Content publishing method, device, computer-readable storage medium and electronic device

Abstract

Embodiments of the present disclosure provide a content publishing method, apparatus, medium, and electronic device, relating to the technical field of data processing. The method includes: obtaining initial recommended content and publishing the initial recommended content using an initial distribution method; the number of recommendations of the initial recommended content is less than or equal to a first threshold; monitoring the number of recommendations of the initial recommended content and determining first recommended content according to the initial recommended content, where the first recommended content is the initial recommended content for which the number of recommendations is greater than the first threshold; determining a first content evaluation parameter corresponding to the first recommended content according to the first user operation data corresponding to the first recommended content; determining second recommended content from the first recommended content according to the first content evaluation parameter, and publishing the second recommended content according to the first content evaluation parameter. The present disclosure can effectively filter out low-quality content and accelerate the distribution of high-quality content during the content recommendation process.

Description

Background Art

[0002] This section aims to provide background or context for the embodiments of the present disclosure described in the claims. The description herein is not admitted to be prior art merely by virtue of its inclusion in this section.

[0003] Recommendation systems can predict users' future behaviors and interests based on their historical behaviors and interests. Therefore, a large amount of user behavior data has become an important part and prerequisite for recommendation systems. For a newly developed website, the problem of cold start is how to design a personalized recommendation system without a large amount of user data and make users satisfied with the recommendation results so that they are willing to use the recommendation system. The cold start problem (cold start) can be mainly divided into three categories: user cold start, item cold start, and system cold start; among them, in information flow recommendation, item cold start mainly refers to content cold start.

[0004] Content-based Recommendation is the continuation and development of information filtering technology. It makes recommendations based on the content information of items, without relying on users' evaluation opinions of items. More importantly, it needs to use machine learning methods to obtain users' interest profiles from examples of feature descriptions of content.

[0005] Collaborative filtering, simply put, is to use the preferences of a group of people with similar interests and common experiences to recommend information that users are interested in. Individuals give a certain degree of response (such as scoring) to information through a cooperation mechanism and record it to achieve the purpose of filtering, thereby helping others to screen information. Summary of the Invention

[0006] The recommendation algorithms adopted by current mainstream information flow recommendation systems are mainly divided into content-based recommendation methods and user behavior-based recommendation methods. However, in the process of making recommendations using existing information flow recommendation systems, there is a problem that low-quality content cannot be effectively filtered and high-quality content cannot be screened out, resulting in the inability to guarantee the quality of the recommended content and reducing the user experience.

[0007] Therefore, in the prior art, there is a lack of a content publishing method that can effectively filter low-quality content and accelerate the distribution of high-quality content.

[0008] For this reason, there is a great need for an improved content publishing method to distinguish high-quality content and low-quality content according to the distribution effects at each stage during the content recommendation process, so as to effectively filter low-quality content and accelerate the distribution of high-quality content.

[0009] In this context, embodiments of the present disclosure are expected to provide a content publishing method, a content publishing device, a computer-readable storage medium, and an electronic device.

[0010] In a first aspect of the embodiments of the present disclosure, a content publishing method is provided, including: obtaining initial recommended content and publishing the initial recommended content in an initial distribution manner; the number of recommendations of the initial recommended content is less than or equal to a first threshold; monitoring the number of recommendations of the initial recommended content and determining first recommended content according to the initial recommended content, where the first recommended content is the initial recommended content with the number of recommendations greater than the first threshold; determining a first content evaluation parameter corresponding to the first recommended content according to the first user operation data corresponding to the first recommended content; determining second recommended content from the first recommended content according to the first content evaluation parameter and publishing the second recommended content according to the first content evaluation parameter.

[0011] In an embodiment of the present disclosure, before obtaining the initial recommended content, the above method further includes: obtaining initial content and extracting content features corresponding to the initial content; determining the feature attention degree of the content features and generating the number of recommendations of the initial content according to the feature attention degree, so as to determine the initial recommended content from the initial content according to the number of recommendations.

[0012] In an embodiment of the present disclosure, publishing the initial recommended content in an initial distribution manner includes: determining a first designated user; the first designated user is a user at a first activity level; in response to a page refresh operation of the first designated user for a first page, determining a target display area in the first page; and displaying the initial recommended content in the target display area.

[0013] In an embodiment of the present disclosure, the first content evaluation parameter includes a first duration quantile ratio. Determining the first content evaluation parameter corresponding to the first recommended content according to the first user operation data corresponding to the first recommended content includes: regarding the stage of content publishing in the initial distribution manner as the first publishing stage; obtaining the first user operation data of the first recommended content; the first user operation data is the user operation data corresponding to the first recommended content in the first publishing stage; determining a first exposure click-through rate and a first average content browsing duration according to the first user operation data; determining a first single-exposure browsing duration of the first recommended content in the first publishing stage according to the first exposure click-through rate and the first average content browsing duration; and determining the first duration quantile ratio according to the first single-exposure browsing duration.

[0014] In an embodiment of the present disclosure, determining the first duration quantile ratio according to the first single-exposure browsing duration includes: obtaining a set of single-exposure durations; the set of single-exposure durations includes the first single-exposure browsing durations of all the first recommended content; obtaining a quantile parameter and determining the first duration quantile ratio corresponding to the set of single-exposure durations according to the quantile parameter and multiple first single-exposure browsing durations.

[0015] In one embodiment of the present disclosure, determining second recommended content from first recommended content according to a first content evaluation parameter includes: obtaining a parameter threshold; the parameter threshold includes a quantile ratio threshold; determining first recommended content with a first duration quantile ratio less than the quantile ratio threshold as filtered content; and determining first recommended content with a first duration quantile ratio greater than or equal to the quantile ratio threshold as second recommended content.

[0016] In one embodiment of the present disclosure, publishing second recommended content according to a first content evaluation parameter includes: determining a second designated user; the second designated user is a user at or above a second activity level; obtaining the duration quantile ratio of the second recommended content, and determining a recommendation coefficient of the second recommended content according to the duration quantile ratio of the second recommended content; determining a recommendation priority of the second recommended content according to the recommendation coefficient of the second recommended content, and publishing the second recommended content to the second designated user according to the recommendation priority.

[0017] In one embodiment of the present disclosure, after publishing second recommended content according to a first content evaluation parameter, the above method further includes: determining the total number of contents in the second stage; the total number of contents in the second stage is the number of all contents published according to the first content evaluation parameter; determining the number of second contents; the number of second contents is the number of contents published according to the first content evaluation parameter and with a recommendation count greater than a second threshold; the second threshold is greater than the first threshold; and determining the recommendation effectiveness rate of the second publishing stage according to the number of second contents and the total number of contents in the second stage.

[0018] In one embodiment of the present disclosure, after publishing second recommended content according to a first content evaluation parameter, the above method further includes: determining second recommended content with a recommendation count greater than a second threshold as third recommended content; the second threshold is greater than the first threshold; taking the stage of content publishing according to the first content evaluation parameter as the second publishing stage; obtaining second user operation data corresponding to the third recommended content; the second user operation data is the user operation data corresponding to the third recommended content in the second publishing stage; determining a second content evaluation parameter corresponding to the third recommended content according to the second user operation data; determining fourth recommended content from the third recommended content according to the second content evaluation parameter, and publishing the fourth recommended content to all users according to the second content evaluation parameter.

[0019] In a second aspect of the embodiments of the present disclosure, a content publishing device is provided, including: a first content publishing module, configured to obtain initial recommended content and publish the initial recommended content by using an initial distribution method; the number of recommendations of the initial recommended content is less than or equal to a first threshold; a first content determination module, configured to monitor the number of recommendations of the initial recommended content and determine first recommended content according to the initial recommended content, where the first recommended content is the initial recommended content with the number of recommendations greater than the first threshold; an evaluation parameter determination module, configured to determine a first content evaluation parameter corresponding to the first recommended content according to first user operation data corresponding to the first recommended content; a second content publishing module, configured to determine second recommended content from the first recommended content according to the first content evaluation parameter and publish the second recommended content according to the first content evaluation parameter.

[0020] In an embodiment of the present disclosure, the content publishing device further includes an initial content determination module, and the initial content determination module is configured to: obtain initial content and extract content features corresponding to the initial content; determine the feature attention degree of the content features, generate the number of recommendations of the initial content according to the feature attention degree, so as to determine the initial recommended content from the initial content according to the number of recommendations.

