Extra interest label blind box recommendation method for breaking information cocoon house, medium and equipment

Abstract

The invention discloses an extra interest label blind box recommendation method for breaking an information cocoon house, a medium, equipment and a computer program product, and relates to the social media platform recommendation field, the extra interest label blind box recommendation method for breaking the information cocoon house mainly comprises the following steps: according to a user-defined interest label, a user browsing content and an opposite label pair; and obtaining an extra interest label, pushing the extra interest label in a blind box form, and obtaining recommended content according to the user feedback and the extra interest label. By implementing the extra interest label blind box recommendation method for breaking the information cocoon house, the medium, the equipment and the computer program product provided by the invention, the personal information cocoon house breaking effect can be improved, the user experience is improved, and the algorithm updating cost is reduced.

Description

Technical Field

[0001] The present invention relates to the field of social media platform recommendations, and more specifically, to a method, medium, device and computer program product for recommending additional interest tags in a blind box that breaks the information cocoon. Background Art

[0002] With the continuous development of Internet technology, the speed at which people obtain information is growing exponentially. Major social media platforms use personalized algorithms to accurately push content that individuals are willing to see to target users, forming an information cocoon over time.

[0003] Existing technologies are basically optimized at the algorithm level. For example, the recommendation algorithm is optimized to enhance the diversity and personalization of recommendations, the boundaries of information fields are broken through cross-domain recommendation strategies, and multimodal information is used to debias and provide a more comprehensive information perspective. However, these are still at the theoretical level and are not very practical. On the other hand, education and guidance for users: existing social media platforms have reported on information cocoons and tried to make users understand the working principles behind recommendation algorithms, so that users can better understand why they see specific types of recommendations, thereby improving their sense of control over information recommendations, but there is a lack of timely action effects and feedback for users.

[0004] In the prior art, there is a blind box navigation method, which is as follows: the user inputs the destination type, and the system selects several points from the POI of the destination as alternative POIs; the alternative POIs are then pushed to the user in the form of a blind box, and after the user selects any blind box, the alternative POI becomes the target POI, increasing the user's travel experience. The application field of this method is limited and is only applicable to the fun guidance of the transportation process. In the form of POI blind boxes, users can increase their travel experience in uncertainty; but it does not touch on the core issue of the information cocoon, namely the personalized recommendations of various social media platforms. There is also a tag-based short video recommendation method, which obtains the tags and tag relevance of the video for the platform; generates a user portrait based on user behavior data, and calculates the user's tag preference for the video; then recommends the video based on the user portrait in a content- and user-collaborative manner; improves the accuracy and surprise level of the recommendation results, and can meet the diverse needs of users; this method labels the short videos and calculates the user's tag preference, and recommends the video based on content- and user-collaborative manner; solves the problem of personalized recommendation on the platform in the information cocoon, and allows users to participate in collaborative management; however, the invention does not provide a specific user collaboration method, and users have no idea where to start when faced with various tag preferences.

[0005] How to provide a recommendation method that is simple and feasible for the platform, and at the same time has a certain degree of participation and incentive effect for users, is a problem that needs to be solved urgently. Summary of the invention

[0006] The purpose of the present invention is to provide a method, medium, device and computer program product for recommending additional interest tags in a blind box to break the information cocoon, which can improve the effect of breaking the personal information cocoon, enhance the user experience and reduce the cost of algorithm updates.

[0007] The present invention provides a method for recommending additional interest tags blind boxes for breaking the information cocoon, comprising the following steps: Step S1: obtaining user-defined interest tags, user browsed content and opposing tag pairs, and obtaining additional interest tags according to the user-defined interest tags, user browsed content and opposing tag pairs; Step S2: preprocessing the additional interest tags to obtain an additional interest tag library; Step S3: obtaining a tag blind box according to the additional interest tag library; Step S4: obtaining user feedback information of the tag blind box, and obtaining recommended content according to the tag blind box user feedback information and the tag blind box.

