Book recommendation method and device, storage medium, equipment and program product

Abstract

The invention discloses a book recommendation method and device, a storage medium, equipment and a program product. The method comprises the steps of obtaining a target topic matched with a target book in multiple modes; a recommendation reason matched with the target book is generated according to post contents of related posts under the target topic, the recommendation reason is used for establishing association between the target book and the target topic, and the related posts are posts related to the target book under the target topic; and displaying the target book and the recommendation reason. According to the method, the target topic highly related to the book is introduced, and the recommendation reason is generated based on the post content of the related post under the target topic, so that the richness and attraction of the recommended copywriting are increased, the book recommendation effect is remarkably improved, and the click conversion rate and the overall reading experience of the book are improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a book recommendation method, apparatus, storage medium, device and program product. Background Art

[0002] In the digital reading era, book recommendation systems are an important bridge between readers and books. Traditional book recommendation systems mainly display book covers, etc., and cannot quickly help users make decisions on whether to read. Users may click in, read a few chapters, and find that they are not interested, so the recommendation efficiency is low. Summary of the invention

[0003] The embodiments of the present application provide a book recommendation method, apparatus, storage medium, device and program product, which can increase the richness and attractiveness of the recommendation copy by introducing a target topic that is highly relevant to the book, and generate recommendation reasons based on the post content of related posts under the target topic, thereby significantly improving the effect of book recommendation, thereby improving the click-through conversion rate of the book and the overall reading experience.

[0004] On the one hand, an embodiment of the present application provides a book recommendation method, the method comprising:

[0005] Obtain target topics that match target books through various means;

[0006] Generate a recommendation reason matching the target book according to the content of the related posts under the target topic, wherein the recommendation reason is used to establish an association between the target book and the target topic, and the related posts are posts under the target topic that are related to the target book;

[0007] The target book and the reason for the recommendation are displayed.

[0008] On the other hand, an embodiment of the present application provides a book recommendation device, the device comprising:

[0009] an acquisition unit, configured to acquire association data between a target book and a plurality of topics, wherein the association data includes at least the number of unique visitors to each topic, the number of replies with the book to each topic, the number of page views of the target book in each topic, the reading conversion rate of the target book in each topic, and the irrelevance score of the target book to each topic;

[0010] a determination unit, configured to determine a target topic matching the target book from the plurality of topics according to a preset matching condition and the associated data;

[0011] A generating unit, configured to generate a recommendation reason matching the target book according to the target topic, wherein the recommendation reason is used to establish an association between the target book and the target topic;

[0012] The display unit is used to display the target book and the reason for recommendation.

[0013] On the other hand, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program is suitable for being loaded by a processor to execute the book recommendation method described in any of the above embodiments.

[0014] On the other hand, an embodiment of the present application provides a computer device, which includes a processor and a memory, wherein a computer program is stored in the memory, and the processor executes the book recommendation method described in any of the above embodiments by calling the computer program stored in the memory.

[0015] On the other hand, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the book recommendation method described in any of the above embodiments.

[0016] The embodiment of the present application obtains a target topic that matches a target book in a variety of ways; generates a recommendation reason that matches the target book based on the content of the relevant posts under the target topic, and the recommendation reason is used to establish an association between the target book and the target topic, and the relevant posts are posts under the target topic that are related to the target book; and displays the target book and the recommendation reason. The embodiment of the present application introduces a target topic that is highly relevant to the book, and generates a recommendation reason based on the content of the relevant posts under the target topic, thereby increasing the richness and attractiveness of the recommendation copy, significantly improving the effect of book recommendation, and thus improving the click-through conversion rate of the book and the overall reading experience. Specifically, target topics matching target books are obtained through a variety of methods, and recommendation reasons are generated based on the content of related posts under the target topics, so that the recommendation copy is no longer limited to the book's brief introduction information, but integrates richer and more specific content. This method not only increases the amount of information in the recommendation copy, but also enhances its attractiveness and persuasiveness, making it easier to stimulate users' interest in reading; the recommendation reason is directly based on the association between the target book and the target topic, and the related posts are content closely related to the target book under the target topic. This strong correlation construction makes the recommendation reason closer to the user's actual needs and interests, and improves the accuracy and effectiveness of the recommendation; due to the improvement in the richness and attractiveness of the recommendation copy, and the enhancement of the correlation between the book and the recommendation reason, users are more likely to have the urge to click and read when they see the recommendation, which can quickly help users make a decision whether to read, thereby improving the click-through rate of the book and the overall reading experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0018] Figure 1 A flowchart of a book recommendation method provided in an embodiment of the present application.

[0019] Figure 2 A schematic diagram of a first application scenario of the book recommendation method provided in an embodiment of the present application.

[0020] Figure 3 A schematic diagram of a second application scenario of the book recommendation method provided in an embodiment of the present application.

[0021] Figure 4 A schematic diagram of a third application scenario of the book recommendation method provided in an embodiment of the present application.

[0022] Figure 5 A schematic diagram of a fourth application scenario of the book recommendation method provided in an embodiment of the present application.

[0023] Figure 6 A schematic diagram of the structure of a book recommendation device provided in an embodiment of the present application.

[0024] Figure 7 A schematic diagram of the structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0025] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0026] The embodiments of the present application provide a method, apparatus, storage medium, device and program product for recommending books. Specifically, the method for recommending books in the embodiments of the present application can be executed by a terminal device or by a server. Among them, the computer device can be a terminal device or a server. The terminal device can be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart TV, a smart speaker, a wearable smart device, a smart car terminal and other devices. The terminal device can also include a client, and the client can be a client of an application that can recommend books. For example, the client includes at least one of a program client and a web client. For example, the application can be a social application, a chat robot application, a customer service application, etc. For example, the application can also be other applications with chat functions, such as reading, shopping, video, music, games, financial management, office and other applications with chat functions. The server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDN), and basic cloud computing services such as big data and artificial intelligence platforms.

[0027] The embodiments of the present application can be applied to various application scenarios such as book recommendation and topic recommendation.

[0028] The following are detailed descriptions. It should be noted that the description order of the following embodiments is not intended to limit the priority order of the embodiments. The following takes the execution subject as a terminal as an example to illustrate the book recommendation method provided in the embodiment of the present application.

[0029] It is understandable that in the specific implementation of this application, related data such as user reading history information and feedback data are involved. When the embodiments of this application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of relevant data need to comply with relevant laws, regulations and standards of relevant countries and regions.

[0030] See also Figures 1 to 5 , Figure 1 A flowchart of a book recommendation method provided in an embodiment of the present application is shown below. Figures 2 to 5 These are schematic diagrams of application scenarios of the book recommendation method provided in the embodiments of the present application. The method comprises the following steps:

[0031] Step 110, obtaining a target topic matching the target book in a variety of ways.

[0032] For example, step 110 can collect topics that match the content, style, audience, and other characteristics of the target book through various channels and strategies. This includes but is not limited to using data analysis technology to identify potential connections between the target book and existing topics, monitoring hot discussions in social media and online forums to determine related topics, and predicting topics that may be of interest by analyzing user reading history and interest preferences. Through these diversified methods, it is ensured that there is a close and meaningful connection between the selected topic and the target book.

[0033] In some embodiments, step 110 may be implemented by the following step 111 or step 112, specifically:

[0034] Step 111 : determining a target topic matching the target book from the multiple topics according to association data between the target book and the multiple topics.

[0035] For example, association data between a target book and multiple topics is obtained, wherein the association data includes at least the number of independent visitors to each topic, the number of replies with books to each topic, the number of page views of the target book in each topic, the reading conversion rate of the target book in each topic, and the irrelevance score of the target book to each topic.

[0036] The book recommendation method provided in the embodiment of the present application can be applied to different scenarios such as electronic reading platforms and social media platforms. Taking the electronic reading platform as an example, the electronic reading platform includes a recommendation page for displaying topics or books. In the recommendation page, the target book and the recommendation reasons matching the target book can be displayed. For example, in response to the user opening the electronic reading platform, before displaying the recommendation page, the association data of the target book and multiple topics can be obtained.

