accuracy of the target push value is improved.

In some embodiments, determining the target feature rejection degree corresponding to the target feature combination according to the first negative-going record number and the second negative-going record number includes: determining the number of content push records corresponding to the content push record set; calculating the ratio of the first negative direction record quantity to the content push record quantity, and obtaining a first negative direction record possibility according to the calculated ratio; calculating the ratio of the second negative direction record quantity to the first negative direction record quantity, and obtaining a second negative direction record possibility according to the calculated ratio; and calculating the target feature repulsion corresponding to the target feature combination based on the first negative direction record possibility and the second negative direction record possibility.

The content push record number refers to the number of content push records included in the content push record set. The first negative-going record likelihood is derived from a ratio of the first negative-going record number to the content-push record number, for example, the ratio of the first negative-going record number to the content-push record number may have a positive correlation with the first negative-going record likelihood. The second negative-going recording possibility is obtained based on a ratio of the second negative-going recording number to the first negative-going recording number, for example, the ratio of the second negative-going recording number to the first negative-going recording number may have a positive correlation with the second negative-going recording possibility. The target feature rejections are positively correlated with the first negative-going record likelihood and the target feature rejections are positively correlated with the second negative-going record likelihood.

Specifically, the server may calculate a ratio of the first negative direction record number to the content push record number, and use the calculated ratio as the first negative direction record possibility. The server may calculate a ratio of the second negative-going record number to the first negative-going record number, and use the calculated ratio as the second negative-going record possibility degree.

In some embodiments, the server may calculate the first negative-going record likelihood and the second negative-going record likelihood, and obtain the target feature rejections corresponding to the target feature combination based on the result obtained by the calculation, for example, the server may perform a product operation on the first negative-going record likelihood and the second negative-going record likelihood, take a result of the product operation as a standard negative-going record likelihood, obtain the target feature rejections corresponding to the target feature combination based on the standard negative-going record likelihood, and have a positive correlation with the standard negative-going record likelihood.

In this embodiment, since the first negative direction record possibility is obtained according to a ratio of the first negative direction record number to the content push record number, and the second negative direction record possibility is obtained according to a ratio of the second negative direction record number to the first negative direction record number, the target feature rejections corresponding to the target feature combination are obtained based on the first negative direction record possibility and the second negative direction record possibility, so that accuracy of the target feature rejections is improved.

In some embodiments, calculating the target feature rejection based on the first negative-going record likelihood and the second negative-going record likelihood to obtain the target feature combination comprises: calculating the product of the first negative direction record possibility and the second negative direction record possibility to obtain a standard negative direction record possibility corresponding to the target feature combination; acquiring the number of content push records corresponding to the target feature combination in the content push record set as the number of combined records; calculating the ratio of the number of the combined records to the number of the content push records, and obtaining the push record possibility corresponding to the target feature combination according to the calculated ratio; and obtaining the target feature repulsion corresponding to the target feature combination based on the standard negative direction record possibility and the push record possibility.

Wherein the standard negative-going recording possibility is a product of the first negative-going recording possibility and the second negative-going recording possibility. The content features in the content push records corresponding to the target feature combinations are consistent with the content features in the target feature combinations, and the object features of the push objects in the content push records corresponding to the target feature combinations are consistent with the object features in the target feature combinations. The number of combined records refers to the number of content push records corresponding to the target feature combination included in the content push record set.

The push record possibility is calculated according to a ratio of the number of the combined records to the number of the content push records, for example, the ratio of the number of the combined records to the number of the content push records has a positive correlation with the push record possibility.

Specifically, the server may compare an object feature of a push object in content push records in the content push record set with an object feature in the target feature combination to obtain an object feature comparison result, compare a content feature of push content in the content push records with a content feature in the target feature combination to obtain a content feature comparison result, and use a content push record whose object feature comparison result is consistent with the content feature comparison result and whose content feature comparison result is consistent with the target feature combination as the content push record corresponding to the target feature combination.

In some embodiments, the server may calculate a ratio of the number of combined records to the number of content push records, use the calculated ratio as a push record likelihood corresponding to the target feature combination, calculate a ratio of a standard negative direction record likelihood to the push record likelihood, and use the calculated ratio as a target feature rejection corresponding to the target feature combination.

In this embodiment, the ratio of the number of combined records to the number of content push records is calculated, the push record possibility corresponding to the target feature combination is obtained according to the calculated ratio, the target feature rejections corresponding to the target feature combination are obtained based on the standard negative direction record possibility and the push record possibility, and the accuracy of the target feature rejections is improved.

In some embodiments, obtaining the target feature combination comprises: sending a feature combination configuration interface to a configuration terminal so that the configuration terminal can display the feature combination configuration interface, and responding to configuration operation aiming at the feature combination configuration interface to obtain a configured target feature combination; and receiving the target characteristic combination sent by the configuration terminal.

The configuration terminal is an interface for configuring the target feature combination. The configuration terminal may be, for example, terminal 102 in fig. 1. The configuration operation refers to an operation of configuring a target feature combination in the feature combination configuration interface. The feature combination configuration interface may display a feature combination editing region, the feature combination editing region may receive a feature combination input or selected by a user, and the feature combination configuration interface may further display a configuration confirmation control, where the configuration confirmation control is used to instruct to perform configuration of a feature combination. The configuration operation may be, for example, a trigger operation to a configuration confirmation control.