[0008] Furthermore, step S1 of the above-mentioned additional interest tag blind box recommendation method for breaking the information cocoon specifically includes: S11: obtaining user-defined interest tags; S12: obtaining user browsing content, and obtaining a user interest tag set based on the user browsing content by using natural language processing technology; S13: obtaining weakly related tags based on the user interest tag set by using a similarity measurement method; S14: obtaining opposing tag pairs, and obtaining opposite tags based on the user interest tag set and the opposing tag pairs; S15: obtaining additional interest tags based on user-defined interest tags, weakly related tags and opposite tags.

[0009] Furthermore, step S12 of the above-mentioned method for recommending additional interest tags in a blind box to break the information cocoon specifically includes: obtaining the user's browsing content, and obtaining a user interest tag set based on the user's browsing content by using natural language processing technology, such as the formula: , , in, Browse content for users, The first time a user browses content i content items, n is the number of content items browsed by the user, A collection of user interest tags. It stands for Natural Language Processing.

[0010] Furthermore, step S13 of the above-mentioned additional interest tag blind box recommendation method for breaking the information cocoon specifically includes: according to the user interest tag set, using the similarity measurement method, obtaining weakly related tags, such as the formula: , , in, Indicates similarity, For potential tags, A collection of user interest tags. Indicates taking the maximum value, represents the cosine similarity, Yes Tags The vector representation of For user interest tags, Yes Tags The vector representation of is a weakly related label, and are the lower and upper bounds of the similarity.

[0011] Furthermore, step S14 of the above-mentioned additional interest tag blind box recommendation method for breaking the information cocoon specifically includes: obtaining an opposing tag pair, and obtaining an opposite tag according to the user interest tag set and the opposing tag pair, such as the formula: , , Among them, C is the opposing label pair, and etc. are pairs of labels with opposite opinions. is a set of opposite labels, For potential tags, For user interest tags, A collection of user interest tags.

[0012] Furthermore, the preprocessing of the above-mentioned additional interest tag blind box recommendation method for breaking the information cocoon includes removing duplicate tags and merging similar tags.

[0013] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-mentioned method for recommending additional interest tags blind boxes that break the information cocoon.

[0014] The present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps of the above-mentioned method for recommending blind boxes of additional interest tags that break the information cocoon are implemented.

[0015] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the above-mentioned method for recommending additional interest tags blind boxes to break the information cocoon.

[0016] The implementation of the additional interest tag blind box recommendation method, medium, device and computer program product provided by the present invention for breaking the information cocoon has the following beneficial effects: The present invention allows users to have greater weight in participating in the management of personal information cocoons by customizing the input of additional interest tag content. At the same time, relevant additional interest tag keywords are extracted based on the previous personalized recommendation data set, and opposite or weakly related additional interest tags are formed according to the user's daily usage preferences, thereby improving the effect of breaking the personal information cocoon. The present invention allows users to directly observe the additional interest tags and their recommended videos by externalizing the additional interest tags, and pushes them to users in combination with blind boxes, which greatly improves the fun of the use process with a high degree of uncertainty, and improves the user's enthusiasm for independently breaking the information cocoon problem. The present invention is simpler and more intuitive and can inspire users to actively participate in the process of breaking the information cocoon governance. It can allow users to see the corresponding content simply and intuitively, reducing the cost of learning and understanding. At the same time, it can be implemented without major algorithm changes. Only additional interest tag interfaces and related content need to be added. For the platform, there is no need to invest too much algorithm update cost, and it can be promoted to major existing self-media platforms. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The present invention will be further described below with reference to the accompanying drawings and embodiments, in which: Figure 1 It is a flow chart of the additional interest tag blind box recommendation method for breaking the information cocoon provided by the present invention; Figure 2 It is a schematic diagram of the additional interest tag blind box recommendation steps for breaking the information cocoon provided by the present invention; Figure 3 It is a schematic diagram of the additional interest tag blind box recommendation system for breaking the information cocoon provided by the present invention; Figure 4 It is a low-fidelity operation interface diagram of the client provided by the present invention; Figure 5 It is a structural block diagram of the computer device provided by the present invention. DETAILED DESCRIPTION

[0018] In order to have a clearer understanding of the technical features, purposes and effects of the present invention, specific embodiments of the present invention are now described in detail with reference to the accompanying drawings.