[0037] For example, a data source may be obtained from a database, log file, or third-party application programming interface (API) corresponding to a platform (such as an electronic reading platform, a social media platform, a UGC platform, etc.) or a forum, and the data source may include metadata of the target book, metadata of the topic, and user reading history information of the user for the target book and topic, etc. Among them, the metadata of the target book may include the title, author, publisher, publication time, classification, content introduction, ratings and comments, recommendation information, etc. For example, the metadata of a topic may include the title of the topic, copywriting information, etc. The user reading history information may include user behavior records in the target book and different topics, such as clicks, readings, comments, sharing, likes, browsing, etc.

[0038] Then, extract the following associated data from the data source:

[0039] Unique visitors per topic: indicates the number of users who are interested in the topic.

[0040] Number of replies with books in each topic: indicates the number of replies that mentioned or discussed the target book in the topic.

[0041] Page views of the target book in each topic: indicates the user's attention to the target book's page in a specific topic.

[0042] The reading conversion rate of the target book in each topic: it indicates the proportion of users who browse the target book and then actually read it.

[0043] Irrelevant score of the target book for each topic: calculated by the algorithm, indicating the degree of mismatch between the target book and the topic content.

[0044] For example, in order to ensure that the extracted related data is accurate, necessary data cleaning and standardization can be performed to facilitate subsequent analysis and matching.

[0045] In some embodiments, the obtaining of associated data between the target book and multiple topics includes: obtaining user reading history information on a target reading platform, the user reading history information including common reading history information between the producer of the topic and the consumer of the target book; and obtaining associated data between the target book and multiple topics by analyzing the user reading history information.

[0046] For example, user reading history information on the target reading platform (such as online bookstores, e-book reading applications, forums, etc.) is collected, including but not limited to the user's browsing history, click history, comment history, collection history, reading history, like history, etc. on the platform. Special attention is paid to the common reading history information of the topic producer (i.e., the user who publishes the topic) and the consumer of the target book (i.e., the user who reads the book), which helps to discover the potential connection and interest points between the two. By analyzing the common reading history information of the topic producer and the consumer of the target book, the associated data of the target book and multiple topics is obtained, including the number of independent visitors, the number of replies with books, the number of page views, the reading conversion rate, the irrelevant score, etc.

[0047] For example, before obtaining the associated data of the target book and multiple topics, the method also includes: preprocessing the topics on the target reading platform, including classifying, deduplicating and labeling the topics, so as to facilitate the subsequent acquisition and analysis of associated data.

[0048] Classification: Classify topics into predefined categories, such as literature, history, technology, entertainment, etc., based on their content, theme, or field. Classification helps in the subsequent analysis and understanding of related data, as well as the generation of reasons for recommendations.

[0049] Deduplication: Identify and remove duplicate or highly similar topics to ensure that each topic is unique in the associated data. Deduplication can reduce data redundancy and improve analysis efficiency.

[0050] Tagging: Assign one or more keyword tags to each topic, such as keywords, themes, styles, etc. Tagging helps to quickly identify the core content of the topic and the potential connection between the target book and the topic.

[0051] The preprocessed topics and their related information are stored in the database for subsequent acquisition and analysis of related data.

[0052] In some embodiments, determining a target topic matching the target book from the multiple topics according to the association data between the target book and the multiple topics includes:

[0053] Determining, from the plurality of topics, a candidate topic that satisfies the preset matching condition according to the preset matching condition and the associated data;

[0054] The candidate topics are sorted from high to low according to their click rates, and the first N candidate topics are selected as target topics matching the target book, where N is a natural number greater than 1.

[0055] In some embodiments, the associated data includes at least the number of unique visitors to each topic, the number of replies with books to each topic, the number of page views of the target book in each topic, the reading conversion rate of the target book in each topic, and the irrelevance score of the target book to each topic. The preset matching conditions include:

[0056] The number of independent visitors is greater than or equal to the visitor number threshold, and the number of replies with books is greater than or equal to the reply number threshold;

[0057] The page views are greater than or equal to the page views threshold;

[0058] The reading conversion rate is greater than or equal to the conversion rate threshold;

[0059] The irrelevance score is less than or equal to a score threshold.

[0060] For example, multiple topics collected are screened according to preset matching conditions (such as the number of independent visitors, the number of replies with books, page views, reading conversion rate, irrelevant score, etc.). Only topics that meet all preset matching conditions at the same time will be included in the candidate topic set. For example, a topic must meet the following conditions: the number of independent visitors is greater than or equal to the visitor number threshold (such as 1000), the number of replies with books is greater than or equal to the reply number threshold (such as 20), the page views are greater than or equal to the page views threshold (such as 2), the reading conversion rate is greater than or equal to the conversion rate threshold (such as 0.2%), and the irrelevant score is less than or equal to the score threshold (such as 0.95, that is, excluding topics with irrelevant scores greater than 0.95) to become a candidate topic.

[0061] Then, the selected candidate topics are sorted from high to low according to their click-through rates. The click-through rate reflects the popularity of the topic and user interest, and is an important indicator for evaluating the attractiveness of the topic. The top N candidate topics after sorting are selected as the target topics that match the target book.

[0062] Among them, N is a natural number greater than 1, and its value can be adjusted according to actual business needs. For example, when more diverse recommendations are needed, the value of N can be increased; when more accurate recommendations are needed, the value of N can be appropriately reduced. For example, the value of N can also be different for different book types or user groups. For example, for popular novels, users may prefer to see more diverse recommendations, so a larger N value can be set; while for books with strong professionalism and a narrow audience, more accurate recommendations may be needed, so a smaller N value can be set.

[0063] Among them, the various thresholds in the preset matching conditions, such as the visitor number threshold, the reply number threshold, the pageview threshold, the conversion rate threshold, and the score threshold, can be dynamically adjusted according to factors such as time, region, and user group to adapt to different recommendation scenarios and needs. For example, during holidays or special events, user activity may increase. At this time, the visitor number threshold and the reply number threshold can be appropriately increased to ensure that the recommended topics have sufficient popularity and discussion. For example, for user groups in different regions, it may be necessary to adjust the threshold settings according to local cultural habits and user preferences.

[0064] In some embodiments, the associated data also includes the first number of comments, the first number of likes, and the first number of favorites of the target book in each topic, and the preset matching condition also includes:

[0065] The first number of comments is greater than or equal to a comment number threshold; or

[0066] The first number of likes is greater than or equal to a likes threshold; or

[0067] The first favorite number is greater than or equal to a favorite number threshold.

[0068] For example, in order to more comprehensively evaluate the relevance between the target book and multiple topics, the associated data can be further enriched and more indicators can be included in the preset matching conditions. Based on the original associated data, the embodiment of the present application adds the following three indicators:

[0069] First number of comments: refers to the number of comments received by the target book in each topic. The number of comments reflects the user's attention and discussion enthusiasm for the books in the topic, and is an important indicator for evaluating the relevance of books and topics.

[0070] First, the number of likes: refers to the number of likes that the target book receives in each topic. The number of likes reflects the degree of recognition and support of users for the books in the topic, and is also a key factor in measuring the relevance of books to topics.

[0071] First, the number of collections: refers to the number of times the target book is collected by users in each topic. The number of collections reflects the user's interest in the books in the topic and their potential willingness to buy, and is also a key factor in measuring the relevance of books to topics.

[0072] The embodiment of the present application adds a matching condition corresponding to the above-mentioned newly added associated data on the basis of the original preset matching condition:

[0073] The first review number condition requires that the number of first reviews of the target book in each topic is greater than or equal to the review number threshold (such as 10). This condition ensures that the selected topic has sufficient user discussion and feedback, thereby improving the credibility and persuasiveness of the book recommendation.

[0074] First likes condition: requires that the first likes of the target book in each topic is greater than or equal to the likes threshold (such as 30). This condition reflects the positive attitude and recognition of users in the topic towards the book, which helps to improve the attractiveness and conversion rate of book recommendations.