[0021] In an embodiment of the present disclosure, the first content publishing module includes a first content publishing unit, and the first content publishing unit is configured to: determine a first specified user; the first specified user is a user in a first activity level; in response to a page refresh operation of the first specified user for a first page, determine a target display area on the first page; and display the initial recommended content in the target display area.

[0022] In an embodiment of the present disclosure, the first content evaluation parameter includes a first duration quantile ratio, the evaluation parameter determination module includes an evaluation parameter determination unit, and the evaluation parameter determination unit includes: a first stage determination subunit, configured to use the stage of content publishing by using the initial distribution method as a first publishing stage; a first data acquisition subunit, configured to acquire first user operation data of the first recommended content; the first user operation data is the user operation data corresponding to the first recommended content in the first publishing stage; a first parameter determination subunit, configured to determine a first exposure click-through rate and a first average content browsing duration according to the first user operation data; a first duration determination subunit, configured to determine a first single-exposure browsing duration of the first recommended content in the first publishing stage according to the first exposure click-through rate and the first average content browsing duration; and a first quantile ratio determination subunit, configured to determine the first duration quantile ratio according to the first single-exposure browsing duration.

[0023] In one embodiment of the present disclosure, the first quantile ratio determination subunit is configured to: obtain a set of single-exposure durations; the set of single-exposure durations includes the first single-exposure viewing durations of all first recommended contents; obtain a quantile parameter, and determine a first duration quantile ratio corresponding to the set of single-exposure durations according to the quantile parameter and a plurality of first single-exposure viewing durations.

[0024] In one embodiment of the present disclosure, the second content publishing module includes a second content publishing unit, and the second content publishing unit is configured to: obtain a parameter threshold; the parameter threshold includes a quantile ratio threshold; determine the first recommended contents with the first duration quantile ratio less than the quantile ratio threshold as filtered contents; determine the first recommended contents with the first duration quantile ratio greater than or equal to the quantile ratio threshold as second recommended contents.

[0025] In one embodiment of the present disclosure, the second content publishing module includes a second content publishing unit, and the second content publishing unit is configured to: determine a second designated user; the second designated user is a user at or above the second activity level; obtain the duration quantile ratio of the second recommended content, and determine a recommendation coefficient of the second recommended content according to the duration quantile ratio of the second recommended content; determine a recommendation priority of the second recommended content according to the recommendation coefficient of the second recommended content, and publish the second recommended content to the second designated user according to the recommendation priority.

[0026] In one embodiment of the present disclosure, the content publishing device further includes a recommendation evaluation module, and the recommendation evaluation module is configured to: determine the total number of contents in the second stage; the total number of contents in the second stage is the number of all contents published according to the first content evaluation parameter; determine the number of second contents; the number of second contents is the number of contents published according to the first content evaluation parameter and with the number of recommendations greater than a second threshold; the second threshold is greater than the first threshold; determine the recommendation effectiveness rate of the second publishing stage according to the number of second contents and the total number of contents in the second stage.

[0027] In one embodiment of the present disclosure, the content publishing device further includes a third content publishing module, and the third content publishing is configured to: determine the second recommended contents with the number of recommendations greater than the second threshold as third recommended contents; the second threshold is greater than the first threshold; regard the stage of content publishing according to the first content evaluation parameter as the second publishing stage; obtain second user operation data corresponding to the third recommended contents; the second user operation data is the user operation data corresponding to the third recommended contents in the second publishing stage; determine a second content evaluation parameter corresponding to the third recommended contents according to the second user operation data; determine fourth recommended contents from the third recommended contents according to the second content evaluation parameter, and publish the fourth recommended contents to all users according to the second content evaluation parameter.

[0028] In a third aspect of the embodiments of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the content publishing method as described above is implemented.

[0029] In a fourth aspect of the embodiments of the present disclosure, an electronic device is provided, including: a processor; and a memory, on which computer-readable instructions are stored, and when the computer-readable instructions are executed by the processor, the content publishing method as described above is implemented.

[0030] According to the technical solution of the embodiments of the present disclosure, on the one hand, the initial recommended content is published by using the initial distribution method, which can ensure that the initial recommended content with a small number of recommendations has sufficient content distribution times to collect user behavior data. On the other hand, content publishing based on the content evaluation parameters determined according to the user operation data can effectively filter out low-quality content and expand the publishing of high-quality content, which is beneficial to improving the recommendation efficiency and the quality of recommended content of content recommendation, and further enhancing the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present disclosure will become readily understood. In the drawings, several embodiments of the present disclosure are shown by way of illustration and not limitation, wherein:

[0032] Figure 1 A schematic block diagram of a system architecture of an exemplary application scenario according to some embodiments of the present disclosure is schematically shown;

[0033] Figure 2 A schematic flowchart of a content publishing method according to some embodiments of the present disclosure is schematically shown;

[0034] Figure 3 A schematic overall flowchart of content publishing in stages according to some embodiments of the present disclosure is schematically shown;

[0035] Figure 4 A schematic flowchart of publishing initial recommended content by using the initial distribution method according to some embodiments of the present disclosure is schematically shown;

[0036] Figure 5 A schematic flowchart of determining the first content evaluation parameter of the first recommended content according to some embodiments of the present disclosure is schematically shown;

[0037] Figure 6 A schematic flowchart of determining the second recommended content from the first recommended content according to some embodiments of the present disclosure is schematically shown;

[0038] Figure 7 Schematically shows a flowchart of publishing a second recommended content according to a first content evaluation parameter according to some embodiments of the present disclosure;

[0039] Figure 8 Schematically shows a flowchart of determining the recommendation effectiveness rate of a second distribution stage according to some embodiments of the present disclosure;

[0040] Figure 9 Schematically shows a schematic block diagram of a content publishing device according to some embodiments of the present disclosure;

[0041] Figure 10 Schematically shows a schematic diagram of a storage medium according to an exemplary embodiment of the present disclosure; and

[0042] Figure 11 Schematically shows a block diagram of an electronic device according to an exemplary embodiment of the disclosure.

[0043] In the drawings, the same or corresponding reference numerals indicate the same or corresponding parts. Detailed Embodiments

[0044] The principles and spirit of the present disclosure will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are given only to enable those skilled in the art to better understand and thus implement the present disclosure, and are not intended to limit the scope of the present disclosure in any way. On the contrary, these embodiments are provided to make the present disclosure more thorough and complete, and to be able to convey the scope of the present disclosure fully to those skilled in the art.

[0045] Those skilled in the art know that the embodiments of the present disclosure can be implemented as a system, a device, an equipment, a method, or a computer program product. Therefore, the present disclosure can be specifically implemented in the following forms, namely: complete hardware, complete software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0046] According to the embodiments of the present disclosure, a content publishing method, a content publishing device, a medium, and an electronic device are provided.

[0047] In this article, it should be understood that the terms involved, such as the Click-through Rate (CTR), can refer to the ratio of the number of times a specified content on a website or application (APP) is clicked and exposed. The click-through rate is usually an important measurement indicator in the recommendation system. The average browsing duration of the content can be the ratio of the browsing duration of the specified content on the website or APP to the number of clicks. The average browsing duration of the content is an important measurement indicator in the recommendation system; among them, the specified content can be any form of readable content such as video content and graphic content. The average stay duration per capita can be the usage duration of each user on the website or APP on average within a certain period of time, which is an important business indicator in the information flow recommendation. In addition, the number of any element in the attached drawings is for illustration rather than limitation, and any naming is only for distinction and does not have any limiting meaning.

[0048] Next, with reference to several representative embodiments of the present disclosure, the principles and spirits of the present disclosure will be elaborated in detail. Summary of the Invention

[0050] The recommendation algorithms adopted by the current mainstream information flow recommendation systems are mainly divided into content-based recommendation methods and user behavior-based recommendation methods, among which the user behavior-based recommendation method has better effects. In the information flow recommendation system, a piece of content generally goes through two stages from entering the recommendation pool to finally expiring: the content cold start stage and the regular recommendation stage.

[0051] In the content cold start stage, due to the lack of user behavior data, it mainly relies on content-based recommendation methods for recommendation to accumulate a certain amount of user behavior data for new content. In the regular stage, after accumulating a certain amount of behavior data, recommendation methods based on behavior (such as collaborative filtering recommendation, etc.) and content-based recommendation methods are used for recommendation.