[0019] Figure 1 A schematic diagram of a blind box recommendation method for additional interest tags that breaks the information cocoon of this embodiment is shown. In this embodiment, the blind box recommendation method for additional interest tags that breaks the information cocoon includes the following steps: S1: Obtain user-defined interest tags, user-browsed content, and opposing tag pairs, and obtain additional interest tags based on the user-defined interest tags, user-browsed content, and opposing tag pairs; Specifically, step S1 includes: S11: Obtain user-defined interest tags; S12: Obtain the user's browsing content, and obtain a user interest tag set based on the user's browsing content using natural language processing technology; Specifically, step S12 includes: obtaining the user browsing content, and obtaining the user interest tag set according to the user browsing content by using natural language processing technology, such as formula: , , in, Browse content for users, The first time a user browses content i content items, n is the number of content items browsed by the user, A collection of user interest tags. stands for Natural Language Processing; Specifically, natural language processing technology can be TF-IDF and word embedding, etc., which are used to extract tags that users are interested in from the content text; each content item corresponds to one or more tags; S13: According to the user interest tag set, a weakly related tag is obtained by using a similarity measurement method; Specifically, step S13 includes: according to the user interest tag set, using a similarity measurement method, obtaining weakly related tags, such as formula: , , in, Indicates similarity, For potential tags, A collection of user interest tags. Indicates taking the maximum value, represents the cosine similarity, Yes Tags The vector representation of For user interest tags, Yes Tags The vector representation of is a weakly related label, and are the lower and upper limits of similarity; S14: Obtaining an opposing tag pair, obtaining an opposite tag according to the user interest tag set and the opposing tag pair; Specifically, step S14 includes: obtaining an opposing tag pair, and obtaining an opposite tag according to the user interest tag set and the opposing tag pair, such as the formula: , , Among them, C is the opposing label pair, and etc. are pairs of labels with opposite opinions. is a set of opposite labels, For potential tags, For user interest tags, A collection of user interest tags; S15: obtaining additional interest tags according to the user-defined interest tags, weakly related tags and opposite tags; S2: preprocess the additional interest tags to obtain an additional interest tag library; Specifically, preprocessing includes removing duplicate tags and merging similar tags; Specifically, step S2 includes: aggregating the added tags into an additional interest tag library and identifying and removing duplicate tags and merging similar tags; the external interface of the additional interest tag library is the "Tag Library" function under the "Today's Additional Interest Tags" page, where users can view and manage historical additional interest tags; Step S3: Obtain a label blind box according to the additional interest label library; Specifically, additional interest tags in the library are pushed to users in the form of blind boxes every day to increase the fun and uncertainty of the push; Step S4: Obtaining user feedback information of the label blind box, and obtaining recommended content according to the user feedback information of the label blind box and the label blind box; Specifically, in step S4, the user clicks to confirm or reselect (limited to y times); the system recommends related content based on this tag and randomly appears on the homepage, ensuring the diversity and richness of the recommended content.

[0020] In some embodiments, the above-mentioned additional interest tag blind box recommendation method for breaking the information cocoon can also be implemented in the following way.

