Method and system for optimizing e-book recommendation algorithm based on tag system

Abstract

The invention relates to the technical field of information, in particular to an electronic book recommendation algorithm optimization method based on a tag system, which comprises the following steps of: S1, collecting reading data of a user from an electronic book application platform and preprocessing the reading data; s2, constructing a user portrait based on the preprocessed user reading data, and constructing a book recommendation index based on a tag system based on metadata of the electronic book; s3, calculating a matching index between the current user and the recommended book based on the user portrait and the book recommendation index; s4, calculating a comprehensive recommendation value of the electronic book based on the matching index in combination with the book recommendation index of the electronic book dimension; and S5, displaying a recommendation result to the current user based on the comprehensive recommendation value of the electronic book. According to the method, the user interests and the electronic book characteristics are deeply mined by constructing the tag system, so that more accurate and personalized recommendation is realized; the matching degree between the user and the electronic book and the popularity degree of the electronic book are comprehensively considered, and the accuracy and satisfaction degree of a recommendation result are improved.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and more specifically, to an optimization method and system for an e - book recommendation algorithm based on a tag system. Background Art

[0002] Currently, with the popularization of digital reading, reading e - books has become an important tool for people to obtain information, learn knowledge, and entertain in their spare time. However, in the face of a vast amount of e - book resources, how to efficiently and accurately recommend books that meet the interests and needs of users has become a major challenge for various e - book applications.

[0003] However, most of the existing e - book recommendation algorithms are based on simple data such as users' browsing history and purchase records, lacking a comprehensive exploration of users' deep - seated interests and a precise description of multi - dimensional features at the book level, resulting in limited accuracy and personalization of the recommendation results.

[0004] Therefore, in view of the problems existing in the prior art, how to design an optimization method and system for an e - book recommendation algorithm based on a tag system is an urgent problem to be solved by those skilled in the art. Summary of the Invention

[0005] In view of this, the present invention provides an optimization method and system for an e - book recommendation algorithm based on a tag system, which deeply explores users' interests and e - book features by constructing a tag system to achieve more accurate and personalized recommendations.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] An optimization method for an e - book recommendation algorithm based on a tag system, comprising the following steps:

[0008] S1. Collect user reading data from an e - book application platform and perform pre - processing;

[0009] S2. Construct a user portrait based on the pre - processed user reading data, and construct a book recommendation index based on the tag system based on the metadata of the e - books;

[0010] S3. Calculate the matching index between the current user and the recommended books based on the user portrait and the book recommendation index;

[0011] S4. Calculate the comprehensive recommendation value of the e - book based on the matching index and the book recommendation index in the e - book dimension;

[0012] S5. Display the recommendation results to the current user based on the comprehensive recommendation value of the e - book.

[0013] Further, in S1, the viewing data of the current user includes the user's behavior data on the e - book application platform and the e - book metadata corresponding to the behavior data;

[0014] Among them, the behavior data includes the user's search records, browsing records, reading duration, reading frequency, likes, comments, and shares; the e - book metadata includes the author, publication year, category, brief introduction, and rating;

[0015] In S1, the pre - processing process includes:

[0016] S11. Perform preliminary processing on the collected viewing data of the current user to remove noise and redundant information; the noise and redundant information include invalid clicks and repeated comments;

[0017] S12. Perform cleaning and normalization operations on the viewing data of the current user after removing noise and redundant information; the cleaning and normalization operations include converting unstructured text data into structured feature vectors for subsequent analysis and processing.

[0018] Further, in S2, the process of constructing a user portrait based on the pre - processed user viewing data includes:

[0019] S21. Statistically analyze the behavior data of the current user; among them, the statistical content includes the reading duration, number of likes, number of comments, and reading frequency of the user for various e - books;

[0020] S22. Cluster the viewing characteristics of the user based on the statistical and analysis results;

[0021] S23. Generate multiple labels for each user according to the clustering results to obtain a user portrait based on the label system, which is used to describe the user's reading interests and preferences.

[0022] Further, in S2, the process of constructing the e - book recommendation index based on the label system includes:

[0023] Parse the metadata of the e - book and extract key information content, and the extracted content includes the author, category, brief introduction, year, rating, and publisher of the e - book;

[0024] Construct corresponding labels for the e - book according to the extracted content; the labels include "science fiction", "novel", and "new book";

[0025] Calculate the popularity of the e - book according to the viewing data of the e - book and generate a book recommendation index based on the label system, which is used to reflect the popularity of the e - book in each label dimension;

[0026] The book recommendation index algorithm based on the tag system includes: the recommendation index calculation based on the tag X dimension and the recommendation index calculation based on the comprehensive dimension;

[0027] The formula for calculating the recommendation index in the label X dimension is:

[0028] Recommendation index of a specific e-book in the dimension of label X = ∑(N data of reading dimension - average value of N reading dimension of high-quality e-books under label X) / average value of N reading dimension of high-quality e-books under label X;

[0029] The calculation formula for the specific e-book recommendation index under the comprehensive dimension is:

[0030] Recommendation index = ∑(tag X recommendation index * e-book tag weight).

[0031] Furthermore, in S3, the matching index is the degree of similarity or overlap between the tags in the current user portrait and the tags in the e-book metadata;

[0032] The formula for calculating the matching index between the current user and a specific e-book is:

[0033] Matching index = ∑(user portrait tag ∩ e-book tag) * user portrait tag weight.

[0034] Furthermore, in S4, the comprehensive recommendation value of the e-book is calculated by combining the matching index and the book recommendation index. The calculation formula is:

[0035] Comprehensive recommendation value = matching index * weight 1 + recommendation index * weight 2.