[0052] However, in the information flow recommendation system, a large amount of new content enters the recommendation pool every day, and only a very small part of it is high-quality content that users are interested in. Using the above-mentioned recommendation methods, there is a problem that low-quality content cannot be effectively filtered out and high-quality content cannot be screened out during the recommendation process, resulting in the inability to guarantee the quality of the recommended content and reducing the user experience.

[0053] Based on the above, the basic idea of the present disclosure is to obtain initial recommended content and publish the initial recommended content using an initial distribution method; the number of recommendations of the initial recommended content is less than or equal to a first threshold; monitor the number of recommendations of the initial recommended content, and determine a first recommended content based on the initial recommended content, where the first recommended content is the initial recommended content with the number of recommendations greater than the first threshold; determine a first content evaluation parameter corresponding to the first recommended content according to the first user operation data corresponding to the first recommended content; determine a second recommended content from the first recommended content according to the first content evaluation parameter, and publish the second recommended content according to the first content evaluation parameter. The present disclosure can determine a content evaluation parameter according to the user operation data of the recommended content, and can effectively filter out low-quality content during the content recommendation process through the content evaluation parameter, and accelerate the distribution of high-quality content, improving the publishing efficiency of content recommendation and the quality of content placement, thereby enhancing the user experience.

[0054] After introducing the basic principle of the present disclosure, the following specifically introduces various non-limiting implementation manners of the present disclosure.

[0055] Overview of Application Scenarios

[0056] First, refer to Figure 1 , Figure 1 which shows a schematic block diagram of a system architecture of an exemplary application scenario of a content publishing method and apparatus to which embodiments of the present disclosure can be applied.

[0057] As Figure 1 shown, the system architecture 100 may include one or more of terminal devices 101, 102, 103, a network 104, and a server 105. The network 104 is used to provide a medium for a communication link between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, etc. The terminal devices 101, 102, 103 may be various electronic devices with a display screen, including but not limited to desktop computers, portable computers, smart phones, and tablet computers, etc. It should be understood that Figure 1 the numbers of terminal devices, networks, and servers in

[0058] The content publishing method provided by the embodiments of the present disclosure is generally executed by the server 105. Correspondingly, the content publishing device is generally disposed in the server 105. However, those skilled in the art can easily understand that the content publishing method provided by the embodiments of the present disclosure can also be executed by the terminal devices 101, 102, 103. Correspondingly, the content publishing device can also be disposed in the terminal devices 101, 102, 103. No special limitation is made in this exemplary embodiment. For example, in an exemplary embodiment, the user uploads the pre-constructed knowledge graph and the user question input by the user to the server 105 through the terminal devices 101, 102, 103. The server 105 sends the initial recommended content to the terminal devices 101, 102, 103, etc. through the content publishing method provided by the embodiments of the present disclosure, so that the terminal devices 101, 102, 103 display the initial recommended content to the user. Further, the server 105 can determine the first recommended content, obtain the first user operation data of the first recommended content from the terminal devices 101, 102, 103, etc., determine the corresponding first content evaluation parameter according to the first user operation data, and determine the second recommended content from the first recommended content, so as to send the second recommended content to the terminal devices 101, 102, 103, etc. through the network 104, so that the terminal devices 101, 102, 103 display the second recommended content to the user.

[0059] It should be understood that Figure 1 The application scenario shown is only an example in which the embodiments of the present disclosure can be implemented. The scope of application of the embodiments of the present disclosure is not limited by any aspect of this application scenario.

[0060] Exemplary Method

[0061] The following combines Figure 1 the application scenario of Figure 2 to describe the content publishing method according to the exemplary embodiments of the present disclosure. It should be noted that the above application scenario is only shown for the convenience of understanding the spirit and principle of the present disclosure, and the embodiments of the present disclosure are not limited in this regard. On the contrary, the embodiments of the present disclosure can be applied to any applicable scenario.

[0062] The present disclosure first provides a content publishing method. The execution subject of this method can be a terminal device or a server. The present disclosure does not make special limitations in this regard. In this exemplary embodiment, the case where the server executes this method is taken as an example for illustration.

[0063] Referring to Figure 2 shown, in step S210, obtain the initial recommended content and publish the initial recommended content by using the initial distribution method; the number of recommendations of the initial recommended content is less than or equal to the first threshold.

[0064] In some example embodiments, the recommended number may be the recommended quantity corresponding to a piece of content to be recommended. The first threshold may be a preset threshold. For example, the first threshold may be 50, 100, etc. The initial recommended content may be new content corresponding to the content cold start phase of the recommendation system, and the recommended number of the initial recommended content is less than or equal to the first threshold. The initial recommended content may be an article, a video, graphic materials, etc. The initial distribution method may be a content distribution method adopted to increase the exposure times of the initial recommended content. For example, the initial distribution method may be a forced distribution method, etc. Forced distribution may be a method of inserting the initial recommended content at a fixed position during each refresh operation of the user.

[0065] For a newly established recommendation system, during the content cold start phase of the system, it is mainly to ensure that new recommended content has a certain number of exposure times. Therefore, for an initial recommended content with a recommended number less than or equal to the first threshold, the initial content distribution method can be adopted to publish the initial recommended content. Refer to Figure 3 , Figure 3 Schematically shows an overall flowchart of content publication in stages according to some embodiments of the present disclosure. When a new recommended content 310 is obtained, in step S301, if the recommended number of this recommended content is less than or equal to the first threshold, this recommended content can be determined as the initial recommended content. In step S302, the initial recommended content can be published by adopting a content-based recommendation method. In step S303, specifically, the initial recommended content can be sent to specified users by adopting a forced publication method.

[0066] In an embodiment of the present disclosure, obtain the initial content, and extract the content features corresponding to the initial content; determine the feature attention degree of the content features, generate the recommended number of the initial content according to the feature attention degree, so as to determine the initial recommended content from the initial content.

[0067] Among them, the initial content may be the latest generated recommended content in the recommendation system. The content features may be the features obtained after extracting key features of the initial content. For example, for an article, by extracting keywords from the text content of the article, the content features of the article, that is, the article keywords, can be determined. The feature attention degree may be the attention degree of the user to the content features.

[0068] Specifically, in the recommendation system, for each newly generated initial content, content extraction processing can be performed on the initial content to obtain the content features of the initial content, so as to calculate the feature attention degree of the determined content features. When calculating the feature attention degree, multiple feature attention degree consideration factors can be combined to comprehensively determine the value of the feature attention degree. According to the calculated feature attention degree, the recommended number of the initial content can be generated, and the recommended number is compared with the first threshold to determine the initial recommended content from the initial content.

[0069] For example, during the period of holding a sports meeting, users may be more interested in the content related to the sports meeting. For an article related to the sports meeting, keywords such as "sports meeting", "holding time", and "opening ceremony" can be extracted from the article, which are the content features. Since people are more interested in such content during this period, after extracting the content features of the article, the feature attention degree of the content features is relatively high. For the content features with a relatively high feature attention degree, the generated number of recommendations is also high. The determined number of recommendations is compared with the first threshold, and the initial content with the number of recommendations less than or equal to the first threshold is used as the initial recommended content.

[0070] In an embodiment of the present disclosure, a first designated user is determined; the first designated user is a user in the first activity level; in response to a page refresh operation of the first designated user for the first page, a target display area is determined in the first page; and the initial recommended content is displayed in the target display area.

[0071] Among them, the activity level can represent the activity degree of the user. The smaller the value of the user activity level, the higher the corresponding activity level. For example, it can be pre-agreed that the user in the first activity level has the highest activity degree, and the activity degree of the user in the second activity level is lower than that of the user in the first activity level, and so on. In a specific application scenario, the activity level can be divided according to requirements. The first designated user can be a user in the first activity level, mainly referring to a user with a relatively high activity degree in the recommendation system. The first page can be a page in the recommendation system for presenting recommended content to the user. The page refresh operation can be a refresh operation of the user for the first page. For example, the page refresh operation can be a refresh operation when the user initially opens the first page; the page refresh operation can also be a refresh operation when the user wants to update the page content displayed on the first page after staying on the first page for a period of time, etc. The target display area can be a display area in the first page for presenting the initial recommended content after receiving the page refresh operation for the first page.