Specifically, the configuration terminal may present an entry of the feature combination configuration interface, and when the configuration terminal receives a trigger operation on the entry of the feature combination configuration interface, the configuration terminal may send a feature combination configuration request to the server, and the server may return the feature combination configuration interface to the configuration terminal in response to the feature combination configuration request. The configuration terminal can display a feature combination configuration interface, when receiving triggering operation on the configuration confirmation control, the configuration terminal can acquire a feature combination input or selected in the feature combination editing area, uses the acquired feature combination as a target feature combination, and sends the target feature combination to the server.

In this embodiment, a feature combination configuration interface is sent to a configuration terminal, so that the configuration terminal displays the feature combination configuration interface, obtains a configured target feature combination in response to a configuration operation for the feature combination configuration interface, and receives the target feature combination sent by the configuration terminal, so that a feature combination with a possibly high feature rejection degree or a possibly low content push value can be configured as the target feature combination through the feature combination configuration interface, thereby reducing the number of the target feature combinations, reducing the computation complexity, and improving the computation efficiency.

In some embodiments, as shown in fig. 5, a content pushing method is provided, where the method may be executed by a terminal or a server, or may be executed by both the terminal and the server, and in this embodiment, the method is described as applied to the server 104 in fig. 1, and includes the following steps: s502, receiving a content push request aiming at a target push object; s504, responding to the content pushing request, and acquiring a candidate pushing content set corresponding to a target pushing object; s506, determining a target rejection characteristic combination based on the target object characteristics corresponding to the target push object; the target rejection feature combination is determined based on the rejection degree of the target features corresponding to the target feature combination; s508, filtering the candidate push contents corresponding to the target rejection characteristic combination from the candidate push content set to obtain the target push contents corresponding to the target push object.

The target push object can be any push object, and when the terminal accesses an interface corresponding to the client corresponding to the server and the interface includes a push content display area, the server can take a user corresponding to the terminal as the target push object. The candidate push content set may be preset or may be set according to needs, for example, a content with a higher popularity may be used as the candidate push content, or the candidate push content may be determined according to the needs of the advertiser. The candidate push content sets corresponding to different target push objects may be the same or different.

Specifically, the server may store a plurality of preset push contents, and the server may obtain push content scores corresponding to the preset push contents, select candidate push contents from the plurality of preset push contents based on the push content scores, and combine the candidate push contents into a candidate push content set. For example, the server may obtain, from a plurality of preset push contents, a preset push content with a push content score greater than a content score threshold as a candidate push content to form a candidate push content set. The content score threshold may be set as needed or preset.

In some embodiments, the pushed content score comprises a first score, and the step of deriving the first score may comprise: the method comprises the steps of obtaining an object portrait corresponding to a target pushing object, determining an object type corresponding to the target pushing object based on the object portrait as a target object type, wherein different object types correspond to different object portraits, obtaining a predicted click possibility and a predicted conversion possibility of the pushing object of the target object type aiming at preset pushing content, carrying out product operation on the predicted click possibility and the predicted conversion possibility, taking the result of the product operation as the target conversion possibility, obtaining unit conversion income, and carrying out product operation on the target conversion possibility and the unit conversion income to obtain a first score. The product of the target conversion likelihood and the unit conversion yield is in positive correlation with the first score, and for example, the product of the target conversion likelihood and the unit conversion yield may be used as the first score.

The Click probability refers to the probability of the content being clicked, and may also be referred to as Click Through Rate (CTR), and a ratio of the actual number of clicks of the content to the total number of exposures may be used as the Click probability. The predicted click likelihood refers to a predicted click likelihood. The predicted click likelihood may also be referred to as a predicted click through rate (pCTR). Conversion Rate (CVR) refers to the Conversion Rate of a user clicking on an advertisement to a valid enabled, registered or paid user, i.e. the actual number of conversions of the advertisement divided by the number of clicks on the advertisement. The estimated conversion (pCVR) is the estimated conversion. The revenue per conversion refers to the cost that the content owner needs to pay to successfully convert a conversion, which may be, for example, a bid (bid), which refers to the price that the advertiser offers for a conversion, i.e., the cost the advertiser pays the traffic owner after a successful conversion.

In some embodiments, the server may calculate the first score using equation (2). Where eCPM (effective cost per minute, revenue that may be obtained per thousand shows) is the first score.

eCPM＝bid*pCTR*pCVR*1000 (2)

In some embodiments, the push content score further includes a second score, and the derived second score may be, for example, a quality score, which may be affected by negative feedback from the user in the social advertisement, social liveness of the user, and the like. The server may calculate the first score and the second score to be summed, and the result of the summation is used as the push content score, for example, the server may calculate the push content score by using formula (3). Where Ranking _ score represents the push content score and quality represents the quality score, i.e., the second score.