[0075] The first collection number condition: requires that the first collection number of the target book in each topic is greater than or equal to the collection number threshold (such as 20). This condition reveals the user's potential interest and willingness to buy the book, which is of great significance for optimizing the book recommendation strategy.

[0076] It should be noted that the above newly added matching conditions do not have to be met at the same time, but you can flexibly select one of them or use them in combination. The specific selection of one or more conditions can be comprehensively considered and dynamically adjusted based on actual business needs, recommendation scenarios, user group characteristics and other factors.

[0077] By expanding the associated data and updating the preset matching conditions, we can more comprehensively evaluate the correlation between the target book and multiple topics, thereby more accurately determining the target topic that matches the target book, further improving the effect of book recommendations and user experience.

[0078] In some embodiments, determining a target topic matching the target book from the multiple topics according to the preset matching condition and the associated data includes:

[0079] Performing cluster analysis on the multiple topics, grouping similar topics into one category, and obtaining clustered topics;

[0080] According to the preset matching condition and the associated data, a target topic matching the target book is determined from the clustered topics.

[0081] For example, cluster analysis technology can be introduced to improve the accuracy and efficiency of topic matching. Cluster analysis is an unsupervised learning method that can classify similar objects into one category, thereby revealing the inherent structure and pattern in the data. In the book recommendation system, cluster analysis can be used to group multiple topics, classify similar topics into one category, and then determine the target topic that matches the target book from these clustered topics.

[0082] For example, before cluster analysis, the collected topics need to be preprocessed, including removing duplicate topics, cleaning invalid data, and text vectorization of topics. Text vectorization is the process of converting text data into numerical features.

[0083] Then, select a suitable clustering algorithm to perform cluster analysis on the preprocessed topics. The selection of clustering algorithm should be based on the characteristics of topic data, the expectation of clustering results, and the limitation of computing resources. Through clustering analysis, similar topics are grouped together to obtain a clustered topic set. Topics in each cluster have high similarity in terms of content, theme, or user group.

[0084] Then, in the clustered topic set, according to the preset matching conditions (such as the number of independent visitors, the number of replies with books, the number of page views, the reading conversion rate, the irrelevant score, the number of comments, the number of likes, the number of collections, etc.) and the related data, the target topics that match the target books are screened out.

[0085] For each cluster, the topic that meets the matching conditions and has the highest relevance can be selected as the representative topic of the cluster. Then, the target topic that best matches the target book can be further screened from these representative topics.

[0086] If diverse recommendation results are needed, multiple topics that meet the requirements can be selected from each cluster as target topics; if the accuracy of the recommendation is pursued, the target topics may be selected only from one or a few clusters that are most relevant to the book.

[0087] Through cluster analysis, similar topics can be grouped together, making it easier to find topics that are highly relevant to the target book. Matching and screening in the clustered topic set can reduce the amount of calculation and improve the response speed of the recommendation system. By selecting target topics from multiple clusters, the diversity of recommendations can be increased to meet the different interests and needs of users.

[0088] Step 112: determining a target topic matching the target book from the multiple topics according to the semantic association between the target book and each post under the multiple topics.

[0089] For example, step 112 may use semantic analysis technology to evaluate the matching degree between the topic and the book by calculating the semantic association between the target book and each post under multiple topics (such as cosine similarity, embedding vector distance, etc.). Then, based on the level of semantic association, the topic that best matches the target book is selected from multiple topics as the target topic. This method can capture deeper semantic information and is suitable for situations where the association between the topic and the book is more obscure or complex.

[0090] In some embodiments, determining a target topic matching the target book from the multiple topics according to the semantic association between the target book and each post under the multiple topics includes:

[0091] Calculating the semantic relevance between the target book and each post under multiple topics;

[0092] Calculate the weighted average of the semantic relevances corresponding to all posts under each topic as the overall relevance between the target book and each topic;

[0093] Topics whose overall relevance exceeds a relevance threshold are determined as target topics that match the target book.

[0094] First, the semantic relevance between the target book and each post under multiple topics is calculated. For example, the abstract or core content of the target book and the posts under each topic are first preprocessed to obtain the preprocessed text data, including the removal of stop words, stem extraction, part-of-speech tagging, etc., so as to perform subsequent semantic analysis. Then, natural language processing techniques such as word embedding, term frequency-inverse document frequency (TF-IDF), etc. are used to perform semantic feature extraction on the preprocessed text data to extract the semantic features of the target book and the semantic features of the posts. Semantic features can capture the semantic meaning and contextual relationship of words in the text. Then, cosine similarity, Euclidean distance or other semantic similarity measurement methods are used to calculate the semantic relevance between the semantic features of the target book and the semantic features of each post. The semantic relevance reflects the semantic similarity between the content of the book and the content of the post.

[0095] Then, the weighted average of the semantic relevance of all posts under each topic is calculated. For example, considering that different posts may have different importance and influence in a topic, a weight can be set for each post, and the weight can be determined based on factors such as the number of likes, the number of replies, and the authority of the publisher. Then, based on the semantic relevance of each post and the weight corresponding to each post, the weighted average of the semantic relevance of all posts under each topic is calculated. This weighted average reflects the overall semantic relevance of the target book to the entire topic.

[0096] Then, the topics whose overall relevance exceeds the relevance threshold are determined as target topics that match the target book. For example, a reasonable relevance threshold can be set according to the needs of the actual application and the distribution of the data. The relevance threshold is used to distinguish between topics that are highly relevant to the target book and those that are less relevant. The overall relevance of each topic is then compared with the relevance threshold. If the overall relevance of a topic exceeds the relevance threshold, it is determined as a target topic that matches the target book. The target topic will be used as part of the recommendation information to improve the richness of the book recommendation copy and the click-through conversion rate of the book.

[0097] Through the above steps, based on the semantic relevance between the target book and each post under multiple topics, the target topic matching the target book can be accurately screened out, thereby providing users with more personalized and accurate book recommendations.

[0098] Step 120, generating a recommendation reason matching the target book according to the content of the related posts under the target topic, wherein the recommendation reason is used to establish an association between the target book and the target topic, and the related posts are posts under the target topic related to the target book.

[0099] For example, after determining the target topics related to the target book, step 120 focuses on extracting valuable information from the relevant posts under these target topics to construct personalized and attractive recommendation reasons. This involves in-depth analysis of the content of the posts, including sentiment, key information extraction, and point of view summary, aiming to capture key points that can highlight the characteristics of the book and stimulate the reader's interest. The generated recommendation reasons not only reflect the close connection between the book and the topic, but also incorporate real user feedback and unique insights in the posts, thereby enhancing the credibility and attractiveness of the recommendation.

[0100] For example, you can analyze the content of related posts under the target topic, extract keywords or sentences related to the target book, and combine the introduction or main content of the target book to generate a concise and clear recommendation reason. The recommendation reason should highlight the connection between the target book and the target topic to stimulate the user's reading interest.

[0101] In some embodiments, the method further includes: obtaining a correlation score between the target book and each post under the target topic, and sorting the posts from high to low according to the correlation score, and selecting the first M posts as the relevant posts, where M is a natural number greater than 1.

[0102] First, obtain the relevance score between the target book and each post under the target topic. For example, perform text preprocessing on the abstract, content overview or key chapters of the target book and each post under the target topic. Text preprocessing may include steps such as removing stop words, performing stem extraction, and part-of-speech tagging for subsequent semantic analysis. Then, natural language processing techniques (such as word embedding, TF-IDF, etc.) are used to perform feature extraction processing to extract the semantic features of the target book and the semantic features of the post. Semantics can capture key words, phrases and their contextual relationships in the text. Then, cosine similarity, Euclidean distance or other semantic similarity measurement methods are used to calculate the similarity or correlation between the feature vector of the target book and the feature vector of each post, thereby obtaining a relevance score between them. This score reflects the semantic similarity or correlation strength between the target book and the post.

