Converged media content updating system and method for realizing intelligent push based on effect feedback
By collecting and analyzing user feedback data and media tags and evaluating user interest preferences, precise push of integrated media content is achieved, and the problem of inability to measure different types of content in the existing technology is solved, and the accuracy and effectiveness of push is improved.
Patent Information
- Application Number
- CN202411433518.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-15
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2044-10-15
AI Technical Summary
The existing integrated media content push update technology cannot measure different types of integrated media content through unified standards, resulting in the inability to accurately push content that users are interested in.
By collecting users' feedback data and media tags on integrated media content, conducting quantitative analysis, evaluating users' browsing proportion among different media types, matching the best push content and updating the content library.
It improves the accuracy and effectiveness of integrated media content push, can push content based on user interest preferences, and expand user interests and understand hot events through real-time popularity.
Smart Images

Figure CN118939818B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of converged media content push and update, and specifically to a converged media content update system and method for realizing intelligent push based on effect feedback. Background Art
[0002] Converged media content push and update technology refers to the use of multiple communication media, such as text, pictures, videos and audio, to push updated content to audiences on different platforms and channels. This technology combines the characteristics of traditional media and emerging digital media. Through intelligent algorithms, personalized recommendation systems and real-time update mechanisms, it can achieve accurate content push and continuous updating based on user interests, preferences and behavioral data, so as to enhance user experience and increase content attractiveness, and ultimately achieve a wider communication effect and influence.
[0003] Converged media is a new type of media that shares information and disseminates information based on the properties of the media. Simply put, it realizes multiple media forms, such as the dissemination of video, audio, text, and pictures in newspapers, Weibo, WeChat public accounts and other different media. However, due to the diversity of resource types in converged media, it is difficult to measure them through unified standards and find the best push solution. For example, in a patent application with publication number CN117009699A, a method for optimizing a resource management mechanism based on converged media is disclosed. This solution cannot push converged media content that users are interested in to users. It will be pushed directly after passing the review. The existing converged media content push and update technology still cannot push and measure different types of converged media content through unified standards, resulting in the inability to accurately push converged media content that users are interested in to users. Summary of the invention
[0004] The present invention aims to solve one of the technical problems in the prior art to at least a certain extent, by collecting users' feedback data on integrated media content and media tags of integrated media content, and then quantifying the feedback data based on users' feedback data on browsing integrated media content, and then analyzing the proportion of different content tags browsed by users in different media types based on browsing time, and then evaluating users' interest preferences, matching the best push content for users based on their interest preferences, and finally updating users' content libraries based on the push content, so as to solve the problem that the existing integrated media content push update technology still cannot measure the push of different types of integrated media content through unified standards, resulting in the inability to accurately push integrated media content that users are interested in to users.
[0005] To achieve the above objectives, in a first aspect, the present application provides a method for updating integrated media content based on effect feedback and intelligent push, comprising the following steps:
[0006] Collect user feedback data on converged media content and media tags of converged media content;
[0007] Analyze feedback data and media tags to evaluate user interests and preferences;
[0008] Match the best push content to users based on their interests and preferences;
[0009] Update the user's content library based on the pushed content.
[0010] Furthermore, collecting user feedback data on the converged media content and media tags of the converged media content includes the following sub-steps:
[0011] Collect user feedback data on converged media content, including whether to like, share, browsing time, and number of single content browsing times;
[0012] When collecting feedback data of the converged media content, media tags of the converged media content are simultaneously obtained, wherein the media tags include media categories and content tags;
[0013] The media categories include video, audio, text and pictures;
[0014] Detect whether there is a preset tag in the integrated media content. If so, output a tag preset signal; if not, output a tag extraction signal;
[0015] If a tag preset signal is output, the preset tag is used as the content tag of the integrated media content; at the same time, the title of the integrated media content is obtained and named as the content title, keywords are extracted from the content title to obtain sample keywords, a keyword sample library is established, and the content tags and corresponding sample keywords are collected;
[0016] If a tag extraction signal is output, the content title of the converged media content is obtained, and keywords are extracted from the content title to obtain title keywords;
[0017] Name the content tag in the keyword sample library as a sample tag, search for the sample tag where the title keyword is located in the keyword sample library, and name it a search tag;
[0018] For all search tags of converged media content, the number of identical search tags is counted and named as the number of tags. The maximum value among the number of search tags is found and the corresponding search tag is marked as a content tag.
[0019] Further, analyzing the feedback data and media tags to evaluate the user's interest preferences includes the following sub-steps:
[0020] Quantify the feedback data based on the feedback data of users browsing the converged media content;
[0021] Based on browsing time, analyze the proportion of different content tags browsed by users in different media types;
[0022] Evaluate user interest preferences.
[0023] Furthermore, based on the feedback data of users browsing the converged media content, quantifying the feedback data includes the following sub-steps:
[0024] For any converged media content browsed by the user, detect whether the number of views of the single content is one, if so, output a single view signal; if not, output a multiple view signal;
[0025] If a single browsing signal is output, the number of single content browsing of the integrated media content by the user is set to 1; if a multiple browsing signal is output, the number of single content browsing of the integrated media content by the user is set to 2;
[0026] Based on the single content views and media tags of the integrated media content by users, the sum of the single content views of different content tags by users is counted, and marked as tag views; at the same time, the sum of the single content views of different media types by users is counted, and marked as type views;
[0027] Based on whether the user likes the converged media content in the feedback data and the media tags corresponding to the converged media content, the number of likes of different content tags by users is counted, marked as tag likes. If the number of single content views of the converged media content by the user is 2, the number of tag likes will increase by one; the number of likes of different media types by users is counted, marked as type likes. If the number of single content views of the converged media content by the user is 2, the type likes will increase by one;
[0028] Based on whether the user shares the integrated media content in the feedback data and the media tags corresponding to the integrated media content, the user's sharing volume of different content tags is counted and marked as tag sharing volume. If the user's single content views of the integrated media content are 2, the tag sharing volume will be increased by one; the user's sharing volume of different media types is counted and marked as type sharing volume. If the user's single content views of the integrated media content are 2, the type sharing volume will be increased by one.
[0029] Furthermore, based on the browsing time, analyzing the proportion of different content tags browsed by users in different media types includes the following sub-steps:
[0030] Mark the converged media content whose media type is video or audio as progress bar content, obtain the video duration or audio duration of the progress bar content, and mark it as the progress bar duration;
[0031] Calculate the browsing time of the progress bar content that the user browses divided by the progress bar time, mark the calculation result as the single content browsing rate, and based on different content tags, count the sum of the single content browsing rates of all progress bar contents in the content tags, and mark it as the tag browsing ratio, where the tag browsing ratio includes the tag video browsing ratio and the tag audio browsing ratio; obtain several tag browsing ratios by counting the tag browsing ratios of different content tags;
[0032] Mark the converged media content whose media type is text and picture as page content, name the bottom of the scroll bar in the page content as browsing limit, and the scroll bar is in the scroll box;
[0033] Divide the scroll box into two parts, upper and lower, according to the browsing limit, mark the part above the browsing limit as the browsed part, obtain the length of the scroll box and the length of the browsed part, and mark them as the total content length and the browsed length respectively;
[0034] Calculate the browsed length divided by the total content length, and mark the calculation result as the single content viewing rate. Based on different content tags, count the sum of the single content viewing rates of all page contents in the content tags, and mark it as the tag viewing ratio. The tag viewing ratio includes the tag text viewing ratio and the tag image viewing ratio. By counting the tag viewing ratios of different content tags, several tag viewing ratios are obtained.