[0072] In the content cold start stage, it is mainly to ensure that the initial recommended content has a certain number of exposure times. In this stage, since the initial recommended content lacks sufficient user behavior data, a content-based recommendation method is mainly used for content distribution. For example, the initial recommended content can be published by means of forced distribution. Refer to Figure 4 , Figure 4A flowchart showing the process of releasing initial recommended content using an initial distribution method according to some embodiments of the present disclosure is schematically illustrated. In step S410, after determining the initial recommended content, a first designated user in the recommendation system, that is, a user at the first active level, can be determined. The first designated user can be the user with the highest level of activity in the recommendation system. In step S420, if the first designated user performs a page refresh operation on the first page, in response to the page refresh operation of the first designated user on the first page, a target display area (such as the upper right area of the page) is determined on the first page. In step S430, the initial recommended content is displayed in the target display area so that the user can view or operate the initial recommended content through the first page.

[0073] In this stage, the recommendation effectiveness rate of the first release stage can be introduced to evaluate the content release effect of this stage, as specifically shown in Formula 1.

[0074]

[0075] Wherein, n1 can be the first threshold; can be the number of contents with the recommended number greater than the first threshold n1 within a preset time period (such as 24 hours); N1 can be the number of all contents released within a preset time period (such as 24 hours); the recommendation effectiveness rate validRate of the first release stage s1 can be the proportion of new contents that reach the first threshold n1 within a preset time period (such as 24 hours).

[0076] In step S220, monitor the recommended number of the initial recommended content, and determine the first recommended content based on the initial recommended content. The first recommended content is the initial recommended content with the recommended number greater than the first threshold.

[0077] In some example embodiments, the first recommended content can be the initial recommended content with the recommended number greater than the first threshold.

[0078] In the cold start stage of the recommendation system, the initial recommended content can be continuously sent to the first designated user using the forced distribution method. As the exposure times of the initial recommended content increase, the recommended number of the initial recommended content also changes simultaneously. Therefore, when releasing the initial recommended content using the initial content distribution method, the recommended number of the initial recommended content can be monitored simultaneously. Compare the recommended number of the initial recommended content with the first threshold, and determine the initial recommended content with the recommended number greater than the first threshold as the first recommended content to further perform content release on the first recommended content. Refer to Figure 3, in step S301, compare the number of recommendations of the initial recommended content with the first threshold. If the number of recommendations of the initial recommended content is greater than the first threshold, then continue to execute step S305. Optionally, the relationship between the number of recommendations and the second threshold can also be determined through step S304 to determine whether to further publish the recommended content.

[0079] In step S230, determine the first content evaluation parameter corresponding to the first recommended content according to the first user operation data corresponding to the first recommended content.

[0080] In some exemplary embodiments, the first user operation data may be the operation data generated when the user views the first recommended content in the first release stage. The first content evaluation parameter may be a parameter for evaluating the content quality of the first recommended content.

[0081] Since during the process of the user browsing the first recommended content, first user operation data corresponding to the first recommended content will be generated. For example, the first user operation data may include the click-through rate of the user on the first recommended content, may also include the average browsing duration of the content of the first recommended content, and may also include the average stay duration per person of the first recommended content, etc. The first content evaluation parameter corresponding to the first recommended content can be determined according to the obtained first user operation data. Refer to Figure 3 , in step S305, the first content evaluation parameter can be determined according to the first user operation data.

[0082] In an embodiment of the present disclosure, the stage of content release using the initial distribution method is used as the first release stage; obtain the first user operation data of the first recommended content; the first user operation data is the user operation data corresponding to the first recommended content in the first release stage; determine the first exposure click-through rate and the first average content browsing duration according to the first user operation data; determine the first single-exposure browsing duration of the first recommended content in the first release stage according to the first exposure click-through rate and the first average content browsing duration; determine the first duration percentile ratio according to the first single-exposure browsing duration.

[0083] Among them, the first release stage can be the stage of releasing recommended content using the initial distribution method. Since there is a lack of sufficient user operation data for the recommended content in the first release stage, the initial distribution method is adopted for content release to ensure that the recommended content has sufficient exposure times. The exposure click-through rate can be the ratio of the number of clicks on a recommended content to the number of exposures. The first exposure click-through rate can be the exposure click-through rate corresponding to a recommended content in the first release stage. The first average content viewing duration can be the average viewing duration of all recommended content released in the first release stage. The first single-exposure viewing duration can be the duration that the first recommended content is viewed by the user during one exposure. The first duration percentile ratio can be the percentile ratio corresponding to the first single-exposure viewing duration.

[0084] Reference Figure 5 , Figure 5 schematically shows a flowchart of determining the first content evaluation parameter of the first recommended content according to some embodiments of the present disclosure. In step S510, the stage of content release using the initial distribution method is taken as the first release stage. In step S520, since after the content release in the first release stage, the recommended content already has certain user behavior data, the first user operation data corresponding to the first recommended content in the first release stage can be obtained; wherein, the first user operation data is the user operation data corresponding to the first recommended content in the first release stage. In order to evaluate the quality of the recommended content released in the first release stage, in step S530, the first exposure click-through rate and the first average content viewing duration, etc. can be determined according to the first user operation data. In step S540, according to the first exposure click-through rate and the first average content viewing duration, the first single-exposure viewing duration of the first recommended content in the first release stage is determined. Specifically, the first single-exposure viewing duration can be calculated by formula 2.

[0085] avgExpDu s1 = ctr s1 × avgDu s1 (Formula 2)

[0086] Wherein, ctr s1 can represent the exposure click-through rate of a recommended content in the first release stage; avgDu s1 can represent the first average content viewing duration corresponding to a recommended content in the first release stage; avgExpDu s1 can represent the first single-exposure viewing duration corresponding to a recommended content in the first release stage.

[0087] In step S550, the first duration percentile ratio can be calculated according to the first single-exposure viewing duration, so as to perform content distribution on the first recommended content according to the calculated first content evaluation parameter.

[0088] In one embodiment of the present disclosure, a set of single-exposure durations is obtained; the set of single-exposure durations includes the first single-exposure browsing durations of all first recommended contents; a quantile parameter is obtained, and a first duration quantile ratio corresponding to the set of single-exposure durations is determined according to the quantile parameter and multiple first single-exposure browsing durations.

[0089] Among them, the set of single-exposure durations may be a set composed of the first single-exposure browsing durations of all first recommended contents corresponding to the first release stage. The quantile parameter may be a preset quantile parameter value, and the quantile parameter may be used to determine the proportion of contents classified as low-quality contents. The first duration quantile ratio may be the quantile ratio of all first single-exposure browsing durations corresponding to the first release stage, and the first duration quantile ratio may be used to determine second recommended contents from the first recommended contents.

[0090] After the first release stage, the first single-exposure browsing durations of all first recommended contents corresponding to the first release stage can be determined, and the obtained all first single-exposure browsing durations are composed into a set of single-exposure durations. The first duration quantile ratio can be determined according to the set of single-exposure durations and the quantile parameter, as specifically shown in Formula 3.

[0091] Q p1 = Quantile({avgExpDu s1} total , p1) (Formula 3)

[0092] Among them, {avgExpDu s1} total can represent the set of single-exposure browsing durations of all recommended contents released in the first release stage; p1 can be a set quantile parameter, used to determine the proportion of contents classified as low-quality contents in the first release stage; Q p1 can be the p1 quantile of the single-exposure browsing durations of all recommended contents in the first release stage.

[0093] According to the calculated first single-exposure browsing durations corresponding to each first recommended content and Q p1 the first duration quantile ratio w s1 can be calculated, as specifically shown in Formula 4.

[0094]

[0095] In step S240, second recommended contents are determined from the first recommended contents according to the first content evaluation parameter, and the second recommended contents are released according to the first content evaluation parameter.

[0096] In some example embodiments, the second recommended content may be the recommended content determined from the first recommended content according to the first content evaluation parameter. The second recommended content may be the high-quality content in the first recommended content.

[0097] After determining the first content evaluation parameter corresponding to the first recommended content, the second recommended content may be determined from the first recommended content according to the first content evaluation parameter, and the second recommended content may be published according to the first content evaluation parameter.

[0098] Due to the content distribution in the first publishing stage, the recommended content already has certain user behavior data. In the second publishing stage after the first publishing stage, various recommendation methods based on content and user behavior data can be further used for content recommendation. Therefore, the quality of the first recommended content can be evaluated according to the content publishing effect in the first publishing stage, and the first content evaluation parameter can be used to identify the content quality of the first recommended content.