[0103] Then, sort the posts from high to low according to the relevance score, and select the first M posts as relevant posts. For example, sort all posts under the target topic according to the calculated relevance score. The sorting is based on the score, that is, the higher the score, the stronger the relevance of the post to the target book. Then, in the sorted post list, select the first M posts as the posts most relevant to the target book (M is a natural number greater than 1). The specific value of M can be dynamically adjusted according to actual business needs, capacity limitations of the recommendation system, and user reading habits.

[0104] For example, when selecting relevant posts, the diversity of the posts can also be considered to ensure that the recommendation reasons can cover multiple aspects and characteristics of the book. For example, different categories or topic tags can be set, and a certain number of posts can be selected from each category or topic to achieve diversity and comprehensiveness of the recommendation.

[0105] In some embodiments, obtaining the relevance score between the target book and each post under the target topic includes: performing keyword matching on each post under the target topic and the target book to obtain the keyword matching degree of each post; performing semantic association analysis on each post under the target topic and the target book to obtain the semantic association degree of each post and the target book; obtaining the popularity of each post under the target topic; and determining the relevance score between the target book and each post under the target topic based on the keyword matching degree, the semantic association degree and the popularity.

[0106] First, a set of keywords closely related to the target book is identified. These keywords include, but are not limited to, the book name, the author's name, the name of a specific character in the book, important events in the book, scene descriptions, book style, book type, etc. By combining exact matching and fuzzy matching, the keyword matching degree between each post and the target book can be calculated. The keyword matching degree can be calculated using a variety of methods, such as word frequency statistics, word frequency-inverse document frequency (TF-IDF), etc. These methods can quantify the frequency and importance of keywords in a post, thereby reflecting the degree of relevance between the post and the target book in terms of content.

[0107] For example, in addition to keyword matching, semantic association analysis is also performed to capture the deeper association between posts and target books, including but not limited to: Semantic role labeling: Identify and label the semantic roles of texts in posts and books to reveal actions, events, and relationships in the text. Named entity recognition: Identify proper nouns in posts and books, such as names of people, places, and institutions, to establish entity associations between texts. Sentiment analysis: Analyze the emotional tendencies in posts and books, such as positive, negative, or neutral, to evaluate their emotional resonance or differences. Through these analyses, the semantic association between each post and the target book can be calculated, reflecting their similarities and differences in content, theme, and sentiment.

[0108] In addition, we also pay attention to the popularity of each post under the target topic. Popularity can be measured by a variety of indicators, such as the number of likes, replies, reposts, shares, comments, comment quality, and user interaction time of the post. These indicators can reflect the popularity and influence of the post under the target topic.

[0109] Then, after obtaining information on the three dimensions of keyword matching, semantic relevance, and post popularity, this information is comprehensively considered. The specific comprehensive method can be weighted average method, linear regression method or other complex machine learning models. These methods can determine the final relevance score based on the relative importance of different dimensions and actual business needs. The relevance score will be used to screen the posts that are most relevant to the target book. This score can fully reflect the comprehensive relevance between the post and the target book in terms of content, theme, emotion and popularity.

[0110] In some embodiments, the method further includes: reading and parsing the text content of each related post to extract key information in each related post; classifying and organizing the key information to obtain processed related information; and constructing the post content of each related post based on the processed key information.

[0111] First, read and parse the text content of each relevant post to extract the key information in each relevant post. For example, first use crawler technology or API interface to crawl the text content of relevant posts from the target topic. The text content may contain a variety of information such as user evaluation of the book, description of the book content, comparison with other books or topics, etc. Then use natural language processing (NLP) technology to parse the text content. This step may include word segmentation, part-of-speech tagging, named entity recognition, etc., so as to extract key information later.

[0112] For example, key information refers to information fragments that can reflect the core content of the post and the user's point of view. Key information can include the evaluation of the target book in related posts, the description of the book content, and the comparison with other books or topics.

[0113] Then, the key information is classified and organized to obtain the processed related information. For example, in order to make better use of these key information, they need to be classified: for example, evaluation information can be divided into positive evaluation, negative evaluation or neutral evaluation; description information can be divided into plot description, character introduction or background setting, etc. This classification helps to understand the content of the post more clearly and lays the foundation for subsequent content construction. For example, based on the classification, the extracted content fragments need to be organized: this includes deduplication to avoid repeated information in the recommendation reasons; and integration, that is, according to the relevance and logic of the content, the fragments are appropriately integrated to form a coherent and organized content collection. This organization helps to improve the readability and attractiveness of the recommendation reasons.

[0114] Then, based on the processed key information, construct the post content of each related post. This step needs to comprehensively consider factors such as the topic of the post, the user's point of view, the logical order of information, etc., to ensure that the constructed post content is both accurate and easy to understand. During the construction process, the post content can also be appropriately optimized. For example, appropriate transition sentences or paragraphs can be added to enhance the coherence of the content; more attractive language or expressions can be used to enhance the attractiveness of the content.

[0115] In some embodiments, the content of the related posts below generates recommendation reasons matching the target book, including:

[0116] Performing content optimization processing on the post contents of the relevant posts under the target topic to obtain optimized post copy information, wherein the content optimization processing includes at least one of the following: semantic analysis, keyword extraction, keyword reorganization, deduplication processing, adjusting copy length, adjusting copy display style, and adjusting copy style;

[0117] A compliance review is performed on the optimized post copy information, and the post copy information that passes the review is determined as a recommendation reason for matching the target book, wherein the compliance review includes at least one of the following: filtering sensitive words and verifying authenticity.

[0118] For example, we can perform semantic analysis on the posts of relevant posts under the target topic to ensure that they are closely related to the theme and content of the target book. We can also use natural language processing technology to identify the core concepts and themes in the posts to ensure their consistency with the book content.

[0119] For example, keywords are extracted and reorganized from the content of relevant posts under the target topic to highlight the relevance and appeal of the target book to the target topic. Reasonable keyword layout and expression can make the recommendation reason more concise and persuasive.

[0120] For example, remove duplicate content in the post, avoid redundant information, and ensure the conciseness and efficiency of the reasons for the recommendation.

[0121] For example, according to the requirements of the recommended scenario and display space, adjust the length of the copy so that it can convey the most effective information in a limited space.

[0122] For example, adjust the display style of the copy, optimize the layout and format of the copy, such as using bold, italics, underline, etc. to highlight key information, and improve the readability and attractiveness of the copy.

[0123] For example, according to the type and style of the target book, the emotional color or language style of the post content can be adjusted to enhance the appeal of the recommendation reason. For example, for a light-hearted and humorous novel, a humorous copywriting style can be adopted; for a professional and rigorous academic book, a formal and rigorous expression method can be adopted.

[0124] For example, the optimized post copy information is filtered for sensitive words to ensure that the post copy information complies with laws, regulations and social ethics and does not contain illegal, vulgar, violent or discriminatory content.

[0125] For example, verify the authenticity, accuracy and completeness of copywriting information to avoid misleading users.

[0126] For example, check whether there are other potential compliance issues in the copywriting information, such as infringement of other people's intellectual property rights, disclosure of personal privacy, etc., to ensure the legality and security of the copywriting information.

[0127] Through the above-mentioned content optimization processing and compliance review steps, it is possible to generate recommendation reasons that are both highly matched with the target books and attractive, while ensuring the legality and authenticity of the copy content, providing users with a better reading recommendation experience.

[0128] In some embodiments, generating a recommendation reason matching the target book based on the content of the related posts under the target topic includes:

[0129] According to the weights of the relevant posts under the target topic and the post contents of the relevant posts, a recommendation reason matching the target book is generated.

[0130] In some embodiments, the method further comprises:

[0131] The weight of the related post is adjusted according to the activity of the related post, where the activity includes the number of posts, the second number of comments, the second number of likes, and the first number of shares of the related post within a set time period.

[0132] In some embodiments, the method further comprises:

[0133] Analyzing multi-dimensional information of the target book, the multi-dimensional information including book type, author popularity, book rating, reader evaluation, and book sales;

[0134] The weights of the related posts are adjusted according to the analysis results of the multi-dimensional information.