[0035] Furthermore, evaluating the user's interest preference includes the following sub-steps:
[0036] Obtain type likes, type shares and type views, wherein the type likes include video likes, audio likes, text likes and picture likes; the type shares include video shares, audio shares, text shares and picture shares; the type views include video views, audio views, text views and picture views;
[0037] According to the order of n from small to large, through the symbol V n represents the number of video views, video likes, and video shares, respectively, where n=1, 2, or 3; symbol A n It represents the number of audio views, audio likes and audio shares respectively; the symbol W n It represents the number of text views, the number of text likes, and the number of text shares respectively; the symbol P n It represents the number of image views, image likes, and image shares respectively;
[0038] The symbol VP represents the percentage of video views, the symbol AP represents the percentage of audio views, the symbol WP represents the percentage of text views, and the symbol PP represents the percentage of image views.
[0039] By formula Calculate the video push index, where VI is the video push index, and the simplified formula is: ; Similarly, through the formula Calculate the audio push index through the formula Calculate the text push index through the formula Calculate image push indicators;
[0040] Then calculate VI+AI+WI+PI, mark the result as the total push index, represented by the symbol TI, set the interest push ratio, represented by the symbol TX, calculate VI / TI×TX to get the video push ratio, calculate AI / TI×TX to get the audio push ratio, calculate WI / TI×TX to get the text push ratio, and calculate PI / TI×TX to get the picture push ratio;
[0041] The video push ratio, audio push ratio, text push ratio and picture push ratio are user's interest preferences for media types, named as type preferences;
[0042] Based on tag views, tag likes, tag shares, tag views ratio, and tag viewing ratio, users’ interest preferences for content tags are analyzed and named tag preference.
[0043] Furthermore, analyzing the user's interest preference for content tags based on tag views, tag likes, tag shares, tag views ratio, and tag viewing ratio includes the following sub-steps:
[0044] Calculate the push priority of different content tags through the push priority calculation formula;
[0045] The push priority calculation formula is configured as:
[0046]
[0047]
[0048]
[0049]
[0050] Among them, FV is the video push priority, TV is the tag views, TL is the tag likes, TS is the tag shares, FA is the audio push priority, FW is the text push priority, FP is the picture push priority, α is the like coefficient, and β is the share coefficient;
[0051] Sort FV in descending order to get the video tag push sequence; Sort FA in descending order to get the audio tag push sequence; Sort FW in descending order to get the text tag push sequence; Sort FP in descending order to get the picture tag push sequence;
[0052] The video tag push sequence, the audio tag push sequence, the text tag push sequence, and the picture tag push sequence are tag preferences.
[0053] Furthermore, matching the best push content for the user based on the user's interest preference includes the following sub-steps:
[0054] Get the number of converged media contents that need to be updated for the next refresh, named update number;
[0055] Calculate the number of updates and multiply it by the video push ratio to get the number of video pushes; calculate the number of updates and multiply it by the audio push ratio to get the number of audio pushes; calculate the number of updates and multiply it by the text push ratio to get the number of text pushes; calculate the number of updates and multiply it by the image push ratio to get the number of image pushes;
[0056] When updating the pushed integrated media content, the media type of the video push number pushed to the user is the integrated media content of the video, the media type of the audio push number pushed is the integrated media content of the audio, the media type of the text push number pushed is the integrated media content of the text, and the media type of the picture push number pushed is the integrated media content of the picture;
[0057] Mark the integrated media content of the first video push number in the video push sequence as pushed content; mark the integrated media content of the first audio push number in the audio push sequence as pushed content; mark the integrated media content of the first text push number in the text push sequence as pushed content; mark the integrated media content of the first picture push number in the picture push sequence as pushed content.
[0058] Further, updating the user's content library based on the pushed content includes the following sub-steps:
[0059] Calculate the sum of the number of video pushes, audio pushes, text pushes, and image pushes, then subtract their sum from the number of updates, and mark the result as the number of expanded pushes;
[0060] Calculate the real-time popularity of all converged media content, and mark the converged media content with the highest number of pushes ranked in real-time popularity as push-expanded content;
[0061] The calculation formula of the real-time heat is: , where RTP is the real-time popularity, NC is the number of comments on the converged media content, NL is the number of likes on the converged media content, NR is the number of reposts of the converged media content, and NV is the number of views of the converged media content. For the like and forwarding coefficient, is the comment coefficient;
[0062] The number of comments, likes, reposts and views are all directly obtained through the integrated media platform;
[0063] When the user needs to update the integrated media content, the pushed content and the pushed extended content will be updated to the user's integrated media platform.
[0064] In the second aspect, the present application provides a media content update system for realizing intelligent push based on effect feedback, including a platform data collection module, an interest preference evaluation module, a push content matching module and a push content update module; the platform data collection module, the interest preference evaluation module and the push content update module are respectively connected to the push content matching module data;
[0065] The platform data collection module is used to collect user feedback data on converged media content and media tags of converged media content;
[0066] The interest preference evaluation module is used to analyze the feedback data and media tags to evaluate the user's interest preferences;
[0067] The push content matching module is used to match the best push content for the user based on the user's interest preferences;
[0068] The push content update module is used to update the user's content library based on the push content.