[0099] In an embodiment of the present disclosure, a parameter threshold is obtained; the parameter threshold includes a quantile ratio threshold; the first recommended content with a first duration quantile ratio less than the quantile ratio threshold is determined as filtered content; the first recommended content with a first duration quantile ratio greater than or equal to the quantile ratio threshold is determined as the second recommended content.

[0100] Among them, the parameter threshold may be the value used to determine the second recommended content from the first recommended content. The quantile ratio threshold may be a pre-set value used to compare with the first duration quantile ratio. For example, the quantile ratio threshold may be 1. The filtered content may be the low-quality content determined from the first recommended content, and the low-quality content will no longer be published in the second publishing stage.

[0101] Reference Figure 6 , Figure 6 Schematically shows a flowchart of determining the second recommended content from the first recommended content according to some embodiments of the present disclosure. In step S610, a parameter threshold is obtained; such as the quantile ratio threshold, to compare the first duration quantile ratio with the quantile ratio threshold. In step S620, the first recommended content with a first duration quantile ratio less than the quantile ratio threshold is determined as filtered content. In step S630, the first recommended content with a first duration quantile ratio greater than or equal to the quantile ratio threshold is determined as the second recommended content.

[0102] Reference Figure 3 ,in step S306, the calculated first duration quantile ratio may be compared with the quantile ratio threshold. For example, the quantile ratio threshold may be set to 1, then in step S307, the w in the first recommended content s1Contents with a value of ≥1 are determined as high-quality contents, i.e., the second recommended contents, for further content distribution of the second recommended contents. In step S308, contents with a value of w s1 <1 in the first recommended contents can be determined as filtered contents (i.e., low-quality contents) and filtered out, and no content recommendation will be made for them.

[0103] In an embodiment of the present disclosure, a second designated user is determined; the second designated user is a user at or above the second activity level; the duration quantile ratio of the second recommended content is obtained, and the recommendation coefficient of the second recommended content is determined according to the duration quantile ratio of the second recommended content; the recommendation priority of the second recommended content is determined according to the recommendation coefficient of the second recommended content, and the second recommended content is published to the second designated user according to the recommendation priority.

[0104] Among them, the second designated user can be a user at or above the second activity level, that is, users with activity levels of the first activity level and the second activity level in the recommendation system. The calculation method of the duration quantile ratio of the second recommended content can be the same as that of the duration quantile ratio of the first recommended content, and it is a specific value. The recommendation coefficient can be a coefficient used to determine the recommendation priority of the second recommended content. The recommendation priority can be a coefficient used when determining the publication priority of the second recommended content for content publication.

[0105] In the recommendation system, after the first publication stage, if the number of recommendations of the recommended content is greater than the first threshold, it enters the second publication stage, that is, the preliminary screening stage. In the second publication stage, low-quality contents can be preliminarily filtered out to expand the distribution of high-quality contents. Refer to Figure 7 , Figure 7 Schematically shows a flowchart of publishing the second recommended content according to the first content evaluation parameter according to some embodiments of the present disclosure. In step S710, when publishing content in the second publication stage, the second designated user can be determined first; the second designated user is a user at or above the second activity level. In step S720, the duration quantile ratio of the second recommended content is obtained, and the recommendation coefficient of the second recommended content is determined according to the duration quantile ratio of the second recommended content. The duration quantile ratio of the second recommended content is w s1 . For the second recommended content with w s1 ≥1, the value of w s1 corresponding to each second recommended content can be used as the weighted weight for subsequent content publication, and the recommendation coefficient is calculated. In step S730, the recommendation priority of the second recommended content is determined according to the recommendation coefficient of the second recommended content, and the second recommended content is published to the second designated user according to the recommendation priority. Specifically, the higher the value of w s1 of the second recommended content, the higher the recommendation priority.

[0106] In one embodiment of the present disclosure, determine the total number of contents in the second stage; the total number of contents in the second stage is the number of all contents published according to the first content evaluation parameter; determine the number of second contents; the number of second contents is the number of contents published according to the first content evaluation parameter and with the recommendation number greater than the second threshold; the second threshold is greater than the first threshold; determine the recommendation effectiveness rate in the second release stage according to the number of second contents and the total number of contents in the second stage.

[0107] Among them, the total number of contents in the second stage can be the number of all contents published according to the first content evaluation parameter, that is, the total number of contents published in the second release stage. The number of second contents can be the number of contents published according to the first content evaluation parameter and with the recommendation number greater than the second threshold, that is, the number of recommended contents published in the second release stage and with the recommendation number greater than the second threshold. The second threshold can be a threshold preset for comparison with the recommendation number, and the second threshold is greater than the first threshold. The recommendation effectiveness rate in the second release stage can be the proportion of the number of recommended contents reaching the second threshold in the second release stage to the number of all recommended contents published in the second release stage.

[0108] Reference Figure 8 , Figure 8 schematically shows a flowchart of determining the recommendation effectiveness rate in the second sub-release stage according to some embodiments of the present disclosure. In step S810, determine the total number of contents in the second stage; the total number of contents in the second stage is the number of all contents published according to the first content evaluation parameter. In step S820, determine the number of second contents; the number of second contents is the number of contents published according to the first content evaluation parameter and with the recommendation number greater than the second threshold; the second threshold is greater than the first threshold. In step S830, determine the recommendation effectiveness rate in the second release stage according to the number of second contents and the total number of contents in the second stage. Specifically, the recommendation effectiveness rate in the second release stage can be calculated using formula 5, and can be represented by validRate s2 It is represented.

[0109]

[0110] Among them, n2 can be the second threshold; the number of contents published in the second release stage and with the recommendation number greater than the second threshold n2, that is, the number of second contents; N2 can be the total number of contents in the second stage, that is, the number of all recommended contents published in the second release stage.

[0111] Through the above two release stages, the purpose of content distribution for recommended contents can already be achieved. In some application scenarios, in order to achieve a better content distribution effect, the recommended contents can be further distributed through optional release stages.

[0112] In one embodiment of the present disclosure, the second recommended content with a recommended number greater than a second threshold is determined as the third recommended content; the second threshold is greater than the first threshold; the stage of content publishing according to the first content evaluation parameter is used as the second publishing stage; the second user operation data corresponding to the third recommended content is obtained; the second user operation data is the user operation data corresponding to the third recommended content in the second publishing stage; the second content evaluation parameter corresponding to the third recommended content is determined according to the second user operation data; the fourth recommended content is determined from the third recommended content according to the second content evaluation parameter, and the fourth recommended content is published to all users according to the second content evaluation parameter.

[0113] Among them, the third recommended content may be the second recommended content with a recommended number greater than the second threshold. The second publishing stage may be the stage of content publishing according to the first content evaluation parameter. For example, the second content publishing stage may be the preliminary content screening stage. The second user operation data may be the user operation data corresponding to the third recommended content published in the second publishing stage. The second content evaluation parameter may be the content evaluation parameter corresponding to the third recommended content published in the second publishing stage. The fourth recommended content may be high-quality content determined from the third recommended content according to the second content evaluation parameter.

[0114] After the second publishing stage, the stage of determining the fourth recommended content from the third recommended content for content publishing may be an optional publishing stage. If a recommendation system only includes three content publishing stages, the optional publishing stage may also be referred to as the final screening stage. This stage is used to finally filter out low-quality content and expand the publishing of high-quality content. In this stage, the corresponding user set is all users in the recommendation system. After the content publishing in the second publishing stage, the third recommended content already has sufficient user behavior data. Therefore, various recommendation methods based on content and user behavior data can be used for content publishing.

[0115] Reference Figure 3 In step S309, the second recommended content with a recommended number greater than the second threshold can be determined as the third recommended content, and the second user operation data corresponding to the third recommended content is obtained to calculate the second content evaluation parameter corresponding to the third recommended content according to the second user operation data. Similar to the first content evaluation parameter, the second content evaluation parameter may include the second user operation data, such as the second exposure click-through rate, the second average content browsing duration, the second single-exposure browsing duration, and the second duration percentile ratio, etc. Specifically, the second single-exposure browsing duration can be calculated using formula 6.