[0135] For example, when there are multiple target topics that match the target book, in order to generate more accurate and attractive book recommendation reasons, a multi-factor fusion strategy can be adopted. This strategy not only considers the weight of related posts under the target topic and the content of related posts, but also combines the activity of related posts and the multi-dimensional information of the target book, thereby realizing the personalization and precision of the recommendation reasons.

[0136] First, extract the content of relevant posts related to books from the target topic, which will serve as the basis for the recommendation reasons. Then, filter and reorganize the extracted content of posts based on the weights of the relevant posts. For relevant posts with high weights, their text information will account for a higher proportion in the recommendation reasons to reflect their importance and influence.

[0137] The activity of related posts can be regularly checked, including the number of posts within a set time period, the second number of comments (the number of existing comments in the related post), the second number of likes (the like behavior of the related post) and the first number of shares (the direct sharing behavior of the related post). Related posts with high activity indicate that they are more popular with users, so they should be given a higher weight in the recommendation reasons.

[0138] For example, the weight of related posts can be adjusted by combining the multi-dimensional information of the target book. For example, a comprehensive analysis of the target book can be conducted, covering multi-dimensional information such as book type, author popularity, book ratings, reader reviews, and book sales. This information helps to deeply understand the characteristics and audience of the book, so as to more accurately match related posts and adjust the weight of related posts.

[0139] Then, based on the results of the multi-dimensional information analysis of the target book, the weight of the relevant posts matching the target book is fine-tuned. For example, if the book has a high rating and good sales, the weight of the relevant posts related to the book's quality should be appropriately increased; if the book type is unique and the audience is small, the weight of the relevant posts related to the book type should also be increased accordingly.

[0140] This multi-factor integrated book recommendation reason generation strategy can provide more accurate and personalized recommendation services, improving users' reading experience and the click-through conversion rate of books.

[0141] Step 130: Display the target book and the reason for recommendation.

[0142] For example, step 130 focuses on presenting the target book and its matching recommendation reasons to the user in a collaborative manner. The display strategy aims to maximize the visibility and influence of the book and the recommendation reason, and may include embedding the recommendation reason in the book detail page, highlighting it in the recommendation list or ad space, and actively pushing it through email or in-app notifications. This collaborative display not only increases the exposure of the book, but also promotes the user's click and reading behavior through persuasive recommendation reasons, thereby optimizing the user experience and improving the conversion rate of the book.

[0143] In some embodiments, the displaying of the target book and the reason for recommendation includes:

[0144] Display the target book and the reason for recommendation on the recommendation page of the target reading platform; or

[0145] The recommendation reason is displayed on the detail page of the target book.

[0146] For example, on the recommendation page of the target reading platform (such as new book recommendations, popular books, editors' picks, etc.), not only the basic information of the target book (such as title, author, cover, etc.) is displayed, but also the reason for the recommendation. This method can help users quickly understand the characteristics and highlights of the book and increase the click-through rate and conversion rate of the book. This can be achieved in the form of a list, grid or carousel, with a brief recommendation reason next to each book to attract user attention.

[0147] like Figure 2 As shown, on the recommendation page 1 of the target reading platform, multiple target books 10 and the recommendation reasons 20 corresponding to each target book 10 are displayed at the same time. For example, book A and the recommendation reason a corresponding to book A are displayed; book B and the recommendation reason b corresponding to book B are displayed; book C and the recommendation reason c corresponding to book C are displayed, etc.

[0148] For example, on the target book's details page, the recommendation reason can be displayed in a prominent position on the details page, such as at the top of the page, above the book introduction, or above the user comment area. Make sure that the display of the recommendation reason does not interfere with the user's acquisition of key information, such as the book's title, author, price, etc.

[0149] like Figure 3 As shown, on the details page 2 of the target book 10 (book A), the basic information of the target book 10 (such as the title, author, cover, book introduction, number of likes, number of comments, number of collections, number of readings, etc.) is displayed, and the recommendation reason 20 is also displayed in a prominent position on the details page 2, such as above the book introduction.

[0150] For example, the recommendation reason 20 may include a concise recommendation reason a1 and a detailed recommendation reason a2.

[0151] For example, concise recommendation reason a1 is used to quickly attract readers' attention and stimulate their interest in target book 10 (book A). Concise recommendation reason a1 can concisely and intuitively convey the core characteristics or uniqueness of the book, attract readers' attention and stimulate their interest in reading; it usually contains several keywords or phrases, which can quickly summarize information such as the book's theme, style, reader group or author's influence.

[0152] For example, detailed recommendation reason a2 can be a further development and deepening of the concise recommendation reason, used to provide readers with a more comprehensive and in-depth introduction to the book. Detailed recommendation reason a2 usually contains more specific content, such as the theme, plot, character creation, writing style, author background, etc. of the book, as well as the potential value and significance of the book to readers.

[0153] For example, the display method can be a variety of display methods such as combining pictures and texts, dynamic display, adding interactive elements, etc.

[0154] For example, use a combination of images and text to display the recommendation reasons, such as adding a brief text of the recommendation reasons next to the book cover, or using icons and labels to highlight the keywords in the recommendation reasons.

[0155] For example, use dynamic effects (such as animation and sliding display) to attract users' attention and make the reasons for recommendation more vivid and attractive.

[0156] For example, add interactive elements such as hovering over the mouse to display more details about the recommendation reason, or clicking on the recommendation reason to expand the full copy.

[0157] In some embodiments, the method further comprises:

[0158] generating a label of the target book, wherein the label is used to describe the attributes of the target book;

[0159] When the recommendation reason is displayed, the label of the target book is also displayed.

[0160] For example, one or more tags are generated for the target book. Tags are used to describe the book's attributes such as theme, style, audience, word count, status, etc. Tags can be keywords, phrases, or short sentences that can quickly convey the core information of the book and help users quickly understand the general outline of the book. These tags can be generated based on multi-dimensional data such as book content analysis, author information, and reader reviews.

[0161] When displaying the recommendation reasons, the target book's tags are also displayed to enhance the user's understanding of the book. Tags can be placed next to the book cover, below the recommendation reasons, or other prominent locations on the page. Different colors, icons, or font styles can be used to distinguish different categories of tags to enhance the identifiability and appeal of the tags. Users can be allowed to click on tags for further searches or filtering to quickly find books that match a specific theme, style, or audience, which can enhance the user's browsing experience and improve search efficiency.

[0162] like Figure 3 As shown, on the details page 2 of the target book 10 (book A), basic information of the target book 10 (such as the title, author, cover, book introduction, number of likes, number of comments, number of collections, number of readings, etc.) is displayed, and the recommendation reason 20 (such as a concise recommendation reason a1 and a detailed recommendation reason a2) is displayed above the book introduction on the details page 2. In addition, a label 30 is displayed above the display position of the recommendation reason 20, for example, the label 30 may include label 1 (the genre is science fiction), label 2 (the status is completed) and label 3 (the word count is 300,000 words).

[0163] By displaying the target book's label while showing the recommendation reason, users can further enhance their understanding of the book and improve the accuracy and attractiveness of the recommendation. Users can quickly determine whether the book meets their interests and needs based on the label, so as to make a more informed reading choice. At the same time, the introduction of labels can also provide more powerful support for the promotion and marketing of books, helping books to better reach the target audience.

[0164] In some embodiments, the method further comprises:

[0165] generating an associated recommendation network of the target book, wherein the associated recommendation network is used to recommend books or authors associated with the target book;

[0166] When displaying the recommendation reason, the associated recommendation network of the target book is also displayed.

[0167] Among them, the association recommendation network is a recommendation system based on the association relationship between books or between books and authors. It aims to recommend one or more related books or authors for the target book, thus forming a more extensive and in-depth reading recommendation network.