[0069] Beneficial effects of the present invention: The present invention collects user feedback data on converged media content and media tags of converged media content, and then quantifies the feedback data based on the feedback data of users browsing the converged media content, and then analyzes the proportion of different content tags browsed by users in different media types based on the browsing time. The advantage is that converged media includes various types of resources. In addition to basic feedback data such as basic browsing volume and comment volume, the browsing degree of users on converged media content also has a great weight in the user's interest analysis. The degree of interest represented by users browsing a part of the content is different from that represented by browsing all the content. Therefore, by analyzing the proportion of different content tags browsed by users in different media types, the basic feedback data can be combined to evaluate the user's interest preference for converged media content of different data types through a unified standard, thereby improving the accuracy and effectiveness of converged media content push;
[0070] The present invention evaluates the user's interest preferences, matches the best pushed content for the user based on the user's interest preferences, and finally updates the user's content library based on the pushed content. The advantage is that when pushing content to the user, most of the content is pushed according to the user's interest preferences, and a small part of the page space is reserved for recommending popular integrated media content to the user according to the real-time popularity, so as to achieve the purpose of expanding the user's interest preferences and understanding hot events, and improve the effectiveness and rationality of integrated media content push. BRIEF DESCRIPTION OF THE DRAWINGS
[0071] Figure 1 is a functional block diagram of the system of the present invention;
[0072] Figure 2 A schematic diagram of the definition of a scroll bar and a scroll box of the present invention;
[0073] Figure 3 is a schematic diagram of another scroll bar of the present invention;
[0074] Figure 4 is a flow chart of the steps of the method of the present invention;
[0075] Figure 5 It is a schematic structural diagram of the electronic device of the present invention. DETAILED DESCRIPTION
[0076] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0077] Example 1, please refer to Figure 1 As shown, the present application provides a media content update system for realizing intelligent push based on effect feedback, including a platform data collection module, an interest preference evaluation module, a push content matching module and a push content update module; the platform data collection module, the interest preference evaluation module and the push content update module are respectively connected to the push content matching module data;
[0078] The platform data collection module is used to collect user feedback data on converged media content and media tags of converged media content;
[0079] The platform data collection module is configured with a platform data collection strategy, which includes:
[0080] Collect user feedback data on converged media content, including whether to like, share, browsing time, and number of single content browsing times;
[0081] When collecting feedback data on converged media content, media tags of the converged media content are also obtained. The media tags include media categories and content tags.
[0082] Media categories include video, audio, text, and images;
[0083] Detect whether there is a preset tag in the integrated media content. If so, output a tag preset signal; if not, output a tag extraction signal;
[0084] If a tag preset signal is output, the preset tag is used as the content tag of the integrated media content; at the same time, the title of the integrated media content is obtained and named as the content title, keywords are extracted from the content title to obtain sample keywords, a keyword sample library is established, and the content tags and corresponding sample keywords are collected;
[0085] In practical applications, for any integrated media content, such as a funny video, the user's feedback data for this funny video is detected as (liked, not shared, viewed for 186s, viewed once), and the media category of this integrated media content is obtained as video. The detected integrated media content has preset tags, and the preset tag signal is output. The preset tags of the integrated media content are obtained as "funny" and "cute pet", that is, the content tags are "funny" and "cute pet"; at the same time, the content title is obtained as "Award of Animal Confusing Behavior", and the sample keywords extracted by the existing keyword extraction technology are "animal", "confusing", "behavior", "big" and "reward", and a keyword sample library is established. Some data of the keyword sample library are shown in Table 1 below:
[0086] Table 1 Partial data of keyword sample library
[0087]
[0088] The different content tags and corresponding sample keywords shown in Table 1 are only partial data, which are intended to show the composition of the keyword sample library. The symbol "|" is a separator to distinguish different sample keywords;
[0089] If a tag extraction signal is output, the content title of the converged media content is obtained, and keywords are extracted from the content title to obtain title keywords;
[0090] Name the content tag in the keyword sample library as a sample tag, search for the sample tag where the title keyword is located in the keyword sample library, and name it a search tag;
[0091] For all search tags of converged media content, count the number of identical search tags, name them as tag numbers, find the maximum value among the tag numbers, and mark the corresponding search tag as a content tag;
[0092] In actual applications, in order to further explain the specific processing process when outputting label extraction signals, this embodiment assumes that the integrated media content watched by the user does not have a preset label, and the title keywords extracted from the content title are "animals", "confusion", "behavior", "big" and "reward". Through searching, it is found that "animals" are in the search tag "cute pets" and "confusion" is in the search tag "funny". The remaining title keywords have no matching search tags, that is, the number of tags for the search tags "cute pets" and "funny" are both 1, and the maximum value is also 1, so they are marked as content tags at the same time.
[0093] The interest preference evaluation module is used to analyze feedback data and media tags to evaluate the user's interest preferences; the interest preference evaluation module includes a data quantification unit, a proportion analysis unit, a type preference analysis unit, and a tag preference analysis unit;
[0094] The data quantification unit is used to quantify the feedback data based on the feedback data of the user browsing the converged media content;
[0095] The data quantization unit is configured with a data quantization strategy, which includes:
[0096] For any converged media content browsed by the user, detect whether the number of views of the single content is one, if so, output a single view signal; if not, output a multiple view signal;
[0097] If a single browsing signal is output, the number of single content browsing of the integrated media content by the user is set to 1; if a multiple browsing signal is output, the number of single content browsing of the integrated media content by the user is set to 2;
[0098] In actual applications, when a user browses the same converged media content multiple times, there are two situations: one is that the user is interested in the converged media content and browses it repeatedly; the other is that the user stays on the same converged media content interface due to other things, and the converged media content is automatically looped and played back multiple times, resulting in multiple browsing. In order to reduce the impact of the second situation, the user's interest preference is evaluated through the user's multiple browsing behavior, and they are uniformly set to 2, which not only reduces the impact of automatic loop playback, but also increases the impact of users actively browsing multiple times. The probability of automatic loop playback is small, and the impact of setting their single content browsing volume to 2 is small and can be ignored. In this embodiment, the number of times the user browses the converged media content is 1, a single browsing signal is output, and the single content browsing volume is set to 1;
[0099] Based on the single content views and media tags of the integrated media content by users, the sum of the single content views of different content tags by users is counted, and marked as tag views; at the same time, the sum of the single content views of different media types by users is counted, and marked as type views;
[0100] In actual applications, statistics show that the number of tag views of the integrated media content with the content tag "funny" by users is 288. Since there are too many content tags, they will not be specifically displayed in this embodiment, and only "funny" is used as an example for illustration; statistics show that the number of type views of the integrated media content with the media type of "video" by users is 438. Similarly, the number of type views of "audio", "text" and "picture" are 32, 78 and 74 respectively;
[0101] Based on whether the user likes the converged media content in the feedback data and the media tags corresponding to the converged media content, the number of likes of different content tags by users is counted, marked as tag likes. If the number of single content views of the converged media content by the user is 2, the number of tag likes will increase by one; the number of likes of different media types by users is counted, marked as type likes. If the number of single content views of the converged media content by the user is 2, the type likes will increase by one;
[0102] Based on the feedback data of whether users share the converged media content and the media tags corresponding to the converged media content, the number of users sharing different content tags is counted, marked as tag sharing amount. If the number of single content views of the converged media content by the user is 2, the tag sharing amount is increased by 1; the number of users sharing different media types is counted, marked as type sharing amount. If the number of single content views of the converged media content by the user is 2, the type sharing amount is increased by 1;
[0103] In actual applications, it is detected whether the user has liked the integrated media content with the content label "funny". If the feedback data shows that the user has liked it, the label likes will be increased by one. At the same time, it is detected whether the user's single content views for this integrated media content are 2. If so, the label likes will be increased by one, that is, the label likes provided by this integrated media content are 2, and the final label likes are 212. Similarly, the label sharing of the user for the integrated media content with the content label "funny" is 138. When the single content views are 2, increasing the label likes and label sharing is to increase the influence of this integrated media content. At this time, it is difficult to judge whether the user watches multiple times because of interest or because of other factors that cause automatic repeated playback, but if the user likes or shares it, then it can be confirmed that the user is interested in this integrated media content, so the number of tag likes and tag shares can be increased to achieve the purpose of increasing influence; the same is true when analyzing the type likes and type shares, it can be judged which different media types are more likely to interest users; statistics show that the type likes of integrated media content with the media type of "video" are 279, and the type shares are 177. For other media types, the type likes and type shares are no longer specifically displayed in this embodiment.