[0021] In this embodiment, the additional interest tag blind box recommendation method for breaking the information cocoon is as follows: Figure 2 As shown, the following steps are included: Step S1: The user enters the client homepage through the client and selects the "Anti-Information Cocoon" function option. If the user clicks on this function option for the first time, the platform will first push the concept of "additional interest tags" and the process of playing to the client in the form of pictures, texts or videos for introduction; Step S2: After the user finishes browsing the platform content or clicks it next time, he can directly enter the "Today's Extra Interest Tags" interface. The user can add extra interest tags by clicking the "Add" button in the center at the bottom of the interface. Among them, there are two main sources of extra interest tags: First, through the extra interest tag custom input module, the user customizes and edits other fields of knowledge that he wants to know, and the value of the added tag is recorded as n1; among them, the upper limit of the number of tags added per day is n, that is, n1≤n, and the upper limit of the number of tag characters that can be input at a time is x, and the value of x is determined by the platform computing power; Second, through the extra interest tag system recommendation generation module, the system automatically generates content tags that are opposite to / weakly related to the user's views other than the information he pays attention to on a daily basis; the value is consistent with the number of tags n1 that the user customizes and edits on the same day, and the tags generated by the system cannot be edited by the user, ensuring that the user has only 50% of the weight to participate in the governance of the information cocoon; if the user weight is 100%, it may cause the user to add content that is of interest to him every time, thus falling into a new information cocoon; therefore, the total value of the extra interest tags added per day is n1+n1=2n1; Additional interest tags generated based on the user's daily interests are randomly placed on the current editing page, including content that is contrary to and weakly related to the content browsed by the user's daily personalized recommendation algorithm; Step S3: Through the tag storage submodule, the system creates and updates the added additional interest tags to the additional interest tag library; through the tag deduplication submodule, the system identifies duplicate and similar tags and merges or removes them; through the tag display submodule, the system visualizes the additional interest tag library; through the tag deletion submodule, the system allows users to delete additional interest tags; Specifically, after the editing is completed, the platform will summarize the added additional interest tags into the additional interest tag library; users can click on different additional interest tags in the "Tag Library" function option in the upper right corner of the "Today's Additional Interest Tags" interface to view tag details such as the time the tag was added, the number of times it has been recommended, the history of recommended related videos, etc.; users can delete additional interest tags; Step S4: Through the additional interest tag random extraction module, when the user logs in to the client or APP for the first time every day, the system will randomly extract a today's additional interest tag from the additional interest tag library and push it to the user in the form of a blind box pop-up window; Step S5: Through the additional interest tag confirmation module, if the user is satisfied with the additional interest tags recommended today, he / she can click the "Confirm" button at the bottom right of the pop-up window; if not satisfied, the user can click the "Change" button at the bottom left of the pop-up window to re-extract additional interest tags, but the number of re-extractions per day is limited to y times to ensure that the user will not fall into a new information cocoon; Step S6: After the user determines today's additional interest tags, the system will randomly recommend related content on the user's homepage based on the tag content through the additional interest tag content recommendation module. The system will also display "Additional interest tag recommendation" next to the relevant content and attach the specific content of today's additional interest tags to encourage users to browse the content completely; Specifically, after the user determines today's additional interest tags, the platform will randomly recommend related videos on the user's homepage based on the tag content; and display "Additional interest tag recommended video" between the "Send barrage" icon and the "@author" icon on the left side of such video, accompanied by the specific content of today's additional interest tags, to encourage users to watch it in full; after the user finishes watching, the system will record the user's viewing times, content, etc. in the additional interest tag library mentioned in step S3 for future users to view.