Ranking_score＝ eCPM+ quality (3)

In the content push method, a content push request for a target push object is received, a candidate push content set corresponding to the target push object is obtained in response to the content push request, a target rejection characteristic combination is determined based on target feature rejection corresponding to the target push object, the target rejection characteristic combination is determined based on target feature rejection corresponding to the target feature combination, the target feature rejection reflects the rejection degree of an object with object features in the target feature combination to content with content features in the target feature combination, the larger the target feature rejection, the more the object rejects the content, and thus the target feature combination with the larger target feature rejection can be used as the target rejection feature combination, so as to filter the candidate push content corresponding to the target rejection feature combination from the candidate push content set, the candidate push contents rejected by the target push object can be filtered, so that the possibility of pushing the rejected contents to the target push object is reduced, and the accuracy of content pushing is improved.

In some embodiments, determining the target rejection feature combination based on the target object feature corresponding to the target push object comprises: acquiring a plurality of candidate rejection feature combinations; the candidate rejection feature combinations are obtained by screening from each target feature combination based on the rejection degree of the target features corresponding to the target feature combinations; and comparing the target object characteristics corresponding to the target push object with the object characteristics corresponding to the push object in the candidate exclusion characteristic combination, and taking the exclusion characteristic combination with consistent comparison as a target exclusion characteristic combination.

Specifically, the server may obtain a plurality of target feature combinations, determine a target feature rejection degree corresponding to each target feature combination, and obtain, from each target feature combination, a target feature combination satisfying a rejection degree screening condition based on the target feature rejection degree, as a candidate rejection feature combination.

In some embodiments, the server may determine target push values corresponding to respective target feature combinations, and obtain, from the respective target feature combinations, a target feature combination satisfying a rejection degree screening condition and a push value screening condition as a candidate rejection feature combination based on the target feature rejection degree and the target push values.

In this embodiment, the target object feature corresponding to the target push object is compared with the object feature corresponding to the push object in the candidate exclusion feature combination, and the exclusion feature combination with the same comparison is used as the target exclusion feature combination, so that the target exclusion feature combination is determined quickly.

In some embodiments, the step of obtaining a candidate combination of exclusion features comprises: acquiring a plurality of target feature combinations; the target characteristic combination is obtained by combining object characteristics and content characteristics; acquiring a content push record set, wherein the content push record set comprises a plurality of content push records, and the content push records comprise push contents and push objects corresponding to the push contents; determining a first negative-going content push record of a push object performing negative-going operation on push content in a content push record set, and acquiring a first negative-going record number corresponding to the first negative-going content push record; determining a second negative-going content push record corresponding to the target feature combination, and acquiring a second negative-going record number corresponding to the second negative-going content push record; the content characteristics of the push content in the second negative-going content push record are consistent with the content characteristics in the target characteristic combination, and the object characteristics of the push object in the second negative-going content push record are consistent with the object characteristics in the target characteristic combination; determining a target feature rejection degree corresponding to the target feature combination according to the first negative direction record quantity and the second negative direction record quantity, wherein the target feature rejection degree represents the rejection degree between the content features and the object features in the target feature combination; and screening candidate rejection feature combinations from each target feature combination based on the target feature rejection degrees corresponding to the target feature combinations.

Specifically, the server may screen each target feature combination to obtain a target feature combination satisfying the rejection degree screening condition and the push value screening condition, and use the target feature combination as a candidate rejection feature combination.

In this embodiment, a first negative-going content push record in which a push object performs negative-going operations on a push content in a content push record set is determined, a first negative-going record number corresponding to the first negative-going content push record is obtained, so that the first negative-going record number reflects the number of content push records performing negative-going operations on the push content, a second negative-going content push record corresponding to a target feature combination is determined, a second negative-going record number corresponding to the second negative-going content push record is obtained, because content features of push content in the second negative-going content push record are consistent with content features in the target feature combination, object features of the push object in the second negative-going content push record are consistent with object features in the target feature combination, so that the second negative-going record number reflects the number of content push records corresponding to the target feature combination in each content push record performing negative-going operations, determining a target feature rejection degree corresponding to the target feature combination according to the first negative direction record quantity and the second negative direction record quantity, wherein the target feature rejection degree represents a rejection degree between the content features and the object features in the target feature combination, so that the target feature rejection degree reflects a rejection degree of the object having the object features in the target feature combination to the content having the content features in the target feature combination, and the rejection degree is larger as the target feature rejection degree is larger, so that a candidate rejection feature combination is screened from each target feature combination based on the target feature rejection degree corresponding to the target feature combination, the target feature combination with the larger target feature rejection degree can be used as the candidate rejection feature combination, and thus when a candidate pushed content corresponding to the target rejection feature combination is filtered from the candidate pushed content set, a target pushed content corresponding to the target pushed object is obtained, candidate push contents with larger rejection degree of the target push object can be filtered from the candidate push content set, so that the possibility of pushing the rejected contents to the target push object is reduced, and the accuracy of content pushing is improved.

In some embodiments, a content push method is provided, comprising the steps of:

1. and the server sends the feature combination configuration interface to the configuration terminal.

The configuration terminal can display the feature combination configuration interface, respond to configuration operation aiming at the feature combination configuration interface, obtain the configured feature combination as a target feature combination, and send the target feature combination to the server. The feature combination is a combination of push user features and advertisement features.