[0104] The proportion analysis unit is used to analyze the proportion of different content tags browsed by users in different media types based on browsing time;
[0105] The proportion analysis unit is configured with a proportion analysis strategy, which includes:
[0106] Mark the converged media content whose media type is video or audio as progress bar content, obtain the video duration or audio duration of the progress bar content, and mark it as the progress bar duration;
[0107] Calculate the browsing time of the progress bar content that the user browses divided by the progress bar time, and mark the calculation result as the single content browsing rate. Based on different content tags, count the sum of the single content browsing rates of all progress bar contents in the content tags, and mark it as the tag browsing ratio. The tag browsing ratio includes the tag video browsing ratio and the tag audio browsing ratio. By counting the tag browsing ratios of different content tags, several tag browsing ratios are obtained.
[0108] In actual applications, since both "video" and "audio" have progress bars, the user's browsing progress of the integrated media content can be directly judged by the progress bar; the integrated media content listed for the first time in this embodiment is the integrated media content of the "video" type, so it is marked as progress bar content, and the progress bar duration is 213s, while the user's browsing duration is 186s. The single content browsing degree is calculated to be 0.87, and the calculation result is retained to two decimal places. Statistics show that the label video browsing ratio of the progress bar content with the content label "funny" is 172.88, and the label audio browsing ratio is 4.33; in addition, there are also some data such as the label browsing ratio of the content label "cute pet" and the label browsing ratio of the content label "food". In this embodiment, only the label browsing ratio of the content label "funny" is listed to explain the specific processing process;
[0109] See also Figure 2 as well as Figure 3 As shown, the converged media content with the media type of text and picture is marked as page content, the bottom of the scroll bar in the page content is named browsing limit, and the scroll bar is in the scroll box;
[0110] Divide the scroll box into two parts, upper and lower, according to the browsing limit, mark the part above the browsing limit as the browsed part, obtain the length of the scroll box and the length of the browsed part, and mark them as the total content length and the browsed length respectively;
[0111] Calculate the browsed length divided by the total content length, and mark the result as the single content viewing rate. Based on different content tags, count the sum of the single content viewing rates of all page contents in the content tags, and mark it as the tag viewing rate. The tag viewing rate includes the tag text viewing rate and the tag image viewing rate. By counting the tag viewing rates of different content tags, several tag viewing rates are obtained.
[0112] In actual applications, the definitions of scroll bars and scroll boxes are as follows Figure 2 As shown, the browsing limit, total content length and browsed length are all in Figure 2 In the text editor, the content of the text and picture types are marked; since there is no progress bar for the "text" type and the "picture" type, the conventional progress bar method of judging the browsing progress cannot be implemented. Whether it is an APP or a web page, scroll bars are set for the "text" type and "picture" type of integrated media content. Although the types of scroll bars are different, the principles are the same, such as Figure 3 Another type of scroll bar, which usually displays content in the form of page turning, can directly obtain the number of pages viewed and the total number of pages for calculation to obtain the tag viewing ratio, for example Figure 3 The gray part represents the browsed part, which is 1 / 4 of the scroll box, that is, the tag viewing ratio is 0.25; Figure 2 For example, the scroll bar in Figure 2 The media type of the integrated media content is "text", and the content label is "funny". The browsed length is 6cm and the total content length is 18cm. The single content viewing rate is 0.33 through calculation. The calculation result is rounded to two decimal places. The statistical calculation shows that the viewing rate of the label text of the integrated media content with the content label "funny" is 10.38%. Similarly, the viewing rate of the label image of the integrated media content with the content label "funny" is 11.42.
[0113] The type preference analysis unit is configured with a type preference analysis strategy, which includes:
[0114] Get type likes, type shares and type views. Type likes include video likes, audio likes, text likes and picture likes; type shares include video shares, audio shares, text shares and picture shares; type views include video views, audio views, text views and picture views;
[0115] According to the order of n from small to large, through the symbol V n represents the number of video views, video likes, and video shares, respectively, where n=1, 2, or 3; symbol A n It represents the number of audio views, audio likes and audio shares respectively; the symbol W nIt represents the number of text views, the number of text likes, and the number of text shares respectively; the symbol P n It represents the number of image views, image likes, and image shares respectively;
[0116] In actual applications, due to the excessive amount of data, this embodiment only displays the video views, video likes and video shares and calculates the video push indicators, aiming to explain the specific analysis and calculation process, and the calculation process of other types of push indicators is omitted; the video views V1 is obtained to be 438, the video likes V2 is 279, and the video shares V3 is 177, that is, the type views, type likes and type shares of the integrated media content with the media type of "video" calculated in this embodiment;
[0117] The symbol VP represents the percentage of video views, the symbol AP represents the percentage of audio views, the symbol WP represents the percentage of text views, and the symbol PP represents the percentage of image views.
[0118] By formula Calculate the video push index, where VI is the video push index, and the simplified formula is: ; Similarly, through the formula Calculate the audio push index through the formula Calculate the text push index through the formula Calculate image push indicators;
[0119] Then calculate VI+AI+WI+PI, mark the result as the total push index, represented by the symbol TI, set the interest push ratio, represented by the symbol TX, calculate VI / TI×TX to get the video push ratio, calculate AI / TI×TX to get the audio push ratio, calculate WI / TI×TX to get the text push ratio, and calculate PI / TI×TX to get the picture push ratio;
[0120] The video push ratio, audio push ratio, text push ratio, and picture push ratio are user interest preferences for media types, named type preferences;
[0121] In actual applications, the video push index VI is calculated to be 78833.28, the audio push index AI is calculated to be 129.9, the text push index WI is 829.68, the picture push index PI is 1039.22, and the total push index TI is further calculated to be 80895.08; the interest push ratio is set to leave a certain space to push real-time popular integrated media content to users. On the one hand, it can expand the user's interest preferences. For example, if the user has not browsed this type of integrated media content before, but after the push, it is found that the user is very interested, then the relevant content can be pushed again. On the other hand, even if the user is not interested, he can still understand the real-time popularity. For higher events, the interest push ratio in this embodiment is set to 80%. The calculation results show that the video push ratio is 0.78, the audio push ratio is 0.00, the text push ratio is 0.01, and the picture push ratio is 0.01. The calculation results retain two decimal places, and the ones less than 0.01 are calculated as 0.01, that is, the audio push ratio is 0.01. It can be seen from the type views provided in this embodiment that users are very interested in video-type integrated media content, while the views of other types of integrated media content are far less than the "video" type. Therefore, this extreme type preference is generated. If the user's type views are more balanced, the type preference actually calculated will not be so extreme.