[0022] This embodiment provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the above-mentioned method for recommending additional interest tags blind boxes that break the information cocoon; Specifically, the computer program includes an additional interest tag blind box recommendation system for breaking the information cocoon, and the additional interest tag blind box recommendation system for breaking the information cocoon is as follows: Figure 3 As shown, it includes the following components: The module for customizing input of additional interest tags allows users to manually input additional interest tags and impose corresponding restrictions and management. Specifically, users can summarize keywords of other content they want to know in this module and enter them for submission, with an upper limit of x characters, and the value of x is determined by the computing power of each platform. The module for customizing input of additional interest tags includes the following two submodules: an input verification submodule, which is used to verify whether the number of characters of the tag input by the user meets the upper limit x set by the platform, and ensure that the number of tags n1 does not exceed the daily upper limit n; a tag formatting submodule, which is used to format the tags input by the user to ensure that they meet the system requirements. The additional interest tag system recommendation generation module automatically generates additional interest tags based on the user's browsing history and content preferences. Specifically, the platform generates tags that are opposite to / weakly related to the user's views based on the content browsed by the daily personalized recommendation algorithm, and inputs them to the user's current editing page. The value is consistent with the number of additional interest tags input by the user on that day. The function of this module can be based on the user's previous personalized recommendation data set to extract relevant additional interest tag keywords. The additional interest tag system recommendation generation module includes the following two sub-modules: a user interest modeling sub-module, which is used to analyze the user's historical data to extract the user's interest tags; a tag generation sub-module, which is used to generate content tags that are opposite to / weakly related to the user's views other than the information that the user pays attention to on a daily basis based on an algorithm. The additional interest tag database is derived from the aggregation and saving of the added additional interest tags in a centralized database; specifically, the external interface of the database is the "Tag Library" function under the "Today's Additional Interest Tags" page, and users can view and manage historical additional interest tags after clicking; the operations that users can perform are: 1. Delete additional interest tags, but limit the daily deletion limit to t to prevent users from only deleting tags that they are not interested in; 2. Click to view the details of each tag, such as the time when the tag was added, the number of times it has been recommended, the history of recommended related videos, etc.; the additional interest tag database includes the following four sub-modules: a tag storage sub-module, which is used to process user-defined and system-generated tags, and insert or update them into the additional interest tag library; a tag deduplication sub-module, which is used to identify and process duplicate and similar tags in the tag library to ensure the uniqueness and rationality of the tags in the library; a tag display sub-module, which is used to display detailed information of each additional interest tag to users; a tag deletion sub-module, which allows users to delete unwanted additional interest tags, but there is a daily deletion limit; The module for randomly extracting additional interest tags. For users who have participated in the program, the system will automatically randomly select an additional interest tag from the additional interest tag library after the user logs into the APP for the first time every day. For users who have not participated in the program, they need to add tags in the "Today's Additional Interest Tags" interface and click the "Draw Blind Box" button at the bottom to extract additional interest tags. The additional interest tag blind box recommendation module is used to push the extracted additional interest tags to users in the form of pop-up windows. The high degree of randomness can greatly increase users' curiosity and participation. Additional interest tag confirmation module, used to provide "Confirm" and "Change" buttons at the bottom of the pop-up window. If the user clicks "Change", the system will randomly select a tag from the tag library for recommendation again until the daily re-extraction limit is reached; The additional interest tag content recommendation module generates and displays relevant content, including videos, articles, pictures, etc., based on the additional interest tags selected by the user: Specifically, after clicking "Confirm" in the previous module, the system randomly recommends relevant content on the homepage based on the additional interest tags confirmed by the user. The recommended content may include videos, articles, pictures, etc., and is marked with "Additional interest tag recommended content" below to encourage users to browse; Among them, the additional interest tag content recommendation module includes the following three sub-modules: a content matching sub-module, which is used to retrieve relevant content and match it with the tags according to the tags selected by the user, and determine the top five contents with the highest matching degree; a content recommendation sub-module, which is used to push qualified content to the APP homepage, and display "Additional interest tag recommendation" next to the relevant content with the specific content of today's additional interest tags; a viewing behavior record sub-module, which is used to record the number of times and content of videos recommended by additional interest tags watched by users, and store them in the additional interest tag library; Specifically, on the client side, the “Anti-Information Cocoon” icon is placed in a prominent position on the platform homepage, which is easy for users to find and click; the “Additional Interest Tags” interface is used to jump to the “Today’s Additional Interest Tags” interface after clicking the icon, which allows users to add additional interest tags. The relevant interactive relationships are as follows: Figure 4 The following is a low-fidelity operation interface diagram of the client.

[0023] This embodiment provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned method for recommending blind boxes of additional interest tags that break the information cocoon are implemented. The storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), a flash memory, a hard disk drive (HDD) or a solid-state drive (SSD), etc.; the storage medium may also include a combination of the above-mentioned types of memory.

[0024] This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps of the above-mentioned method for recommending blind boxes of additional interest tags that break the information cocoon are implemented.