[0168] For example, a large amount of data on books and authors is collected from the book recommendation system, including the book's title, author, category, content introduction, rating, comment, etc. The collected data is then preprocessed, such as removing duplicate data, cleaning invalid information, and performing text vectorization, etc., for subsequent analysis. Then, association rule mining, collaborative filtering, content filtering and other algorithms are used to mine the associations between books and between books and authors from the preprocessed data. These associations can be determined based on a variety of factors, such as the similarity of the subject of the book, the author relationship (such as the same author, different authors belonging to the same field, etc.), and reader behavior (such as co-purchase, co-evaluation, co-reading, etc.). Based on the mined associations, an association recommendation network is constructed for the target book. The network is centered on the target book and connects other books or authors associated with it. Each node in the network represents a book or an author, and each edge represents the association relationship and its strength between books or authors.

[0169] For example, when displaying the recommendation reason of the target book, the related recommendation network can be displayed next to or below it. The display method can adopt a graphical interface, such as a network diagram, a tree diagram or a list, so that the user can intuitively see other books or authors related to the target book.

[0170] like Figure 4 As shown, on the detail page 2 of the target book 10 (book A), the basic information of the target book 10 is displayed, and the recommendation reason 20 (such as a concise recommendation reason a1, and a detailed recommendation reason a2) is displayed above the book introduction on the detail page 2, and a label 30 is displayed above the display position of the recommendation reason 20. In addition, the associated recommendation network 40 is displayed below the display position of the recommendation reason 20. For example, book A and book H have similar themes, and the other works associated with the author G of book H are book L; for example, book A and book G have common reader behavior; for example, based on the author relationship, through the author A of book A, other works associated with author A also include book I, book J and book K.

[0171] like Figure 5 As shown, on the recommendation page 1 of the target reading platform, a plurality of target books 10 and the recommendation reasons 20 corresponding to each target book 10 are displayed simultaneously, as well as the tags 30 and the associated recommendation network 40 corresponding to each target book 10 .

[0172] For example, add interactive functions to the association recommendation network, such as clicking on a node to view detailed information, dragging a node to adjust the layout, filtering related books of a specific type, etc.

[0173] For example, a search function may also be provided to allow users to quickly find related books or authors of interest based on keywords or conditions.

[0174] For example, the associated recommendation network can also be personalized based on the user's reading history, preferences, and current context information (such as time, location, device, etc.).

[0175] For example, for users who like science fiction novels, science fiction books can be displayed preferentially in the associated recommendation network; for users browsing on mobile devices, short books suitable for reading on small screens can be recommended.

[0176] For example, a user is browsing a book A about programming. When showing the reasons for recommending the book, a related recommendation network can be shown at the same time, including other books related to programming (such as book B, book C, etc.), other books recommended by the author of the book (such as book D), and other works by the same author (such as E). In this way, users can not only understand the highlights and features of the current book, but also discover more information about books and authors related to their interests, thereby broadening their reading horizons and selection range.

[0177] By introducing an associated recommendation network and presenting it to users together with the reasons for the recommendation, we can provide richer and more personalized reading recommendation services and enhance users' reading experience and satisfaction.

[0178] In some embodiments, the method further comprises:

[0179] Acquire user feedback data on the recommendation reason, the feedback data including at least one of the following: a click rate, a stay time, a third number of comments, a third number of likes, and a second number of shares of the user on the recommendation reason;

[0180] The preset matching condition or the method for generating the recommendation reason is adjusted according to the feedback data.

[0181] For example, the collected user feedback data include but are not limited to the click rate of users on the recommendation reasons, the dwell time, the third number of comments (the number of comments of users on the recommendation reasons), the third number of likes (the number of users who liked the recommendation reasons), and the second number of shares (the number of users who shared the recommendation reasons). These data can fully reflect the user's acceptance and satisfaction with the recommendation reasons.

[0182] For example, the collection of feedback data can be achieved by setting up corresponding tracking codes or plug-ins on the user interface. For example, when a user clicks on a recommendation reason, views the details of a recommendation reason, posts a comment or likes a recommendation reason, or shares a recommendation reason, the system will automatically record these actions and generate corresponding feedback data.

[0183] For example, the collected feedback data will be used to analyze users' acceptance and satisfaction with the recommendation reasons. For example, a high click-through rate and a long stay time may indicate that the recommendation reason is highly attractive to users; while a large number of positive comments and likes may mean that the recommendation reason is widely recognized by users.

[0184] Based on the analysis results of the feedback data, the preset matching conditions can be dynamically adjusted. For example, if it is found that the user is particularly interested in certain types of topics or books, the weight of these topics or books in the matching conditions can be appropriately increased so that more relevant content can be presented in subsequent recommendations.

[0185] Feedback data can also guide the optimization of the generation of recommendation reasons. For example, if it is found that users frequently mention certain keywords or topics in their comments on certain recommendation reasons, you can consider incorporating more of these elements when generating recommendation reasons to improve the attractiveness and relevance of the recommendation reasons.

[0186] By continuously collecting and analyzing user feedback data, the performance and user experience of the recommendation system can be continuously iterated and improved.

[0187] All of the above technical solutions can be arbitrarily combined to form optional embodiments of the present application, which will not be described in detail here.

[0188] The embodiment of the present application obtains a target topic that matches a target book in a variety of ways; generates a recommendation reason that matches the target book based on the content of the relevant posts under the target topic, and the recommendation reason is used to establish an association between the target book and the target topic, and the relevant posts are posts under the target topic that are related to the target book; and displays the target book and the recommendation reason. The embodiment of the present application introduces a target topic that is highly relevant to the book, and generates a recommendation reason based on the content of the relevant posts under the target topic, thereby increasing the richness and attractiveness of the recommendation copy, significantly improving the effect of book recommendation, and thus improving the click-through conversion rate of the book and the overall reading experience. Specifically, target topics matching target books are obtained through a variety of methods, and recommendation reasons are generated based on the content of related posts under the target topics, so that the recommendation copy is no longer limited to the book's brief introduction information, but integrates richer and more specific content. This method not only increases the amount of information in the recommendation copy, but also enhances its attractiveness and persuasiveness, making it easier to stimulate users' interest in reading; the recommendation reason is directly based on the association between the target book and the target topic, and the related posts are content closely related to the target book under the target topic. This strong correlation construction makes the recommendation reason closer to the user's actual needs and interests, and improves the accuracy and effectiveness of the recommendation; due to the improvement in the richness and attractiveness of the recommendation copy, and the enhancement of the correlation between the book and the recommendation reason, users are more likely to have the urge to click and read when they see the recommendation, which can quickly help users make a decision whether to read, thereby improving the click-through rate of the book and the overall reading experience.

[0189] In order to better implement the book recommendation method of the present application embodiment, the present application embodiment also provides a book recommendation device. Figure 6 , Figure 6 This is a schematic diagram of the structure of a book recommendation device provided in an embodiment of the present application. The book recommendation device 200 may include:

[0190] An acquisition unit 210 is used to acquire a target topic matching a target book in a variety of ways;

[0191] A generating unit 220, configured to generate a recommendation reason matching the target book according to the content of the related posts under the target topic, wherein the recommendation reason is used to establish an association between the target book and the target topic, and the related posts are posts under the target topic that are related to the target book;

[0192] The display unit 230 is used to display the target book and the reason for recommendation.

[0193] In some embodiments, the generating unit 220 is further used to: obtain the relevance score between the target book and each post under the target topic, sort them from high to low according to the relevance score, and select the first M posts as the relevant posts, where M is a natural number greater than 1.

[0194] In some embodiments, when obtaining the relevance score between the target book and each post under the target topic, the generation unit 220 can be used to: perform keyword matching on each post under the target topic and the target book to obtain the keyword matching degree of each post; perform semantic association analysis on each post under the target topic and the target book to obtain the semantic association degree of each post and the target book; obtain the popularity of each post under the target topic; and determine the relevance score between the target book and each post under the target topic based on the keyword matching degree, the semantic association degree and the popularity.