As shown in fig. 6, the feature may be configured at the configuration terminal according to expert knowledge to obtain a target feature combination, and specifically, the configuration terminal may obtain a feature combination configured by the configuration user based on the expert knowledge, use the configured feature combination as the target feature combination, and send the target feature combination to the server.

2. And the server receives the target characteristic combination sent by the configuration terminal.

The target feature combination may be plural.

3. Acquiring an advertisement push record set; the advertisement push record set comprises a plurality of advertisement push records, and the advertisement push records comprise push advertisements and push users to which the push advertisements are pushed.

Wherein the advertisement push record in the advertisement push record set can be a record generated by pushing the advertisement from a historical time period. Or determining a push advertisement set and a push user set based on the target feature combination, and pushing the push advertisement in the push advertisement set to the push user in the push user set. The push advertisement set comprises a plurality of push advertisements, and the push user set comprises a plurality of push users. The pushed advertisement set comprises the pushed advertisement with the advertisement characteristics in the target characteristic combination. The push user set comprises push users with the push user characteristics in the target characteristic combination. The server may generate an advertisement push record for each push, for example, the server may log the advertisement pushed each time and the user to which the advertisement is pushed, and the server may generate the advertisement push record based on the log. The advertisement push record may further include an operation type of the user for pushing the advertisement, where the operation type may be any one of a positive feedback behavior and a negative feedback behavior. For example, negative feedback behavior may be represented by C1 and positive feedback behavior by C2.

In some embodiments, the number of push advertisements included in the set of push advertisements may be the same as the number of push users in the set of push users, each push user pushing one push advertisement at the time of push. At the time of pushing, a pushed advertisement having an advertisement feature in the targeted combination of features may be pushed to a pushing user having a pushing user feature in the targeted combination of features.

4. Determining a first negative advertisement push record of a push user performing a negative feedback action on a pushed advertisement in an advertisement push record set, and acquiring a first negative record number corresponding to the first negative advertisement push record.

The server may use the advertisement push records whose operation types are negative feedback behaviors as first negative advertisement push records, count the number of each first negative advertisement push record, and use the counted number as the number of the first negative records.

5. And determining a second negative direction advertisement push record corresponding to the target feature combination, and acquiring a second negative direction record quantity corresponding to the second negative direction advertisement push record.

The advertisement characteristics of the pushed advertisements in the first negative advertisement pushing record are the same as the advertisement characteristics in the target characteristic combination, and the user characteristics of the pushing users in the second negative advertisement pushing record are the same as the user characteristics in the target characteristic combination.

6. Determining the number of advertisement push records corresponding to an advertisement push record set, calculating the ratio of the number of first negative direction records to the number of advertisement push records, taking the calculated ratio as a first negative direction record possibility, calculating the ratio of a second negative direction record number to the first negative direction record number, taking the calculated ratio as a second negative direction record possibility, calculating the product of the first negative direction record possibility and the second negative direction record possibility, taking the calculated result as a standard negative direction record possibility corresponding to a target feature combination, obtaining the number of advertisement push records corresponding to the target feature combination in the advertisement push record set, taking the number of advertisement push records as a combined record number, calculating the ratio of the combined record number to the number of advertisement push records, taking the calculated ratio as a push record possibility corresponding to the target feature combination, calculating the ratio of the standard negative direction record possibility to the push record possibility, and taking the calculated ratio as the target feature rejection degree corresponding to the target feature combination.

7. The method comprises the steps of obtaining a reference advertisement push record set corresponding to a target characteristic combination, obtaining an advertisement push value corresponding to the reference advertisement push record, summing the advertisement push values corresponding to all the reference advertisement push records in the reference advertisement push record set respectively, taking the result of the summing operation as a total advertisement push value corresponding to the target characteristic combination, obtaining the number of the reference advertisement push records in the reference advertisement push records as the number of the reference advertisement records, calculating the ratio of the total advertisement push value to the number of the reference advertisement records, taking the calculated ratio as a unit advertisement push value, obtaining the number of preset advertisement records, carrying out product operation on the unit advertisement push value and the number of the preset advertisement records, and taking the result of the product operation as the target push value corresponding to the target characteristic combination.

And the pushed advertisement in the reference advertisement pushing record has the advertisement characteristics in the target characteristic combination.

8. And selecting target feature combinations meeting rejection degree screening conditions and pushing value screening conditions from the target feature combinations according to the target feature rejection degrees and the target pushing values corresponding to the target feature combinations, taking the target feature combinations as rejection feature combinations, and storing the rejection feature combinations.

The push value screening condition comprises at least one of the target push value being greater than the push value threshold or the target push value ranking being before the value ranking threshold; the repulsion degree screening condition includes at least one of the target feature repulsion degree being greater than the repulsion degree threshold or the target repulsion degree ranking being before the repulsion degree ranking threshold. For example, the rejection degree screening condition corresponding to the target feature combination (ai, bj) may be expressed as p (c1/(ai, bj)) > δ, where δ represents a rejection degree threshold. C1 indicates that the operation type is negative feedback behavior. The push value screening condition corresponding to the target feature combination (ai, bj) can be expressed as GPM (ai, bj) < epsilon, wherein epsilon represents a push value threshold. The target push value may also be referred to as a distribution value.