[0122] The tag preference analysis unit is used to analyze the user's interest preference for content tags based on tag views, tag likes, tag shares, tag views ratio, and tag viewing ratio, which is named tag preference;
[0123] The label preference analysis unit is configured with a label preference analysis strategy, which includes:
[0124] Calculate the push priority of different content tags through the push priority calculation formula;
[0125] The push priority calculation formula is configured as:
[0126]
[0127]
[0128]
[0129]
[0130] Among them, FV is the video push priority, TV is the tag views, TL is the tag likes, TS is the tag shares, FA is the audio push priority, FW is the text push priority, FP is the picture push priority, α is the like coefficient, and β is the share coefficient;
[0131] In practical applications, the value of α needs to satisfy α×TL greater than or equal to half of TV, and the value of β needs to satisfy β×TS greater than or equal to half of TV. The purpose is to increase the weight of tag likes and tag shares in the priority calculation process. Usually, the tag likes and tag shares are much smaller than the tag views, and in most cases they are less than half. However, the role of tag likes and tag shares in the interest analysis process is greater than that of tag views. Therefore, α and β are set so that the sum of the weights of tag likes and tag shares can be greater than or equal to tag views. First, let α and β be equal to 1, and then determine whether they can make β×TS greater than or equal to half of TV and α×TL greater than or equal to T in different content tags. Half of V, when all are satisfied, the values of α and β are fixed, if not satisfied, α and β are increased by one; FV, FA, FW and FP are the push priorities of the same content tag in different media types; α and β in this embodiment are 2 and 3; this embodiment takes the content tag "funny" as an example, the tag browsing volume TV is 288, the tag like volume TL is 212, the tag sharing volume TS is 138, and the tag video browsing ratio VP is 172.88. By calculation, the video push priority of the content tag "funny" is 194662.88. Similarly, the audio push priority is 4875.58, the text push priority is 11687.88, and the picture push priority is 12858.92;
[0132] Sort FV in descending order to get the video tag push sequence; Sort FA in descending order to get the audio tag push sequence; Sort FW in descending order to get the text tag push sequence; Sort FP in descending order to get the picture tag push sequence;
[0133] The video tag push sequence, the audio tag push sequence, the text tag push sequence, and the image tag push sequence are the tag preferences;
[0134] In actual applications, the FVs of various content tags such as "funny", "cute pets", "food", "beautiful scenery" and "travel" are calculated, and they are sorted in order from large to small, and the video tag push sequence is "'funny', 'cute pets', 'travel', 'food', 'beautiful scenery'". The content tags are not limited to the five listed in this embodiment. This embodiment only uses them as examples to explain the specific meaning and application process of the video tag push sequence. The analysis and application process of the audio tag push sequence, the text tag push sequence and the picture tag push sequence are the same as the video tag push sequence, so they are omitted in this embodiment.
[0135] The push content matching module is used to match the best push content for the user based on the user's interest preferences;
[0136] The push content matching module is configured with push content matching strategies, which include:
[0137] Get the number of converged media contents that need to be updated for the next refresh, named update number;
[0138] Calculate the number of updates and multiply it by the video push ratio to get the number of video pushes; calculate the number of updates and multiply it by the audio push ratio to get the number of audio pushes; calculate the number of updates and multiply it by the text push ratio to get the number of text pushes; calculate the number of updates and multiply it by the image push ratio to get the number of image pushes;
[0139] In actual applications, when a user refreshes the push homepage or re-enters the push homepage, the integrated media content needs to be updated. At this time, the number of integrated media content that needs to be displayed on the push homepage is the update number. The update number is 40. The number of video pushes is calculated to be 40×0.78=32, and the calculation result is rounded up to 1 to retain the integer. Similarly, the number of audio pushes is calculated to be 1, the number of text pushes is 1, and the number of picture pushes is 1.
[0140] When updating the pushed integrated media content, the media type of the video push number pushed to the user is the integrated media content of the video, the media type of the audio push number pushed is the integrated media content of the audio, the media type of the text push number pushed is the integrated media content of the text, and the media type of the picture push number pushed is the integrated media content of the picture;
[0141] Mark the integrated media content of the first video push number in the video push sequence as push content; mark the integrated media content of the first audio push number in the audio push sequence as push content; mark the integrated media content of the first text push number in the text push sequence as push content; mark the integrated media content of the first picture push number in the picture push sequence as push content;
[0142] In actual applications, when updating the pushed integrated media content, 32 video-type integrated media contents, 1 audio-type integrated media content, 1 text-type integrated media content, and 1 picture-type integrated media content are pushed to users; the integrated media contents corresponding to the top 32 content tags in the video push sequence are marked as pushed contents, the integrated media contents corresponding to the first content tag in the audio push sequence are marked as pushed contents, the integrated media contents corresponding to the first content tag in the text push sequence are marked as pushed contents, and the integrated media contents corresponding to the first content tag in the picture push sequence are marked as pushed contents.
[0143] The push content update module is used to update the user's content library based on the push content;
[0144] The push content update module is configured with a push content update strategy, which includes:
[0145] Calculate the sum of the number of video pushes, audio pushes, text pushes, and image pushes, then subtract their sum from the number of updates, and mark the result as the number of expanded pushes;
[0146] Calculate the real-time popularity of all converged media content, and mark the converged media content with the highest number of pushes ranked in real-time popularity as push-expanded content;
[0147] The calculation formula for real-time heat is: , where RTP is the real-time popularity, NC is the number of comments on the converged media content, NL is the number of likes on the converged media content, NR is the number of reposts of the converged media content, and NV is the number of views of the converged media content. For the like and forwarding coefficient, is the comment coefficient;
[0148] The number of comments, likes, reposts, and views are all obtained directly through the integrated media platform;
[0149] When the user needs to update the converged media content, the pushed content and the pushed extended content will be updated to the user's converged media platform;
[0150] In actual applications, the number of expanded pushes is calculated to be 40-(32+1+1+1)=5. The algorithm for calculating real-time popularity is similar to the hot search ranking algorithm, which is used to evaluate the spread and influence of integrated media content among users. as well as This is because usually, the number of likes and reposts by users is roughly the same, while the number of comments is usually much larger than the number of likes and reposts, and the number of views is much larger than the number of likes, reposts and comments. as well as Assign weights to them to ensure that their influence on real-time popularity is close; after analyzing historical big data, it is found that under normal circumstances, the frequency of users in the integrated media platform liking and forwarding is about 0.4 of the frequency of browsing, and the frequency of commenting is about 0.6 of the frequency of browsing, so Set to 0.4, Set to 0.6; since the calculation of real-time heat is similar to the calculation of hot search ranking, this embodiment will no longer display the specific calculation process. Please refer to the hot search ranking for specific calculation results; obtain several real-time heats through calculation, and mark the 5 largest integrated media contents among them as push extended content; when the user needs to update the integrated media content, update the push content and push extended content to the user's integrated media platform.