[0195] In some embodiments, the generation unit 220 is further used to: read and parse the text content of each related post to extract key information in each related post; classify and organize the key information to obtain processed related information; and construct the post content of each related post based on the processed key information.

[0196] In some embodiments, the acquisition unit 210 is used to: determine a target topic matching the target book from the multiple topics based on association data between the target book and the multiple topics; or determine a target topic matching the target book from the multiple topics based on semantic associations between the target book and each post under the multiple topics.

[0197] In some embodiments, when determining a target topic that matches the target book from the multiple topics based on the associated data between the target book and the multiple topics, the acquisition unit 210 is used to: determine candidate topics that meet the preset matching conditions from the multiple topics based on the preset matching conditions and the associated data; sort the candidate topics from high to low according to their click-through rates, and select the top N candidate topics as target topics that match the target book, where N is a natural number greater than 1.

[0198] In some embodiments, the associated data includes at least the number of independent visitors for each topic, the number of replies with book links for each topic, the number of page views of the target book in each topic, the reading conversion rate of the target book in each topic, and the irrelevant score of the target book to each topic, and the preset matching conditions include: the number of independent visitors is greater than or equal to the visitor number threshold, and the number of replies with book links is greater than or equal to the reply number threshold; the page views are greater than or equal to the view count threshold; the reading conversion rate is greater than or equal to the conversion rate threshold; the irrelevant score is less than or equal to the score threshold.

[0199] In some embodiments, the associated data also includes the first number of comments, the first number of likes and the first number of collections of the target book in each topic, and the preset matching conditions also include: the first number of comments is greater than or equal to a comment number threshold; or the first number of likes is greater than or equal to a like number threshold; or the first number of collections is greater than or equal to a collection number threshold.

[0200] In some embodiments, the book recommendation device 200 also includes a processing unit for: obtaining user feedback data on the recommendation reason, the feedback data including at least one of the following: user click rate, dwell time, third number of comments, third number of likes, and second number of shares on the recommendation reason; adjusting the preset matching condition or the method of generating the recommendation reason according to the feedback data.

[0201] In some embodiments, when determining a target topic matching the target book from multiple topics based on the semantic associations between the target book and each post under multiple topics, the acquisition unit 210 is used to: calculate the semantic associations between the target book and each post under the multiple topics; calculate the weighted average of the semantic associations corresponding to all posts under each topic as the overall association between the target book and each topic; and determine a topic whose overall association exceeds a association threshold as a target topic matching the target book.

[0202] In some embodiments, the generation unit 220 is used to: perform content optimization processing on the post content of the relevant posts under the target topic to obtain optimized post copy information, wherein the content optimization processing includes at least one of the following: semantic analysis, keyword extraction, keyword reorganization, deduplication processing, adjusting the copy length, adjusting the copy display style, and adjusting the copy style; perform a compliance review on the optimized post copy information, and determine the post copy information that has passed the review as the recommended reason for matching the target book, wherein the compliance review includes at least one of the following: sensitive word filtering and authenticity verification.

[0203] In some embodiments, the generating unit 220 is used to generate a recommendation reason matching the target book according to the weights of the related posts under the target topic and the post contents of the related posts.

[0204] In some embodiments, the generating unit 220 is further used to adjust the weight of the related post according to the activity of the related post, where the activity includes the second number of comments, the second number of likes and the first number of shares of the related post within a set time period.

[0205] In some embodiments, the generation unit 220 is further used to: analyze multi-dimensional information of the target book, the multi-dimensional information including book type, author popularity, book rating, reader reviews, and book sales; and adjust the weight of the related posts according to the analysis results of the multi-dimensional information.

[0206] In some embodiments, the display unit 230 is used to: display the target book and the recommendation reason on a recommendation page of a target reading platform; or display the recommendation reason on a detail page of the target book.

[0207] In some embodiments, the display unit 230 is further used to: generate a label for the target book, where the label is used to describe the attributes of the target book; and display the label of the target book simultaneously when displaying the recommendation reason.

[0208] In some embodiments, the display unit 230 is further used to: generate an associated recommendation network of the target book, wherein the associated recommendation network is used to recommend books or authors associated with the target book; and when displaying the recommendation reason, simultaneously display the associated recommendation network of the target book.

[0209] The book recommendation device 200 may be integrated in a terminal device or a server that has a storage device and a processor and has computing capabilities, or the book recommendation device 200 is the terminal device or the server.

[0210] In some embodiments, the present application further provides a computer device, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0211] In some embodiments, Figure 7 As shown, Figure 7Another structural diagram of a computer device provided in an embodiment of the present application, the computer device 300 also includes a processor 301 having one or more processing cores, a memory 302 having one or more computer-readable storage media, and a computer program stored in the memory 302 and executable on the processor. The processor 301 is electrically connected to the memory 302. Those skilled in the art will appreciate that the computer device structure shown in the figure does not constitute a limitation on the computer device, and may include more or fewer components than shown, or combine certain components, or arrange components differently.

[0212] The processor 301 is the control center of the computer device 300. It uses various interfaces and lines to connect the various parts of the entire computer device 300, executes various functions of the computer device 300 and processes data by running or loading software programs and / or modules stored in the memory 302, and calling data stored in the memory 302, thereby monitoring the computer device 300 as a whole.

[0213] Optionally, the computer device 300 may be a terminal device, and the processor 301 may call a software program and a module stored in the memory 302 to perform the following operations:

[0214] A target topic matching a target book is obtained in a variety of ways; a recommendation reason matching the target book is generated based on the content of related posts under the target topic, wherein the recommendation reason is used to establish an association between the target book and the target topic, and the related posts are posts under the target topic related to the target book; and the target book and the recommendation reason are displayed.

[0215] The specific implementation of the above operations can be found in the previous embodiments, which will not be described in detail here.

[0216] In some embodiments, Figure 7 As shown, the computer device 300 further includes: a display unit 303, a radio frequency circuit 304, an audio circuit 305, an input unit 306, and a power supply 307. The processor 301 is electrically connected to the display unit 303, the radio frequency circuit 304, the audio circuit 305, the input unit 306, and the power supply 307, respectively. Those skilled in the art will appreciate that Figure 7 The computer device structure shown in the figure does not constitute a limitation on the computer device, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0217] The display unit 303 can be used to display information input by the user or information provided to the user and various graphical user interfaces of the computer device, which can be composed of graphics, text, icons, videos and any combination thereof. The display unit 303 can include a display panel and a touch panel.

[0218] The radio frequency circuit 304 may be used to send and receive radio frequency signals, so as to establish wireless communication with a network device or other computer devices through wireless communication, and to send and receive signals between the network device or other computer devices.

[0219] The audio circuit 305 can be used to provide an audio interface between the user and the computer device through a speaker and a microphone. The audio circuit 305 can transmit the electrical signal converted from the received audio data to the speaker, which is converted into a sound signal for output; on the other hand, the microphone converts the collected sound signal into an electrical signal, which is received by the audio circuit 305 and converted into audio data. The audio data is then processed by the output processor 301 and sent to, for example, another computer device through the radio frequency circuit 304, or the audio data is output to a memory for further processing. The audio circuit 305 may also include an earplug jack to provide communication between an external headset and a computer device.

[0220] The input unit 306 may be used to receive input numbers, character information or object feature information (such as fingerprint, iris, facial information, etc.), and generate keyboard, mouse, joystick, optical or trackball signal input related to user settings and function control.

[0221] The power supply 307 is used to supply power to various components of the computer device 300 .

[0222] although Figure 7 Not shown, the computer device 300 may also include a camera, a sensor, a wireless fidelity module, a Bluetooth module, etc., which will not be described in detail here.

[0223] In some embodiments, the present application further provides a computer-readable storage medium for storing a computer program. The computer-readable storage medium can be applied to a terminal device or a server, and the computer program enables the terminal device or the server to execute the corresponding process in the book recommendation method in the embodiment of the present application, which will not be described here for the sake of brevity.