[0116] avgExpDu s2 =ctr s2 ×avgDu s2 (Formula 6)

[0117] Among them, ctr s2 can represent the exposure click-through rate of a recommended content in the second release stage; avgDu s2 can represent the average viewing duration of the second content corresponding to a recommended content in the second release stage; avgExpDu s2 can represent the viewing duration per single exposure of the second content corresponding to a recommended content in the second release stage.

[0118] After calculating the viewing duration per single exposure of the second content corresponding to all the third recommended contents, the second duration quantile ratio can be calculated using Formula 7.

[0119] Q p2 = Quantile({avgExpDu s2} total , p2) (Formula 7)

[0120] Among them, {avgExpDu s2} total can represent the set of the viewing durations per single exposure of all the recommended contents released in the second release stage; p2 can be a set quantile parameter used to determine the proportion of the contents classified as low-quality contents in the second release stage; Q p2 can be the p2 quantile of the viewing durations per single exposure of all the recommended contents in the second release stage.

[0121] According to the calculated viewing duration per single exposure of the second content corresponding to each second recommended content and Q p2 the second duration quantile ratio w s2 can be calculated, as specifically shown in Formula 8.

[0122]

[0123] Reference Figure 3 , in step S309, the calculated second duration quantile ratio can be compared with the quantile ratio threshold. For example, the quantile ratio threshold can be set to 1, then in step S311, the contents in the third recommended contents with w s2 ≥1 can be determined as high-quality contents, that is, the fourth recommended contents, so as to perform the next content distribution for the fourth recommended contents. In step S312, the contents in the third recommended contents with w s2 <1 can be determined as filtered contents (i.e., low-quality contents), and the filtered contents can be filtered out and no longer recommended for content.

[0124] During the process of content release for the fourth recommended contents, for the fourth recommended contents with w s2 ≥1, for each w s2The value is used as the weighted weight for subsequent content publication, calculates the recommendation coefficient, and determines the corresponding recommendation priority according to the recommendation coefficient. Specifically, w s2 The higher the value of w, the higher the recommendation priority of the fourth recommended content. In this content publication stage, the exposure click-through rate and the average stay duration of all content in the recommendation system can be used to evaluate the distribution efficiency.

[0125] It is easy for those skilled in the art to understand that in other application scenarios, the number of content publication stages can be set according to specific content publication requirements. For example, the content publication stage can also be divided into four stages, five stages, etc. The present disclosure does not make any special limitations on the specific number of content publication stages.

[0126] The content publication method of the present disclosure obtains initial recommended content and publishes the initial recommended content using an initial distribution method; the number of recommendations of the initial recommended content is less than or equal to a first threshold; monitors the number of recommendations of the initial recommended content, and determines first recommended content according to the initial recommended content, where the first recommended content is the initial recommended content with the number of recommendations greater than the first threshold; determines a first content evaluation parameter corresponding to the first recommended content according to the first user operation data corresponding to the first recommended content; determines second recommended content from the first recommended content according to the first content evaluation parameter, and publishes the second recommended content according to the first content evaluation parameter. On the one hand, in the process of content publication, introducing a phased content evaluation parameter (such as the duration percentile ratio) to distinguish high-quality content from low-quality content can effectively filter out low-quality content, and for high-quality content, the recommendation priority of high-quality content can be further determined to further accelerate the distribution of high-quality content. On the other hand, in different content publication stages, introducing a phased content publication evaluation index, that is, the recommendation efficiency of different publication stages, can effectively determine the distribution effect of each content publication stage. On the other hand, by publishing content in stages, as the number of publication stages increases, the number of recommended content can be gradually reduced, and the number of recommended users can be expanded, which can effectively improve the recommendation efficiency of content and enhance the user experience.

[0127] Exemplary Apparatus

[0128] After introducing the method of the exemplary embodiment of the present disclosure, next, reference is made to Figure 9 to describe the content publication device of the exemplary embodiment of the present disclosure.

[0129] In Figure 9 , the content publication device 900 may include a first content publication module 910, a first content determination module 920, an evaluation parameter determination module 930, and a second content publication module 940. Among them:

[0130] The first content publishing module 910 is configured to obtain initial recommended content and publish the initial recommended content in an initial distribution manner; the number of recommendations of the initial recommended content is less than or equal to a first threshold; the first content determination module 920 is configured to monitor the number of recommendations of the initial recommended content and determine first recommended content based on the initial recommended content, where the first recommended content is the initial recommended content with the number of recommendations greater than the first threshold; the evaluation parameter determination module 930 is configured to determine a first content evaluation parameter corresponding to the first recommended content according to first user operation data corresponding to the first recommended content; the second content publishing module 940 is configured to determine second recommended content from the first recommended content according to the first content evaluation parameter and publish the second recommended content according to the first content evaluation parameter.

[0131] In an embodiment of the present disclosure, the content publishing device further includes an initial content determination module, and the initial content determination module is configured to: obtain initial content and extract content features corresponding to the initial content; determine the feature attention degree of the content features, generate the number of recommendations of the initial content according to the feature attention degree, so as to determine initial recommended content from the initial content according to the number of recommendations.

[0132] In an embodiment of the present disclosure, the first content publishing module includes a first content publishing unit, and the first content publishing unit is configured to: determine a first designated user; the first designated user is a user at a first activity level; in response to a page refresh operation of the first designated user for a first page, determine a target display area in the first page; and display the initial recommended content in the target display area.

[0133] In an embodiment of the present disclosure, the first content evaluation parameter includes a first duration quantile ratio, the evaluation parameter determination module includes an evaluation parameter determination unit, and the evaluation parameter determination unit includes: a first stage determination subunit, configured to use the stage of content publishing in the initial distribution manner as the first publishing stage; a first data acquisition subunit, configured to acquire first user operation data of the first recommended content; the first user operation data is the user operation data corresponding to the first recommended content in the first publishing stage; a first parameter determination subunit, configured to determine a first exposure click-through rate and a first average content browsing duration according to the first user operation data; a first duration determination subunit, configured to determine a first single-exposure browsing duration of the first recommended content in the first publishing stage according to the first exposure click-through rate and the first average content browsing duration; and a first quantile ratio determination subunit, configured to determine the first duration quantile ratio according to the first single-exposure browsing duration.

[0134] In one embodiment of the present disclosure, the first quantile ratio determination subunit is configured to: obtain a set of single-exposure durations; the set of single-exposure durations includes the first single-exposure browsing durations of all the first recommended contents; obtain a quantile parameter, and determine a first duration quantile ratio corresponding to the set of single-exposure durations according to the quantile parameter and the multiple first single-exposure browsing durations.

[0135] In one embodiment of the present disclosure, the second content publishing module includes a second content publishing unit, and the second content publishing unit is configured to: obtain a parameter threshold; the parameter threshold includes a quantile ratio threshold; determine the first recommended contents with the first duration quantile ratio less than the quantile ratio threshold as filtered contents; and determine the first recommended contents with the first duration quantile ratio greater than or equal to the quantile ratio threshold as second recommended contents.

[0136] In one embodiment of the present disclosure, the second content publishing module includes a second content publishing unit, and the second content publishing unit is configured to: determine a second specified user; the second specified user is a user at or above the second activity level; obtain the duration quantile ratio of the second recommended contents, and determine a recommendation coefficient of the second recommended contents according to the duration quantile ratio of the second recommended contents; determine a recommendation priority of the second recommended contents according to the recommendation coefficient of the second recommended contents, and publish the second recommended contents to the second specified users according to the recommendation priority.

[0137] In one embodiment of the present disclosure, the content publishing device further includes a recommendation evaluation module, and the recommendation evaluation module is configured to: determine the total number of contents in the second stage; the total number of contents in the second stage is the number of all contents published according to the first content evaluation parameter; determine the number of second contents; the number of second contents is the number of contents published according to the first content evaluation parameter and with the number of recommendations greater than a second threshold; the second threshold is greater than the first threshold; and determine the recommendation effectiveness rate of the second publishing stage according to the number of second contents and the total number of contents in the second stage.

[0138] In one embodiment of the present disclosure, the content publishing device further includes a third content publishing module, and the third content publishing is configured to: determine the second recommended contents with the number of recommendations greater than the second threshold as third recommended contents; the second threshold is greater than the first threshold; regard the stage of content publishing according to the first content evaluation parameter as the second publishing stage; obtain the second user operation data corresponding to the third recommended contents; the second user operation data is the user operation data corresponding to the third recommended contents in the second publishing stage; determine the second content evaluation parameter corresponding to the third recommended contents according to the second user operation data; determine fourth recommended contents from the third recommended contents according to the second content evaluation parameter, and publish the fourth recommended contents to all users according to the second content evaluation parameter.