[0151] Example 2, please refer to Figure 4 As shown, the present application provides a method for updating integrated media content based on effect feedback to realize intelligent push, comprising the following steps:
[0152] Step S1, collecting user feedback data on the converged media content and media tags of the converged media content; Step S1 includes the following sub-steps:
[0153] Step S101, collecting user feedback data on converged media content, the feedback data including whether to like, whether to share, browsing time, and number of single content browsing times;
[0154] Step S102, when collecting feedback data of the converged media content, simultaneously obtaining media tags of the converged media content, the media tags including media categories and content tags;
[0155] Step S103, media categories include video, audio, text and picture;
[0156] Step S104, detecting whether the integrated media content has a preset tag, if so, outputting a tag preset signal; if not, outputting a tag extraction signal;
[0157] Step S105: if a tag preset signal is output, the preset tag is used as the content tag of the integrated media content; at the same time, the title of the integrated media content is obtained, named as the content title, keywords are extracted from the content title to obtain sample keywords, a keyword sample library is established, and the content tag and the corresponding sample keywords are collected;
[0158] Step S106: if a tag extraction signal is output, the content title of the integrated media content is obtained, and keywords are extracted from the content title to obtain title keywords;
[0159] Step S107, naming the content tag in the keyword sample library as a sample tag, searching for the sample tag where the title keyword is located in the keyword sample library, and naming it a search tag;
[0160] Step S108, for all search tags of the converged media content, count the number of identical search tags, name it as the tag number, find the maximum value among the tag numbers, and mark the corresponding search tag as the content tag;
[0161] Step S2, analyzing the feedback data and media tags to evaluate the user's interest preferences; Step S2 includes the following sub-steps:
[0162] Step S201, quantifying the feedback data based on the feedback data of the user browsing the converged media content;
[0163] Step S201 includes the following sub-steps:
[0164] Step S2011, for any converged media content browsed by the user, detecting whether the number of single content browsing times is one, if so, outputting a single browsing signal; if not, outputting a multiple browsing signal;
[0165] Step S2012: if a single browsing signal is output, the number of single content browsing of the integrated media content by the user is set to 1; if a multiple browsing signal is output, the number of single content browsing of the integrated media content by the user is set to 2;
[0166] Step S2013: based on the single content views of the integrated media content and the media tags of the users, the sum of the single content views of the users for different content tags is counted, and marked as tag views; at the same time, the sum of the single content views of the users for different media types is counted, and marked as type views;
[0167] Step S2014: based on whether the user likes the converged media content in the feedback data and the media tags corresponding to the converged media content, the number of likes of different content tags by the user is counted, marked as the tag likes, and if the number of single content views of the converged media content by the user is 2, the tag likes is increased by one; the number of likes of different media types by the user is counted, marked as the type likes, and if the number of single content views of the converged media content by the user is 2, the type likes is increased by one;
[0168] Step S2015: Based on whether the user shares the converged media content in the feedback data and the media tags corresponding to the converged media content, the user's sharing amount of different content tags is counted, marked as tag sharing amount. If the user's single content browsing amount of the converged media content is 2, the tag sharing amount is increased by 1; the user's sharing amount of different media types is counted, marked as type sharing amount. If the user's single content browsing amount of the converged media content is 2, the type sharing amount is increased by 1;
[0169] Step S202, analyzing the proportion of different content tags browsed by users in different media types based on the browsing time;
[0170] Step S202 includes the following sub-steps:
[0171] Step S2021, marking the converged media content whose media type is video and audio as progress bar content, obtaining the video duration or audio duration of the progress bar content, and marking it as the progress bar duration;
[0172] Step S2022, calculate the browsing time of the progress bar content that the user browses divided by the progress bar time, mark the calculation result as the single content browsing degree, and based on different content tags, count the sum of the single content browsing degrees of all progress bar contents in the content tags, mark it as the tag browsing ratio, the tag browsing ratio includes the tag video browsing ratio and the tag audio browsing ratio; by counting the tag browsing ratios of different content tags, obtain several tag browsing ratios;
[0173] Step S2023, marking the integrated media content with the media type of text and picture as page content, naming the bottom of the scroll bar in the page content as the browsing limit, and the scroll bar is in the scroll box;
[0174] Step S2024, dividing the scroll box into upper and lower parts according to the browsing limit, marking the part above the browsing limit as the browsed part, obtaining the length of the scroll box and the length of the browsed part, and marking them as the total content length and the browsed length respectively;
[0175] Step S2025, calculating the browsed length divided by the total content length, marking the calculation result as the single content viewing rate, and based on different content tags, counting the sum of the single content viewing rates of all page contents in the content tags, marking it as the tag viewing rate, where the tag viewing rate includes the tag text viewing rate and the tag image viewing rate; by counting the tag viewing rates of different content tags, several tag viewing rates are obtained;
[0176] Step S203, evaluating the user's interest preference;
[0177] Step S203 includes the following sub-steps:
[0178] Step S2031, obtaining type likes, type shares and type views, where type likes include video likes, audio likes, text likes and picture likes; type shares include video shares, audio shares, text shares and picture shares; type views include video views, audio views, text views and picture views;
[0179] Step S2032: in the order of n from small to large, by symbol V n represents the number of video views, video likes, and video shares, respectively, where n=1, 2, or 3; symbol A n It represents the number of audio views, audio likes and audio shares respectively; the symbol W n It represents the number of text views, the number of text likes, and the number of text shares respectively; the symbol P n It represents the number of image views, image likes, and image shares respectively;
[0180] Step S2033, the symbol VP represents the browsing ratio of the tag video, the symbol AP represents the browsing ratio of the tag audio, the symbol WP represents the viewing ratio of the tag text, and the symbol PP represents the viewing ratio of the tag picture;
[0181] Step S2034, by formula Calculate the video push index, where VI is the video push index, and the simplified formula is: ; Similarly, through the formula Calculate the audio push index through the formula Calculate the text push index through the formula Calculate image push indicators;
[0182] Step S2035, calculate VI+AI+WI+PI again, mark the calculation result as the total push index, represented by the symbol TI, set the interest push ratio, represented by the symbol TX, calculate VI / TI×TX to get the video push ratio, calculate AI / TI×TX to get the audio push ratio, calculate WI / TI×TX to get the text push ratio, and calculate PI / TI×TX to get the picture push ratio;
[0183] Step S2036, the video push ratio, audio push ratio, text push ratio and picture push ratio are the user's interest preference for media types, named type preference;
[0184] Step S2037, analyzing the user's interest preference for content tags based on tag views, tag likes, tag shares, tag views ratio, and tag viewing ratio, which is named tag preference;
[0185] Step S2037 includes the following sub-steps:
[0186] Step S2037.a, calculating the push priorities of different content tags using a push priority calculation formula;
[0187] The push priority calculation formula is configured as:
[0188]
[0189]
[0190]
[0191]
[0192] Among them, FV is the video push priority, TV is the tag views, TL is the tag likes, TS is the tag shares, FA is the audio push priority, FW is the text push priority, FP is the picture push priority, α is the like coefficient, and β is the share coefficient;
[0193] Step S2037.b, sort FV in descending order to obtain a video tag push sequence; sort FA in descending order to obtain an audio tag push sequence; sort FW in descending order to obtain a text tag push sequence; sort FP in descending order to obtain a picture tag push sequence;
[0194] Step S2037.c: the video tag push sequence, the audio tag push sequence, the text tag push sequence, and the picture tag push sequence are the tag preferences;