[0224] In some embodiments, the present application further provides a computer program product, which includes a computer program stored in a computer-readable storage medium. The processor of the computer device reads the computer program from the computer-readable storage medium, and the processor executes the computer program, so that the computer device executes the corresponding process in the book recommendation method in the embodiment of the present application, which will not be described in detail for the sake of brevity.

[0225] The present application also provides a computer program, which includes a computer program, and the computer program is stored in a computer-readable storage medium. The processor of the computer device reads the computer program from the computer-readable storage medium, and the processor executes the computer program, so that the computer device executes the corresponding process in the book recommendation method in the embodiment of the present application, which will not be described in detail for the sake of brevity.

[0226] It should be understood that the processor of the embodiment of the present application may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method embodiment can be completed by the hardware integrated logic circuit or software instructions in the processor. The above processor can be a general processor, a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in the embodiment of the present application can be directly embodied as a hardware decoding processor to perform, or the hardware and software modules in the decoding processor are combined and performed. The software module can be located in a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, and other mature storage media in the art. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0227] It can be understood that the memory in the embodiments of the present application can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct RAM bus random access memory (DR RAM). It should be noted that the memory of the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0228] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0229] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0230] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components 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 interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0231] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0232] In addition, each functional unit in the embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0233] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. 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 and includes several instructions for a computer device (which can be a personal computer, a server) to perform all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage media include: various media that can store program codes, such as USB flash drives, mobile hard disks, ROM, RAM, magnetic disks or optical disks.

[0234] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A book recommendation method, characterized in that: The method comprises: Obtain target topics that match target books through various means; Generate a recommendation reason matching the target book according to the content of the related posts under the target topic, wherein the recommendation reason is used to establish an association between the target book and the target topic, and the related posts are posts under the target topic that are related to the target book; The target book and the reason for the recommendation are displayed.

2. The book recommendation method according to claim 1, characterized in that: The method further comprises: Obtain the relevance scores between the target book and each post under the target topic, sort them from high to low according to the relevance scores, and select the first M posts as the relevant posts, where M is a natural number greater than 1.

3. The book recommendation method according to claim 2, characterized in that: The step of obtaining the relevance scores between the target book and each post under the target topic includes: Perform keyword matching on each post under the target topic and the target book to obtain the keyword matching degree of each post; Performing semantic association analysis on each post under the target topic and the target book to obtain the semantic association between each post and the target book; Get the popularity of each post under the target topic; The relevance score between the target book and each post under the target topic is determined according to the keyword matching degree, the semantic relevance degree and the popularity.

4. The book recommendation method according to claim 1, characterized in that: The method further comprises: Read and parse the text content of each relevant post to extract key information in each relevant post; Classifying and arranging the key information to obtain processed related information; Based on the processed key information, construct the post content of each related post.

5. The book recommendation method according to claim 1, characterized in that: The target topic matching the target book is obtained in a variety of ways, including: Determining a target topic matching the target book from the multiple topics according to association data between the target book and the multiple topics; or According to the semantic relevance between the target book and each post under the multiple topics, a target topic matching the target book is determined from the multiple topics.

6. The book recommendation method according to claim 5, characterized in that: The step of determining a target topic matching the target book from the multiple topics according to the association data between the target book and the multiple topics includes: Determining, from the plurality of topics, a candidate topic that satisfies the preset matching condition according to the preset matching condition and the associated data; The candidate topics are sorted from high to low according to their click rates, and the first N candidate topics are selected as target topics matching the target book, where N is a natural number greater than 1.

7. The book recommendation method according to claim 6, characterized in that: The associated data at least includes the number of independent visitors to each topic, the number of replies with books to each topic, the number of page views of the target book in each topic, the reading conversion rate of the target book in each topic, and the irrelevance score of the target book to each topic. The preset matching conditions include: The number of independent visitors is greater than or equal to the visitor number threshold, and the number of replies with books is greater than or equal to the reply number threshold; The page views are greater than or equal to the page views threshold; The reading conversion rate is greater than or equal to the conversion rate threshold; The irrelevance score is less than or equal to a score threshold.

8. The book recommendation method according to claim 7, characterized in that: The associated data also includes the first number of comments, the first number of likes, and the first number of favorites of the target book in each topic. The preset matching conditions also include: The first number of comments is greater than or equal to a comment number threshold; or The first number of likes is greater than or equal to a likes threshold; or The first favorite number is greater than or equal to a favorite number threshold.

9. The book recommendation method according to claim 6, characterized in that: The method further comprises: Acquire user feedback data on the recommendation reason, the feedback data including at least one of the following: a click rate, a stay time, a third number of comments, a third number of likes, and a second number of shares of the user on the recommendation reason; The preset matching condition or the method for generating the recommendation reason is adjusted according to the feedback data.

10. The book recommendation method according to claim 5, characterized in that: The step of determining a target topic matching the target book from the multiple topics according to the semantic association between the target book and each post under the multiple topics includes: Calculating the semantic relevance between the target book and each post under multiple topics; Calculate the weighted average of the semantic relevances corresponding to all posts under each topic as the overall relevance between the target book and each topic; Topics whose overall relevance exceeds a relevance threshold are determined as target topics that match the target book.

11. The book recommendation method according to claim 1, characterized in that: The generating of recommendation reasons matching the target book according to the content of the relevant posts under the target topic includes: Performing content optimization processing on the post contents of the relevant posts under the target topic to obtain optimized post copy information, wherein the content optimization processing includes at least one of the following: semantic analysis, keyword extraction, keyword reorganization, deduplication processing, adjusting copy length, adjusting copy display style, and adjusting copy style; A compliance review is performed on the optimized post copy information, and the post copy information that passes the review is determined as a recommendation reason for matching the target book, wherein the compliance review includes at least one of the following: filtering sensitive words and verifying authenticity.

12. The book recommendation method according to claim 1, characterized in that: The generating of recommendation reasons matching the target book according to the content of the relevant posts under the target topic includes: According to the weights of the relevant posts under the target topic and the post contents of the relevant posts, a recommendation reason matching the target book is generated.

13. The book recommendation method according to claim 12, characterized in that: The method further comprises: The weight of the related post is adjusted according to the activity of the related post, where the activity includes the second number of comments, the second number of likes, and the first number of shares of the related post within a set time period.

14. The book recommendation method according to claim 12, characterized in that: The method further comprises: Analyzing multi-dimensional information of the target book, the multi-dimensional information including book type, author popularity, book rating, reader evaluation, and book sales; The weights of the related posts are adjusted according to the analysis results of the multi-dimensional information.

15. The book recommendation method according to any one of claims 1 to 14, characterized in that: The display of the target book and the reason for recommendation includes: Display the target book and the reason for recommendation on the recommendation page of the target reading platform; or The recommendation reason is displayed on the detail page of the target book.

16. The book recommendation method according to any one of claims 1 to 14, characterized in that: The method further comprises: generating a label of the target book, wherein the label is used to describe the attributes of the target book; When the recommendation reason is displayed, the label of the target book is also displayed.

17. The book recommendation method according to any one of claims 1 to 14, characterized in that: The method further comprises: generating an associated recommendation network of the target book, wherein the associated recommendation network is used to recommend books or authors associated with the target book; When displaying the recommendation reason, the associated recommendation network of the target book is also displayed.

18. A book recommendation device, characterized in that: The device comprises: An acquisition unit, used for acquiring a target topic matching a target book through a variety of methods; A generating unit, configured to generate a recommendation reason matching the target book according to the content of the related posts under the target topic, wherein the recommendation reason is used to establish an association between the target book and the target topic, and the related posts are posts under the target topic that are related to the target book; The display unit is used to display the target book and the reason for recommendation.

19. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is suitable for being loaded by a processor to execute the book recommendation method according to any one of claims 1 to 17.

20. A computer device, characterized in that: The computer device includes a processor and a memory, wherein a computer program is stored in the memory, and the processor is used to execute the book recommendation method according to any one of claims 1 to 17 by calling the computer program stored in the memory.

21. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the book recommendation method according to any one of claims 1 to 17 is implemented.

Images