[0139] In a third aspect of the embodiments of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the content publishing method described in the first aspect above is implemented.

[0140] Exemplary Medium

[0141] After introducing the apparatus of the exemplary embodiments of the present disclosure, next, reference is made to Figure 10 the storage medium of the exemplary embodiments of the present disclosure is described.

[0142] In some embodiments, each aspect of the present disclosure may also be implemented as a medium on which program code is stored, and when the program code is executed by a processor of a device, it is used to implement the steps in the content publishing method according to various exemplary embodiments of the present disclosure described in the "Exemplary Method" section above of this specification.

[0143] For example, when the processor of the device executes the program code, it may implement steps such as Figure 2 step S210 described in, obtain initial recommended content, and publish the initial recommended content by using an initial distribution method; the number of recommendations of the initial recommended content is less than or equal to a first threshold; step S220, monitor the number of recommendations of the initial recommended content, and determine a first recommended content according to the initial recommended content, where the first recommended content is the initial recommended content with the number of recommendations greater than the first threshold; step S230, determine a first content evaluation parameter corresponding to the first recommended content according to the first user operation data corresponding to the first recommended content; step S240, determine a second recommended content from the first recommended content according to the first content evaluation parameter, and publish the second recommended content according to the first content evaluation parameter.

[0144] Reference is made to Figure 10 shown, a program product 1000 for implementing the above content publishing method or implementing the above content publishing method according to an embodiment of the present disclosure is described, which may adopt a portable compact disc read-only memory (CD-ROM) and include program code, and may run on a terminal device, such as a personal computer. However, the program product of the present disclosure is not limited thereto.

[0145] The program product may adopt any combination of one or more readable media. The readable media may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0146] The computer-readable signal medium may include a data signal propagated in a baseband or as a part of a carrier wave, which carries the readable program code. Such a propagated data signal may take various forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The readable signal medium may also be any readable medium other than the readable storage medium.

[0147] The program code for performing the operations of the present disclosure may be written in any combination of one or more programming languages, including object-oriented programming languages such as Java, C++, etc., and also including conventional procedural programming languages such as the "C" language or similar programming languages. The program code may be executed entirely on the user's computing device, partially on the user's device, partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user's computing device through any type of network, including a local area network (LAN) or a wide area network (WAN).

[0148] Exemplary Computing Device

[0149] After introducing the content publishing method, content publishing apparatus, and storage medium of the exemplary embodiments of the present disclosure, next, reference is made to Figure 11 describe the electronic device of the exemplary embodiments of the present disclosure.

[0150] Those skilled in the art can understand that various aspects of the present disclosure can be implemented as a system, method, or program product. Therefore, various aspects of the present disclosure can be specifically implemented in the following forms, namely: a complete hardware embodiment, a complete software embodiment (including firmware, microcode, etc.), or an embodiment combining hardware and software aspects, which can be collectively referred to as "circuitry", "module", or "system" here.

[0151] In some possible embodiments, the electronic device according to the present disclosure may at least include at least one processing unit and at least one storage unit. Among them, the storage unit stores program code, and when the program code is executed by the processing unit, the processing unit executes the steps in the content publishing method according to various exemplary embodiments of the present disclosure described in the above "Exemplary Method" section of this specification. For example, the processing unit may execute steps such as Figure 2 shown in step S210, obtain initial recommended content, and publish the initial recommended content using an initial distribution method; the number of recommendations of the initial recommended content is less than or equal to a first threshold; step S220, monitor the number of recommendations of the initial recommended content, and determine a first recommended content according to the initial recommended content, where the first recommended content is the initial recommended content with the number of recommendations greater than the first threshold; step S230, determine a first content evaluation parameter corresponding to the first recommended content according to the first user operation data corresponding to the first recommended content; step S240, determine a second recommended content from the first recommended content according to the first content evaluation parameter, and publish the second recommended content according to the first content evaluation parameter.

[0152] Next, refer to Figure 11 to describe the electronic device 1100 according to an exemplary embodiment of the present invention. Figure 11 The electronic device 1100 shown is merely an example and should not impose any limitation on the functions and usage scope of the embodiments of the present invention.

[0153] As Figure 11 shown, the electronic device 1100 is presented in the form of a general-purpose computing device. The components of the electronic device 1100 may include, but are not limited to: the above at least one processing unit 1101, the above at least one storage unit 1102, a bus 1103 connecting different system components (including the storage unit 1102 and the processing unit 1101), and a display unit 1107.

[0154] The bus 1103 represents one or more of several types of bus structures, including a memory bus or a memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the multiple bus structures.

[0155] The storage unit 1102 may include a readable medium in the form of volatile memory, such as a random access memory (RAM) 1121 and / or a cache memory 1122, and may further include a read-only memory (ROM) 1123.

[0156] The storage unit 1102 may also include a program / utility 1125 having a set (at least one) of program modules 1124. Such program modules 1124 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include the implementation of a network environment.

[0157] The electronic device 1100 may also communicate with one or more external devices 1104 (such as a keyboard, a pointing device, a Bluetooth device, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 1100, and / or may communicate with any device that enables the electronic device 1100 to communicate with one or more other computing devices (such as a router, a modem, etc.). Such communication may be carried out through an input / output (I / O) interface 1105. Moreover, the electronic device 1100 may also communicate with one or more networks (such as a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through a network adapter 1106. As shown in the figure, the network adapter 1106 communicates with other modules of the electronic device 1100 through a bus 1103. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 1100, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, etc.

[0158] It should be noted that, although several units / modules or sub-units / modules of the content publishing device are mentioned in the above detailed description, this division is merely exemplary and not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more of the above-described units / modules may be embodied in one unit / modules. Conversely, the features and functions of one unit / modules described above may be further divided and embodied by multiple units / modules.

[0159] In addition, although the operations of the method of the present disclosure are described in a specific order in the drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step for execution, and / or one step may be decomposed into multiple steps for execution.

[0160] Although the spirit and principles of the present disclosure have been described with reference to several specific embodiments, it should be understood that the present disclosure is not limited to the specific embodiments disclosed, and the division of each aspect does not mean that the features in these aspects cannot be combined for benefits. This division is only for the convenience of expression. The present disclosure aims to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.

Claims

1. A content publishing method, characterized in that, Includes: Obtain initial recommended content and publish the initial recommended content using an initial distribution method; The number of recommendations for the initial recommended content is less than or equal to a first threshold; Monitor the number of recommendations for the initial recommended content and determine a first recommended content based on the initial recommended content, where the first recommended content is the initial recommended content for which the number of recommendations is greater than the first threshold; Determine a first content evaluation parameter corresponding to the first recommended content based on first user operation data corresponding to the first recommended content. The first content evaluation parameter is a parameter for evaluating the content quality of the first recommended content. The first content evaluation parameter includes a first duration percentile ratio, and the first duration percentile ratio is the percentile ratio corresponding to the first single-exposure viewing duration. The first single-exposure viewing duration is determined based on the exposure click-through rate and the first content average viewing duration of the first recommended content in a first publishing stage. The first duration percentile ratio is determined based on a percentile parameter and the first single-exposure viewing durations of multiple first recommended contents; Determine a second recommended content from the first recommended contents based on the first content evaluation parameter and publish the second recommended content based on the first content evaluation parameter.

2. The method according to claim 1, wherein Before obtaining the initial recommended content, the method further includes: Obtain initial content and extract content features corresponding to the initial content; Determine the feature attention degree of the content features, generate the number of recommendations for the initial content based on the feature attention degree, and determine the initial recommended content from the initial content based on the number of recommendations.

3. The method according to claim 1, wherein The step of publishing the initial recommended content using the initial distribution method includes: Determine a first designated user; the first designated user is a user at a first activity level; In response to a page refresh operation of the first designated user for a first page, determine a target display area on the first page; Display the initial recommended content in the target display area.