[0195] Step S3, matching the best push content for the user based on the user's interest preference; Step S3 includes the following sub-steps:
[0196] Step S301, obtaining the number of converged media contents that need to be updated for the next refresh, named update number;
[0197] Step S302, calculate the update number and multiply it by the video push ratio to obtain the video push number; calculate the update number and multiply it by the audio push ratio to obtain the audio push number; calculate the update number and multiply it by the text push ratio to obtain the text push number; calculate the update number and multiply it by the picture push ratio to obtain the picture push number;
[0198] Step S303: when updating the pushed integrated media content, the media type of the video push number is the integrated media content of video, the media type of the audio push number is the integrated media content of audio, the media type of the text push number is the integrated media content of text, and the media type of the picture push number is the integrated media content of picture;
[0199] Step S304: Mark the integrated media content of the first video push number in the video push sequence as push content; mark the integrated media content of the first audio push number in the audio push sequence as push content; mark the integrated media content of the first text push number in the text push sequence as push content; mark the integrated media content of the first picture push number in the picture push sequence as push content;
[0200] Step S4, updating the user's content library based on the pushed content; Step S4 includes the following sub-steps:
[0201] Step S401, calculating the sum of the number of video pushes, the number of audio pushes, the number of text pushes, and the number of picture pushes, then subtracting the sum from the number of updates, and marking the calculation result as the number of extended pushes;
[0202] Step S402, calculating the real-time popularity of all converged media contents, and marking the converged media contents with the highest number of expanded pushes ranked in real-time popularity as push expanded contents;
[0203] Step S403: the real-time heat is calculated as: , where RTP is the real-time popularity, NC is the number of comments on the converged media content, NL is the number of likes on the converged media content, NR is the number of reposts of the converged media content, and NV is the number of views of the converged media content. For the like and forwarding coefficient, is the comment coefficient;
[0204] Step S404, the number of comments, likes, reposts and views are all directly obtained through the integrated media platform;
[0205] Step S405: When the user needs to update the integrated media content, the pushed content and the pushed extended content are updated to the user's integrated media platform.
[0206] Example 3, please refer to Figure 5 As shown, Figure 5 The structural diagram of an electronic device is illustrated, and the electronic device may include: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus. The memory stores computer-readable instructions, and the processor can call the instructions in the memory. When the computer-readable instructions are executed by the processor, the steps in the method for updating the integrated media content for intelligent push based on effect feedback are executed to achieve the following functions: collecting user feedback data on the integrated media content and the media tags of the integrated media content; analyzing the feedback data and the media tags to evaluate the user's interest preferences; matching the best push content for the user based on the user's interest preferences; and updating the user's content library based on the push content.
[0207] In addition, the logic instructions in the above-mentioned memory can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.
[0208] Embodiment 4, the present application also provides a computer-readable storage medium, the present application provides a storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps in the method for updating integrated media content for intelligent push based on effect feedback are executed to achieve the following functions: collecting user feedback data on integrated media content and media tags of integrated media content; analyzing the feedback data and media tags to evaluate the user's interest preferences; matching the best push content for the user based on the user's interest preferences; and updating the user's content library based on the push content.
[0209] Through the description of the above implementation methods, the embodiments of the present invention can be provided as methods, systems or computer program products. Based on such an understanding, the above technical solutions can be essentially or partly contributed to the prior art in the form of software products, which can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and include several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.
[0210] In the embodiments provided in the present application, it should be understood that the disclosed system or method can be implemented in other ways. The embodiments described above are merely schematic. For example, the division of modules or units is only a logical function division. There may be other division methods in actual implementation. For example, multiple modules or units can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of systems, modules and units can be electrical, mechanical or other forms.
[0211] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.
Claims
1. A method for updating integrated media content based on effect feedback and intelligent push, characterized in that: The steps include: Collect user feedback data on converged media content and media tags of converged media content; Analyze feedback data and media tags to evaluate user interests and preferences; Match the best push content to users based on their interests and preferences; Update the user's content library based on the pushed content; The media tags include media categories and content tags, and the media categories include video, audio, text, and pictures; Analyzing feedback data and media tags to evaluate user interest preferences includes the following sub-steps: Quantify the feedback data based on the feedback data of users browsing the converged media content; Based on browsing time, analyze the proportion of different content tags browsed by users in different media types; Assess user interests and preferences; Based on the feedback data of users browsing the converged media content, quantifying the feedback data includes the following sub-steps: For any converged media content browsed by the user, detect whether the number of views of the single content is one, if so, output a single view signal; if not, output a multiple view signal; If a single browsing signal is output, the number of single content browsing of the integrated media content by the user is set to 1; if a multiple browsing signal is output, the number of single content browsing of the integrated media content by the user is set to 2; Based on the single content views and media tags of the integrated media content by users, the sum of the single content views of different content tags by users is counted, and marked as tag views; at the same time, the sum of the single content views of different media types by users is counted, and marked as type views; Based on whether the user likes the converged media content in the feedback data and the media tags corresponding to the converged media content, the number of likes of different content tags by users is counted, marked as tag likes. If the number of single content views of the converged media content by the user is 2, the number of tag likes will increase by one; the number of likes of different media types by users is counted, marked as type likes. If the number of single content views of the converged media content by the user is 2, the type likes will increase by one; Based on the feedback data of whether users share the converged media content and the media tags corresponding to the converged media content, the number of users sharing different content tags is counted, marked as tag sharing amount. If the number of single content views of the converged media content by the user is 2, the tag sharing amount is increased by 1; the number of users sharing different media types is counted, marked as type sharing amount. If the number of single content views of the converged media content by the user is 2, the type sharing amount is increased by 1; Based on browsing time, analyzing the proportion of different content tags browsed by users in different media types includes the following sub-steps: Mark the converged media content whose media type is video or audio as progress bar content, obtain the video duration or audio duration of the progress bar content, and mark it as the progress bar duration; Calculate the browsing time of the progress bar content that the user browses divided by the progress bar time, mark the calculation result as the single content browsing rate, and based on different content tags, count the sum of the single content browsing rates of all progress bar contents in the content tags, and mark it as the tag browsing ratio, where the tag browsing ratio includes the tag video browsing ratio and the tag audio browsing ratio; By counting the tag browsing ratios of different content tags, we can get the tag browsing ratios of several tags; Mark the converged media content whose media type is text and picture as page