In some embodiments, the server may format the exclusion feature combination to obtain a formatted exclusion feature combination, and store the formatted exclusion feature combination. As shown in fig. 6, the server may store the formatted combination of exclusion characteristics in a common sense repository. The formatted combination of exclusion features is understandable by the program, in that the formatted combination of exclusion features has a format understandable by the program. The common sense knowledge base is constructed to support two major categories of dimensions, wherein one category is an advertising dimension and mainly supports pluggable advertising range selection including but not limited to at least one of advertising industry, advertisers or advertising groups and the like, and the other category is a user dimension including but not limited to at least one of user age, academic calendar or love state and the like. The common sense repository may be located in the refinement engine, for example, the common sense repository may be configured in a database that can be directly read by the refinement engine, and the database stores the types of data results that can be directly read and understood by the refinement engine.

9. Receiving an advertisement pushing request aiming at a target pushing user, and responding to the advertisement pushing request to obtain a candidate pushing advertisement set corresponding to the target pushing user.

10. And acquiring the user characteristics corresponding to the target pushing user, comparing the user characteristics of the target pushing user with the user characteristics in the rejection characteristic combination, and taking the rejection characteristic combination with consistent comparison as the target rejection characteristic combination.

Wherein, the target exclusion feature combination may be plural.

11. And filtering out advertisements with the advertisement characteristics in the target exclusion characteristic combination from the candidate pushed advertisement set, and taking the remaining candidate pushed advertisements after filtering in the candidate pushed advertisement set as target pushed advertisements corresponding to the target pushed users.

The fine queue corresponding to the target push user can be generated in the fine queue engine. The server may store each candidate push advertisement in the set of candidate push advertisements in a fine queue, and the server may filter out candidate push advertisements from the fine queue that have the advertisement characteristics in the target rejection characteristic combination. For example, assuming that the target exclusion feature combination is (a1, b2), the advertisement with advertisement feature b2 is removed from the fine queue, and the remaining advertisements in the fine queue are used as the target push advertisements corresponding to the target push users.

12. And pushing the target push advertisement to the target push user.

According to the content push method, the target exclusion characteristic combination is obtained, and the advertisements excluded by the user can be predicted, so that the advertisements excluded by the user can be filtered from the advertisements needing to be pushed when the advertisements are pushed to the user, the condition of pushing the excluded advertisements to the user is reduced, and the accuracy of pushing the advertisements is improved.

It should be understood that although the various steps in the flowcharts of fig. 2-6 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 2-6 may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed in turn or alternately with other steps or at least some of the other steps.

In some embodiments, as shown in fig. 7, a content pushing apparatus is provided, which may be a part of a computer device using a software module or a hardware module, or a combination of the two, and specifically includes: a target feature combination obtaining module 702, a content push record set obtaining module 704, a first negative record number obtaining module 706, a second negative record number obtaining module 708, and a target feature exclusion degree determining module 710, where:

a target feature combination obtaining module 702, configured to obtain a target feature combination, where the target feature combination is obtained by combining object features and content features;

a content push record set obtaining module 704, configured to obtain a content push record set, where the content push record set includes multiple content push records, and each content push record includes a push content and a push object corresponding to the push content;

a first negative direction record number obtaining module 706, configured to determine a first negative direction content push record in which a push object performs a negative direction operation on a push content in a content push record set, and obtain a first negative direction record number corresponding to the first negative direction content push record;

a second negative-going record number obtaining module 708, configured to determine a second negative-going content push record corresponding to the target feature combination, and obtain a second negative-going record number corresponding to the second negative-going content push record; the content characteristics of the push content in the second negative-going content push record are consistent with the content characteristics in the target characteristic combination, and the object characteristics of the push object in the second negative-going content push record are consistent with the object characteristics in the target characteristic combination;

the target feature exclusion degree determining module 710 is configured to determine a target feature exclusion degree corresponding to the target feature combination according to the first negative direction record number and the second negative direction record number, where the target feature exclusion degree indicates a degree of exclusion between a content feature in the target feature combination and an object feature, and screen candidate content to be pushed based on the target feature exclusion degree corresponding to the target feature combination.

In some embodiments, the target feature combinations are multiple, and the apparatus is further configured to select, as the rejection feature combination, a target feature combination that meets a rejection degree screening condition according to a target feature rejection degree corresponding to the target feature combination, where the rejection feature combination is used to filter, when content pushing is performed on a push object having object features in the rejection feature combination, push content having content features in the rejection feature combination; the repulsion degree screening condition includes at least one of the target feature repulsion degree being greater than the repulsion degree threshold or the target repulsion degree ranking being before the repulsion degree ranking threshold.

In some embodiments, the apparatus further comprises: a reference push record set obtaining module, configured to obtain a reference push record set, where the reference push record set includes multiple reference push records, and the reference push records are content push records corresponding to a target feature combination; the content push value acquisition module is used for acquiring a content push value corresponding to the reference push record; the target push value obtaining module is used for counting the content push values corresponding to the reference push records in the reference push record set to obtain the target push value corresponding to the target feature combination; and screening candidate contents to be pushed according to the target feature rejection degree and the target pushing value corresponding to the target feature combination.