[0025] like Figure 5As shown, the computer device may include: at least one processor 121, such as a CPU (Central Processing Unit), at least one communication interface 123, a memory 124, and at least one communication bus 122. Among them, the communication bus 122 is used to realize the connection and communication between these components. Among them, the communication interface 123 may include a display screen (Display), a keyboard (Keyboard), and the optional communication interface 123 may also include a standard wired interface and a wireless interface. The memory 124 may be a high-speed RAM memory (Random Access Memory, volatile random access memory), or a non-volatile memory (non-volatile memory), such as at least one disk memory. The memory 124 may also be at least one storage device located away from the aforementioned processor 121. Among them, the memory 124 stores an application program, and the processor 121 calls the program code stored in the memory 124 to perform any of the above method steps. Among them, the communication bus 122 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The communication bus 122 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 5Only one line is used to represent it, but it does not mean that there is only one bus or one type of bus. Among them, the memory 124 may include a volatile memory (English: volatile memory), such as a random access memory (English: random-access memory, abbreviated: RAM); the memory may also include a non-volatile memory (English: non-volatile memory), such as a flash memory (English: flash memory), a hard disk drive (English: hard disk drive, abbreviated: HDD) or a solid-state drive (English: solid-state drive, abbreviated: SSD); the memory 124 may also include a combination of the above-mentioned types of memory. Among them, the processor 121 may be a central processing unit (English: central processing unit, abbreviated: CPU), a network processor (English: network processor, abbreviated: NP) or a combination of CPU and NP. Among them, the processor 121 may further include a hardware chip. The above-mentioned hardware chip may be an application-specific integrated circuit (English: application-specific integrated circuit, abbreviated: ASIC), a programmable logic device (English: programmable logic device, abbreviated: PLD) or a combination thereof. The above-mentioned PLD can be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof. Optionally, the memory 124 is also used to store program instructions. The processor 121 can call program instructions to implement the additional interest tag blind box recommendation method for breaking the information cocoon as in this embodiment.

[0026] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the enlightenment of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the purpose of the present invention and the claims, which all fall within the protection of the present invention.

Claims

1. A blind box recommendation method for additional interest tags that breaks the information cocoon, characterized in that: The following steps are involved: S1: Obtain user-defined interest tags, user-browsed content, and opposing tag pairs, and obtain additional interest tags according to the user-defined interest tags, user-browsed content, and opposing tag pairs; S2: preprocessing the additional interest tags to obtain an additional interest tag library; S3: Obtain a label blind box according to the additional interest label library; S4: Obtain user feedback information of the label blind box, and obtain recommended content based on the user feedback information of the label blind box and the label blind box.

2. The method for recommending blind boxes of additional interest tags for breaking the information cocoon according to claim 1 is characterized in that: Step S1 specifically includes: S11: Obtain user-defined interest tags; S12: Obtaining user browsing content, and obtaining a user interest tag set based on the user browsing content using natural language processing technology; S13: Obtain weakly related tags using a similarity measurement method according to the user interest tag set; S14: Obtain an opposing tag pair, and obtain an opposite tag according to the user interest tag set and the opposing tag pair; S15: Obtain additional interest tags according to the user-defined interest tags, the weakly related tags and the opposite tags.

3. The method for recommending blind boxes of additional interest tags for breaking the information cocoon according to claim 2 is characterized in that: Step S12 specifically includes: obtaining the user browsing content, and obtaining the user interest tag set based on the user browsing content by using natural language processing technology, such as formula: ， ， in, Browse content for users, The first time a user browses content i content items, n is the number of content items browsed by the user, A collection of user interest tags. It stands for Natural Language Processing.

4. The method for recommending blind boxes of additional interest tags for breaking the information cocoon according to claim 2 is characterized in that: Step S13 specifically includes: according to the user interest tag set, using a similarity measurement method, obtaining weakly related tags, such as formula: ， ， in, Indicates similarity, For potential tags, A collection of user interest tags. Indicates taking the maximum value, represents the cosine similarity, Yes Tags The vector representation of For user interest tags, Yes Tags The vector representation of is a weakly related label, and are the lower and upper bounds of the similarity.

5. The method for recommending blind boxes of additional interest tags for breaking the information cocoon according to claim 2 is characterized in that: Step S14 specifically includes: obtaining an opposing tag pair, and obtaining an opposite tag according to the user interest tag set and the opposing tag pair, such as the formula: ， ， Among them, C is the opposing label pair, and etc. are pairs of labels with opposite opinions. is a set of opposite labels, For potential labels, For user interest tags, A collection of user interest tags.

6. The method for recommending blind boxes of additional interest tags for breaking the information cocoon according to claim 2 is characterized in that: The preprocessing includes removing duplicate tags and merging similar tags.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for recommending additional interest tag blind boxes to break the information cocoon as described in any one of claims 1-6 are implemented.

8. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method for recommending additional interest tag blind boxes to break the information cocoon as described in any one of claims 1-6 are implemented.

9. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method for recommending additional interest tags blind boxes to break the information cocoon as described in any one of claims 1-6 are implemented.

Images