content, name the bottom of the scroll bar in the page content as browsing limit, and the scroll bar is in the scroll box; Divide the scroll box into two parts, upper and lower, according to the browsing limit, mark the part above the browsing limit as the browsed part, obtain the length of the scroll box and the length of the browsed part, and mark them as the total content length and the browsed length respectively; Calculate the browsed length divided by the total content length, and mark the calculation result as the single content viewing rate. Based on different content tags, count the sum of the single content viewing rates of all page contents in the content tags, and mark it as the tag viewing rate. The tag viewing rate includes the tag text viewing rate and the tag image viewing rate. By counting the tag viewing rates of different content tags, several tag viewing rates are obtained. Evaluating user interest preferences includes the following sub-steps: Obtain type likes, type shares and type views, wherein the type likes include video likes, audio likes, text likes and picture likes; the type shares include video shares, audio shares, text shares and picture shares; the type views include video views, audio views, text views and picture views; According to the order of n from small to large, through the symbol V n represents the number of video views, video likes, and video shares, respectively, where n=1, 2, or 3; symbol A n It represents the number of audio views, audio likes and audio shares respectively; the symbol W n It represents the number of text views, the number of text likes, and the number of text shares respectively; the symbol P n It represents the number of image views, image likes, and image shares respectively; The symbol VP represents the percentage of video views, the symbol AP represents the percentage of audio views, the symbol WP represents the percentage of text views, and the symbol PP represents the percentage of image views. By formula Calculate the video push index, where VI is the video push index, and the simplified formula is: ; Similarly, through the formula Calculate the audio push index through the formula Calculate the text push index through the formula Calculate image push indicators; Then calculate VI+AI+WI+PI, mark the result as the total push index, represented by the symbol TI, set the interest push ratio, represented by the symbol TX, calculate VI / TI×TX to get the video push ratio, calculate AI / TI×TX to get the audio push ratio, calculate WI / TI×TX to get the text push ratio, and calculate PI / TI×TX to get the picture push ratio; The video push ratio, audio push ratio, text push ratio and picture push ratio are user's interest preferences for media types, named as type preferences; Analyze users' interest preferences for content tags based on tag views, tag likes, tag shares, tag views and tag views, and name it tag preference; Analyzing users' interest preferences for content tags based on tag views, tag likes, tag shares, tag views ratio, and tag viewing ratio includes the following sub-steps: Calculate the push priority of different content tags through the push priority calculation formula; The push priority calculation formula is configured as: Among them, FV is the video push priority, TV is the tag views, TL is the tag likes, TS is the tag shares, FA is the audio push priority, FW is the text push priority, FP is the picture push priority, α is the like coefficient, and β is the share coefficient; Sort FV in descending order to get the video tag push sequence; Sort FA in descending order to get the audio tag push sequence; Sort FW in descending order to get the text tag push sequence; Sort FP in descending order to get the picture tag push sequence; The video tag push sequence, the audio tag push sequence, the text tag push sequence, and the picture tag push sequence are tag preferences.
2. The method for updating integrated media content based on effect feedback and realizing intelligent push according to claim 1, characterized in that: Collecting user feedback data on converged media content and media tags of converged media content includes the following sub-steps: Collect user feedback data on converged media content, including whether to like, share, browsing time, and number of single content browsing times; When collecting feedback data on converged media content, the media tags of the converged media content are also obtained; Detect whether there is a preset tag in the integrated media content. If so, output a tag preset signal; if not, output a tag extraction signal; If a tag preset signal is output, the preset tag is used as the content tag of the integrated media content; at the same time, the title of the integrated media content is obtained and named as the content title, keywords are extracted from the content title to obtain sample keywords, a keyword sample library is established, and the content tags and corresponding sample keywords are collected; If a tag extraction signal is output, the content title of the converged media content is obtained, and keywords are extracted from the content title to obtain title keywords; Name the content tag in the keyword sample library as a sample tag, search for the sample tag where the title keyword is located in the keyword sample library, and name it a search tag; For all search tags of converged media content, the number of identical search tags is counted and named as the number of tags. The maximum value among the number of search tags is found and the corresponding search tag is marked as a content tag.
3. The method for updating integrated media content based on effect feedback and realizing intelligent push according to claim 2, characterized in that: Matching the best push content to users based on their interests includes the following sub-steps: Get the number of converged media contents that need to be updated for the next refresh, named update number; Calculate the number of updates and multiply it by the video push ratio to get the number of video pushes; Calculate the update quantity and multiply it by the audio push ratio to get the audio push quantity; Calculate the number of updates and multiply it by the text push ratio to get the number of text pushes; Calculate the number of updates and multiply it by the image push ratio to get the number of image pushes. When updating the pushed integrated media content, the media type of the video push number pushed to the user is the integrated media content of the video, the media type of the audio push number pushed is the integrated media content of the audio, the media type of the text push number pushed is the integrated media content of the text, and the media type of the picture push number pushed is the integrated media content of the picture; Mark the integrated media content of the first video push number in the video push sequence as pushed content; mark the integrated media content of the first audio push number in the audio push sequence as pushed content; mark the integrated media content of the first text push number in the text push sequence as pushed content; mark the integrated media content of the first picture push number in the picture push sequence as pushed content.
4. The method for updating integrated media content based on effect feedback and realizing intelligent push according to claim 3 is characterized in that: Updating the user's content library based on pushed content includes the following sub-steps: Calculate the sum of the number of video pushes, audio pushes, text pushes, and image pushes, then subtract their sum from the number of updates, and mark the result as the number of expanded pushes; Calculate the real-time popularity of all converged media content, and mark the converged media content with the highest number of pushes ranked in real-time popularity as push-expanded content; The calculation formula of the real-time heat is: , where RTP is the real-time popularity, NC is the number of comments on the converged media content, NL is the number of likes on the converged media content, NR is the number of reposts of the converged media content, and NV is the number of views of the converged media content. For the like and forwarding coefficient, is the comment coefficient; The number of comments, likes, reposts and views are all directly obtained through the integrated media platform; When the user needs to update the integrated media content, the pushed content and the pushed extended content will be updated to the user's integrated media platform.
5. A system for updating integrated media content based on effect feedback and realizing intelligent push, which is used to implement the method for updating integrated media content based on effect feedback and realizing intelligent push according to any one of claims 1 to 4, characterized in that: It includes a platform data collection module, an interest preference evaluation module, a push content matching module and a push content update module; the platform data collection module, the interest preference evaluation module and the push content update module are respectively connected to the push content matching module data; The platform data collection module is used to collect user feedback data on converged media content and media tags of converged media content; The interest preference evaluation module is used to analyze the feedback data and media tags to evaluate the user's interest preferences; The push content matching module is used to match the best push content for the user based on the user's interest preferences; The push content update module is used to update the user's content library based on the push content.