4. The method according to claim 1, characterized in that, The first content evaluation parameter includes a first duration percentile ratio. The step of determining the first content evaluation parameter corresponding to the first recommended content based on the first user operation data corresponding to the first recommended content includes: Regard the stage of content publishing using the initial distribution method as the first publishing stage; Obtain first user operation data for the first recommended content; the first user operation data is the user operation data corresponding to the first recommended content in the first publishing stage; Determine a first exposure click-through rate and a first content average viewing duration based on the first user operation data; Determine the first single-exposure viewing duration of the first recommended content in the first publishing stage based on the first exposure click-through rate and the first content average viewing duration; Determine the first duration percentile ratio based on the first single-exposure viewing duration.

5. The method according to claim 4, wherein The step of determining the first duration percentile ratio based on the first single-exposure viewing duration includes: Obtain a set of single-exposure durations; the set of single-exposure durations includes the first single-exposure viewing durations of all the first recommended contents; Obtain the quantile parameter, and determine the first duration quantile ratio corresponding to the set of single-exposure durations according to the quantile parameter and multiple first single-exposure viewing durations.

6. The method according to claim 5, characterized in that, The determining the second recommended content from the first recommended content according to the first content evaluation parameter includes: Obtain a parameter threshold; the parameter threshold includes a quantile ratio threshold; Determine the first recommended content with the first duration quantile ratio less than the quantile ratio threshold as filtered content; Determine the first recommended content with the first duration quantile ratio greater than or equal to the quantile ratio threshold as the second recommended content.

7. The method according to claim 1, wherein The publishing the second recommended content according to the first content evaluation parameter includes: Determine a second designated user; the second designated user is a user above the second activity level; Obtain the duration quantile ratio of the second recommended content, and determine the recommendation coefficient of the second recommended content according to the duration quantile ratio of the second recommended content; Determine the recommendation priority of the second recommended content according to the recommendation coefficient of the second recommended content, and publish the second recommended content to the second designated user according to the recommendation priority.

8. The method according to claim 1 or 7, characterized in that, After the publishing the second recommended content according to the first content evaluation parameter, the method further includes: Determine the total number of second-stage content; the total number of second-stage content is the number of all content published according to the first content evaluation parameter; Determine the number of second content; the number of second content is the number of content published according to the first content evaluation parameter and with the number of recommendations greater than a second threshold; the second threshold is greater than the first threshold; Determine the recommendation effectiveness rate of the second publishing stage according to the number of second content and the total number of second-stage content.

9. The method according to claim 1 or 7, characterized in that, After the publishing the second recommended content according to the first content evaluation parameter, the method further includes: Determine the third recommended content as the second recommended content with the number of recommendations greater than the second threshold; the second threshold is greater than the first threshold; Take the stage of content publishing according to the first content evaluation parameter as the second publishing stage; Obtain the second user operation data corresponding to the third recommended content; the second user operation data is the user operation data corresponding to the third recommended content in the second publishing stage; Determine the second content evaluation parameter corresponding to the third recommended content according to the second user operation data; Determine the fourth recommended content from the third recommended content according to the second content evaluation parameter, and publish the fourth recommended content to all users according to the second content evaluation parameter.

10. A content publishing device, characterized in that, Includes: A first content publishing module, configured to obtain initial recommended content and publish the initial recommended content using an initial distribution method; The number of recommendations of the initial recommended content is less than or equal to a first threshold; A first content determination module, configured to monitor the number of recommendations of the initial recommended content and determine the first recommended content according to the initial recommended content, where the first recommended content is the initial recommended content with the number of recommendations greater than the first threshold; An evaluation parameter determination module, configured to determine a first content evaluation parameter corresponding to the first recommended content according to the first user operation data corresponding to the first recommended content, where the first content evaluation parameter is a parameter for evaluating the content quality of the first recommended content, and the first content evaluation parameter includes a first duration percentile ratio, where the first duration percentile ratio is the percentile ratio corresponding to the first single-exposure browsing duration, and the first single-exposure browsing duration is determined according to the exposure click-through rate and the first content average browsing duration of the first recommended content in the first release stage, and the first duration percentile ratio is determined according to the percentile parameter and the first single-exposure browsing durations of multiple first recommended contents; A second content release module, configured to determine a second recommended content from the first recommended contents according to the first content evaluation parameter, and release the second recommended content according to the first content evaluation parameter.

11. The device according to claim 10, characterized in that, The content release device further includes an initial content determination module, and the initial content determination module is configured to: Obtain initial content, and extract the content features corresponding to the initial content; Determine the feature attention degree of the content features, generate the recommended number of the initial content according to the feature attention degree, and determine the initial recommended content from the initial content according to the recommended number.

12. The device according to claim 10, characterized in that, The first content release module includes a first content release unit, and the first content release unit is configured to: Determine a first designated user; the first designated user is a user in the first activity level; In response to a page refresh operation of the first designated user on a first page, determine a target display area on the first page; Display the initial recommended content in the target display area.

13. The device according to claim 10, characterized in that, The first content evaluation parameter includes a first duration percentile ratio, the evaluation parameter determination module includes an evaluation parameter determination unit, and the evaluation parameter determination unit includes: A first stage determination subunit, configured to use the stage of content release using the initial distribution method as the first release stage; A first data acquisition subunit, configured to acquire the first user operation data of the first recommended content; the first user operation data is the user operation data corresponding to the first recommended content in the first release stage; A first parameter determination subunit, configured to determine a first exposure click-through rate and a first content average browsing duration according to the first user operation data; A first duration determination subunit, configured to determine the first single-exposure browsing duration of the first recommended content in the first release stage according to the first exposure click-through rate and the first content average browsing duration; A first percentile ratio determination subunit, configured to determine the first duration percentile ratio according to the first single-exposure browsing duration.

14. The device according to claim 13, characterized in that, The first percentile ratio determination subunit is configured to: Obtain a set of single-exposure durations; the set of single-exposure durations includes the first single-exposure browsing durations of all the first recommended contents; Obtain a percentile parameter, and determine the first duration percentile ratio corresponding to the set of single-exposure durations according to the percentile parameter and multiple first single-exposure browsing durations.

15. The device according to claim 14, characterized in that, The second content publishing module includes a second content publishing unit, and the second content publishing unit is configured to: Obtain a parameter threshold; the parameter threshold includes a quantile ratio threshold; Determine the first recommended content with a first duration quantile ratio less than the quantile ratio threshold as filtered content; Determine the first recommended content with a first duration quantile ratio greater than or equal to the quantile ratio threshold as the second recommended content.

16. The device according to claim 10, wherein The second content publishing module includes a second content publishing unit, and the second content publishing unit is configured to: Determine a second designated user; the second designated user is a user at or above the second activity level; Obtain the duration quantile ratio of the second recommended content, and determine the recommendation coefficient of the second recommended content according to the duration quantile ratio of the second recommended content; Determine the recommendation priority of the second recommended content according to the recommendation coefficient of the second recommended content, and publish the second recommended content to the second designated user according to the recommendation priority.

17. The device according to claim 10 or 16, characterized in that, The content publishing device further includes a recommendation evaluation module, and the recommendation evaluation module is configured to: Determine the total number of contents in the second stage; the total number of contents in the second stage is the number of all contents published according to the first content evaluation parameter; Determine the number of second contents; the number of second contents is the number of contents published according to the first content evaluation parameter and with a recommendation number greater than a second threshold; The second threshold is greater than the first threshold; Determine the recommendation effectiveness rate of the second publishing stage according to the number of second contents and the total number of contents in the second stage.

18. The device according to claim 10 or 16, characterized in that The content publishing device further includes a third content publishing module, and the third content publishing is configured to: Determine the second recommended content with a recommendation number greater than the second threshold as the third recommended content; The second threshold is greater than the first threshold; Regard the stage of content publishing according to the first content evaluation parameter as the second publishing stage; Obtain the second user operation data corresponding to the third recommended content; The second user operation data is the user operation data corresponding to the third recommended content in the second publishing stage; Determine the second content evaluation parameter corresponding to the third recommended content according to the second user operation data; Determine the fourth recommended content from the third recommended content according to the second content evaluation parameter, and publish the fourth recommended content to all users according to the second content evaluation parameter.

19. An electronic device, characterized in that, Including: A processor; And A memory, on which computer-readable instructions are stored, and when the computer-readable instructions are executed by the processor, the content publishing method described in any one of claims 1 to 9 is implemented.

20. A computer-readable medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the content publishing method described in any one of claims 1 to 9 is implemented.

Images