In some embodiments, the target feature combinations are multiple, and the device is further configured to select, as the rejection feature combination, a target feature combination that satisfies rejection degree screening conditions and push value screening conditions according to the target feature rejection degree and the target push value corresponding to the target feature combination; the exclusion feature combination is used for filtering the push content with the content features in the exclusion feature combination when the push object with the object features in the exclusion feature combination is subjected to content push; the push value screening condition comprises that the target push value is larger than the push value threshold value or the target push value is sequenced before the value sequencing threshold value; the repulsion degree screening condition includes at least one of the target feature repulsion degree being greater than the repulsion degree threshold or the target repulsion degree ranking being before the repulsion degree ranking threshold.

In some embodiments, the target push value derivation module comprises: the overall content push value obtaining unit is used for summing the content push values corresponding to the reference push records in the reference push record set to obtain the overall content push value; a reference content record number obtaining unit, configured to obtain a number of reference push records in a reference push record set as a reference content record number; and the target push value obtaining unit is used for calculating the ratio of the total content push value to the reference content record number and obtaining the target push value corresponding to the target feature combination based on the calculated ratio.

In some embodiments, the target feature rejections determination module 710 includes: the content push record number determining unit is used for determining the content push record number corresponding to the content push record set; a first negative direction record possibility obtaining unit, configured to calculate a ratio of the first negative direction record number to the content push record number, and obtain a first negative direction record possibility according to the calculated ratio; a second negative-direction record possibility degree obtaining unit configured to calculate a ratio of the second negative-direction record number to the first negative-direction record number, and obtain a second negative-direction record possibility degree according to the calculated ratio; and the target feature exclusion degree obtaining unit is used for obtaining the target feature exclusion degree corresponding to the target feature combination through calculation based on the first negative direction record possibility degree and the second negative direction record possibility degree.

In some embodiments, the target feature exclusion degree obtaining unit is further configured to calculate a product of the first negative direction record possibility degree and the second negative direction record possibility degree to obtain a standard negative direction record possibility degree corresponding to the target feature combination; acquiring the number of content push records corresponding to the target feature combination in the content push record set as the number of combined records; calculating the ratio of the number of the combined records to the number of the content push records, and obtaining the push record possibility corresponding to the target feature combination according to the calculated ratio; and obtaining the target feature repulsion corresponding to the target feature combination based on the standard negative direction record possibility and the push record possibility.

In some embodiments, the target feature combination acquisition module 702 includes: the device comprises a feature combination configuration interface sending unit, a feature combination configuration interface sending unit and a feature combination configuration interface sending unit, wherein the feature combination configuration interface sending unit is used for sending a feature combination configuration interface to a configuration terminal so that the configuration terminal can display the feature combination configuration interface and respond to configuration operation aiming at the feature combination configuration interface to obtain a configured target feature combination; and the target characteristic combination receiving unit is used for receiving the target characteristic combination sent by the configuration terminal.

In some embodiments, as shown in fig. 8, there is provided a content pushing apparatus, which may be a part of a computer device using a software module or a hardware module, or a combination of the two, and specifically includes: a content push request receiving module 802, a candidate push content set obtaining module 804, a target rejection characteristic combination determining module 806, and a target push content obtaining module 808, where:

a content push request receiving module 802, configured to receive a content push request for a target push object;

a candidate push content set obtaining module 804, configured to, in response to the content push request, obtain a candidate push content set corresponding to the target push object;

a target exclusion feature combination determining module 806, configured to determine a target exclusion feature combination based on a target object feature corresponding to the target push object, where the target exclusion feature combination is determined based on a target feature exclusion degree corresponding to the target feature combination;

a target push content obtaining module 808, configured to filter candidate push content corresponding to the target exclusion feature combination from the candidate push content set, so as to obtain target push content corresponding to the target push object.

In some embodiments, the target-exclusive feature combination determination module 806 includes:

a first candidate repulsive feature combination obtaining unit configured to obtain a plurality of candidate repulsive feature combinations; the candidate rejection feature combinations are obtained by screening from each target feature combination based on the rejection degree of the target features corresponding to the target feature combinations;

and the target rejection characteristic combination obtaining unit is used for comparing the target object characteristics corresponding to the target push object with the object characteristics corresponding to the push objects in the candidate rejection characteristic combination, and taking the rejection characteristic combination with consistent comparison as the target rejection characteristic combination. In some embodiments, the apparatus further comprises a candidate repulsive feature combination derivation module, which comprises:

a target feature combination obtaining unit configured to obtain a plurality of target feature combinations; the target characteristic combination is obtained by combining object characteristics and content characteristics;

the content push record set acquisition unit is used for acquiring a content push record set, wherein the content push record set comprises a plurality of content push records, and each content push record comprises push content and a push object corresponding to the push content;

a first negative direction record number obtaining unit, configured to determine a first negative direction content push record in which a push object performs a negative direction operation on a push content in a content push record set, and obtain a first negative direction record number corresponding to the first negative direction content push record;

a second negative direction record number obtaining unit, configured to determine a second negative direction content push record corresponding to the target feature combination, and obtain a second negative direction record number corresponding to the second negative direction content push record; the content characteristics of the push content in the second negative-going content push record are consistent with the content characteristics in the target characteristic combination, and the object characteristics of the push object in the second negative-going content push record are consistent with the object characteristics in the target characteristic combination;

the target feature exclusion degree determining unit is used for determining a target feature exclusion degree corresponding to the target feature combination according to the first negative direction record quantity and the second negative direction record quantity, and the target feature exclusion degree represents the exclusion degree between the content features and the object features in the target feature combination;

and the second candidate rejection feature combination obtaining unit is used for screening and obtaining candidate rejection feature combinations from each target feature combination based on the target feature rejection degree corresponding to the target feature combination.

For specific limitations of the content push device, reference may be made to the above limitations of the content push method, which is not described herein again. The modules in the content pushing apparatus may be implemented in whole or in part by software, hardware, and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.

In some embodiments, a computer device is provided, which may be a terminal, and its internal structure diagram may be as shown in fig. 9. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless communication can be realized through WIFI, an operator network, NFC (near field communication) or other technologies. The computer program is executed by a processor to implement a content push method. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.

In some embodiments, a computer device is provided, which may be a server, the internal structure of which may be as shown in fig. 10. The computer device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used for storing data involved in the content push method. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a content push method.

Those skilled in the art will appreciate that the configurations shown in fig. 9 and 10 are merely block diagrams of some configurations relevant to the present disclosure, and do not constitute a limitation on the computing devices to which the present disclosure may be applied, and that a particular computing device may include more or less components than those shown, or combine certain components, or have a different arrangement of components.

In some embodiments, there is further provided a computer device comprising a memory and a processor, the memory having stored therein a computer program, the processor implementing the steps of the above method embodiments when executing the computer program.

In some embodiments, a computer-readable storage medium is provided, in which a computer program is stored, which computer program, when being executed by a processor, carries out the steps of the above-mentioned method embodiments.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others.

The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.

The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.

Claims

1. A method for pushing content, the method comprising:

acquiring a target feature combination, wherein the target feature combination is obtained by combining object features and content features;

acquiring a content push record set, wherein the content push record set comprises a plurality of content push records, and each content push record comprises push content and a push object corresponding to the push content;

determining a first negative-going content push record of a push object performing negative-going operation on push content in the content push record set, and acquiring a first negative-going record number corresponding to the first negative-going content push record;

determining a second negative-going content push record corresponding to the target feature combination, and acquiring a second negative-going record quantity corresponding to the second negative-going content push record; wherein, the content features of the push content in the second negative-going content push record are consistent with the content features in the target feature combination, and the object features of the push object in the second negative-going content push record are consistent with the object features in the target feature combination;

and determining a target feature rejection degree corresponding to the target feature combination according to the first negative direction record quantity and the second negative direction record quantity, wherein the target feature rejection degree represents a rejection degree between the content features and the object features in the target feature combination, and screening candidate contents to be pushed based on the target feature rejection degree corresponding to the target feature combination.

2. The method of claim 1, wherein the target feature combination is a plurality, the method further comprising:

selecting a target feature combination meeting rejection degree screening conditions according to the target feature rejection degree corresponding to the target feature combination to serve as a rejection feature combination, wherein the rejection feature combination is used for filtering push contents with content features in the rejection feature combination when pushing the contents of the push objects with the object features in the rejection feature combination;

the repulsion degree screening condition includes at least one of the target feature repulsion degree being greater than the repulsion degree threshold or the target repulsion degree ranking being before the repulsion degree ranking threshold.

3. The method of claim 1, further comprising:

acquiring a reference push record set, wherein the reference push record set comprises a plurality of reference push records, and the reference push records are content push records corresponding to the target feature combination;

acquiring a content push value corresponding to the reference push record;

counting content push values corresponding to the reference push records in the reference push record set to obtain target push values corresponding to the target feature combinations; and screening candidate contents to be pushed according to the target feature rejection degree corresponding to the target feature combination and the target pushing value.

4. The method of claim 3, wherein the target feature combination is a plurality, the method further comprising:

selecting a target feature combination meeting rejection degree screening conditions and pushing value screening conditions as a rejection feature combination according to the target feature rejection degree and the target pushing value corresponding to the target feature combination; the exclusion characteristic combination is used for filtering the push content with the content characteristics in the exclusion characteristic combination when the push object with the object characteristics in the exclusion characteristic combination is subjected to content push;

the push value screening condition comprises at least one of the target push value being greater than the push value threshold or the target push value ranking being before the value ranking threshold; the repulsion degree screening condition includes at least one of the target feature repulsion degree being greater than the repulsion degree threshold or the target repulsion degree ranking being before the repulsion degree ranking threshold.

5. The method according to claim 3, wherein the counting content push values corresponding to the reference push records in the reference push record set to obtain a target push value corresponding to the target feature combination comprises:

summing and calculating content push values corresponding to the reference push records in the reference push record set to obtain a total content push value;

acquiring the number of reference push records in a reference push record set as the number of reference content records;

and calculating the ratio of the total content push value to the number of the reference content records, and obtaining a target push value corresponding to the target feature combination based on the calculated ratio.

6. The method of claim 1, wherein the determining the target feature rejections for the target feature combination according to the first number of negative-going records and the second number of negative-going records comprises:

determining the number of content push records corresponding to the content push record set;

calculating the ratio of the first negative direction record quantity to the content push record quantity, and obtaining a first negative direction record possibility according to the calculated ratio;

calculating the ratio of the second negative direction record quantity to the first negative direction record quantity, and obtaining a second negative direction record possibility according to the calculated ratio;

and calculating a target feature rejection degree corresponding to the target feature combination based on the first negative direction recording possibility degree and the second negative direction recording possibility degree.

7. The method of claim 6, wherein the calculating a target feature rejection based on the first negative recording likelihood and the second negative recording likelihood for the target feature combination comprises:

calculating the product of the first negative direction record possibility and the second negative direction record possibility to obtain a standard negative direction record possibility corresponding to the target feature combination;

acquiring the number of content push records corresponding to the target feature combination in the content push record set as the number of combined records;

calculating the ratio of the number of the combined records to the number of the content push records, and obtaining the push record possibility corresponding to the target feature combination according to the calculated ratio;

and obtaining a target feature rejection degree corresponding to the target feature combination based on the standard negative direction recording possibility degree and the push recording possibility degree.

8. The method of claim 1, wherein the obtaining a target feature combination comprises:

sending a feature combination configuration interface to a configuration terminal so that the configuration terminal displays the feature combination configuration interface, and responding to configuration operation aiming at the feature combination configuration interface to obtain a configured target feature combination;

and receiving the target characteristic combination sent by the configuration terminal.

9. A method for pushing content, the method comprising:

receiving a content push request aiming at a target push object;

responding to the content pushing request, and acquiring a candidate pushing content set corresponding to the target pushing object;

determining a target rejection characteristic combination based on the target object characteristics corresponding to the target push object, wherein the target rejection characteristic combination is determined based on the target characteristic rejection degree corresponding to the target characteristic combination;

and filtering the candidate push contents corresponding to the target rejection characteristic combination from the candidate push content set to obtain the target push contents corresponding to the target push object.

10. The method of claim 9, wherein determining a target exclusive feature combination based on the target object feature corresponding to the target push object comprises:

acquiring a plurality of candidate rejection feature combinations; the candidate rejection feature combinations are obtained by screening from each target feature combination based on the target feature rejection degrees corresponding to the target feature combinations;

and comparing the target object characteristics corresponding to the target push object with the object characteristics corresponding to the push objects in the candidate exclusion characteristic combination, and taking the exclusion characteristic combination with consistent comparison as a target exclusion characteristic combination.

11. The method of claim 10, wherein the step of obtaining the candidate repulsive feature combination comprises:

acquiring a plurality of target feature combinations; the target feature combination is obtained by combining object features and content features;

determining a target feature rejection degree corresponding to the target feature combination according to the first negative direction record quantity and the second negative direction record quantity, wherein the target feature rejection degree represents a rejection degree between the content features and the object features in the target feature combination;

and screening candidate rejection feature combinations from the target feature combinations based on the target feature rejection degrees corresponding to the target feature combinations.

12. A content pushing apparatus, characterized in that the apparatus comprises:

the target feature combination acquisition module is used for acquiring a target feature combination, wherein the target feature combination is obtained by combining object features and content features;

the content push record set acquisition module is used for acquiring a content push record set, wherein the content push record set comprises a plurality of content push records, and each content push record comprises push content and a push object corresponding to the push content;

a first negative direction record number obtaining module, configured to determine a first negative direction content push record in which a push object performs a negative direction operation on a push content in the content push record set, and obtain a first negative direction record number corresponding to the first negative direction content push record;

a second negative direction record number obtaining module, configured to determine a second negative direction content push record corresponding to the target feature combination, and obtain a second negative direction record number corresponding to the second negative direction content push record; wherein, the content features of the push content in the second negative-going content push record are consistent with the content features in the target feature combination, and the object features of the push object in the second negative-going content push record are consistent with the object features in the target feature combination;

a target feature exclusion degree determining module, configured to determine a target feature exclusion degree corresponding to the target feature combination according to the first negative direction record number and the second negative direction record number, where the target feature exclusion degree indicates a degree of exclusion between a content feature and an object feature in the target feature combination, and screen candidate content to be pushed based on the target feature exclusion degree corresponding to the target feature combination.

13. A content pushing apparatus, characterized in that the apparatus comprises:

the content pushing request receiving module is used for receiving a content pushing request aiming at a target pushing object;

a candidate push content set acquisition module, configured to respond to the content push request, and acquire a candidate push content set corresponding to the target push object;

the target rejection characteristic combination determining module is used for determining a target rejection characteristic combination based on the target object characteristics corresponding to the target push object, wherein the target rejection characteristic combination is determined based on the target characteristic rejection degree corresponding to the target characteristic combination;

and the target push content obtaining module is used for filtering the candidate push content corresponding to the target rejection characteristic combination from the candidate push content set to obtain the target push content corresponding to the target push object.

14. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor realizes the steps of the method of any one of claims 1 to 11 when executing the computer program.

15. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 11.