[0036] Furthermore, in S5 , the comprehensive recommendation values of all the e-books that the current user may be interested in are calculated and sorted in descending order, and the e-book with the highest comprehensive recommendation value score is displayed to the user as a recommendation result.

[0037] The present invention discloses an optimization system for an e-book recommendation algorithm based on a tag system, comprising:

[0038] Data collection module: used to collect users' reading data on the e-book application platform and metadata of related e-books;

[0039] Data processing module: used to pre-process the collected user reading data and e-book metadata. The pre-processing process includes removing noise and redundant information, data cleaning and normalization operations;

[0040] User portrait building module: used to collect statistics and analyze user behavior data and generate a user portrait based on a tag system for each user;

[0041] Book recommendation index calculation module: used to calculate the e-book recommendation index based on the tag system;

[0042] Matching index calculation module: used to calculate the matching index between the user and multiple e-books;

[0043] Comprehensive recommendation value calculation module: used to calculate the comprehensive recommendation value of e-books according to the matching index and book recommendation index;

[0044] Recommendation result display module: used to sort the comprehensive recommendation values of multiple e-books, and display the e-book with the highest comprehensive recommendation score as the recommendation result to the user.

[0045] As can be seen from the above technical solutions, compared with the prior art, the present invention has the following beneficial effects:

[0046] (1) By constructing a tag system, the present invention deeply explores the user's interests and e-book features, and realizes more accurate and personalized recommendations.

[0047] (2) The present invention comprehensively considers the matching degree between the user and the e-book and the popularity of the e-book, and improves the accuracy and satisfaction of the recommendation results.

[0048] (3) The optimization method of the e-book recommendation algorithm based on the tag system of the present invention can be applied to the scenario of e-book recommendation on the e-book platform, providing users with a more convenient and efficient information acquisition experience. Description of the Drawings

[0049] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0050] Figure 1 Flowchart of the optimization method of the e-book recommendation algorithm based on the tag system provided by the present invention;

[0051] Figure 2 Structural schematic diagram of the optimization system of the e-book recommendation algorithm based on the tag system provided by the present invention. Detailed Embodiments

[0052] The following will combine the attached From Figure 1 to Figure 2 in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0053] Figure 1 Among them, an embodiment of the present invention discloses an optimization method for an e - book recommendation algorithm based on a tag system, including the following steps:

[0054] S1. Collect the reading data of users from the e - book application platform and perform pre - processing;

[0055] S2. Build a user portrait based on the pre - processed user reading data, and build a book recommendation index under the tag system based on the metadata of the e - books;

[0056] S3. Calculate the matching index between the current user and the recommended books based on the user portrait and the book recommendation index;

[0057] S4. Calculate the comprehensive recommendation value of the e - book based on the matching index combined with the book recommendation index in the e - book dimension;

[0058] S5. Display the recommendation results to the current user based on the comprehensive recommendation value of the e - book.

[0059] In this embodiment, in S1, the reading data of the current user includes the behavior data of the current user on the e - book application platform and the e - book metadata corresponding to the behavior data;

[0060] Among them, the behavior data includes the user's search records, browsing records, reading duration, reading frequency, likes, comments and shares; the e - book metadata includes the author, publication year, category, introduction and rating;

[0061] In S1, the pre - processing process includes:

[0062] S11. Perform preliminary processing on the collected reading data of the current user to remove noise and redundant information; the noise and redundant information include invalid clicks and repeated comments;

[0063] S12. Perform cleaning and normalization operations on the current user reading data after removing noise and redundant information; the cleaning and normalization operations include converting unstructured text data into structured feature vectors for subsequent analysis and processing.

[0064] In this embodiment, in S2, the process of building a user portrait based on the pre - processed user reading data includes:

[0065] S21. Statistically analyze the behavior data of the current user; among them, the statistical content includes the reading duration, number of likes, number of comments and reading frequency of the user for various types of e - books;

[0066] S22. Cluster the reading characteristics of the user based on the statistical and analysis results;

[0067] S23. Generate multiple tags for each user according to the clustering results, and obtain a user portrait based on the tag system to describe the reading interests and preferences of the users.

[0068] In this embodiment, in S2, the construction process of the e-book recommendation index based on the tag system includes:

[0069] Parse the metadata of the e-book and extract key information content, and the extracted content includes the author, category, introduction, year, rating, and publisher of the e-book;

[0070] Construct corresponding tags for the e-book according to the extracted content; the tags include "science fiction", "novel", and "new book";

[0071] Calculate its popularity according to the reading data of the e-book and generate a book recommendation index based on the tag system to reflect the popularity of the e-book in each tag dimension;

[0072] The book recommendation index algorithm based on the tag system includes: calculating the recommendation index in the X dimension of the tag and calculating the recommendation index in the comprehensive dimension;

[0073] The formula for calculating the recommendation index in the X dimension of the tag is:

[0074] The recommendation index of a specific e-book in the X dimension of the tag = ∑(reading dimension N data - average value of reading dimension N of high-quality e-books under tag X) / average value of reading dimension N of high-quality e-books under tag X;

[0075] The formula for calculating the recommendation index of a specific e-book in the comprehensive dimension is:

[0076] Recommendation index = ∑(recommendation index of tag X * e-book tag weight).

[0077] In this embodiment, in S3, the matching index is the degree of similarity or coincidence between the tags in the current user portrait and the tags in the e-book metadata; %]]

[0078] The formula for calculating the matching index between the current user and a specific e-book is:

[0079] Matching index = ∑(user portrait tag ∩ e-book tag) * user portrait tag weight.

[0080] In this embodiment, in S4, combine the matching index and the book recommendation index to calculate the comprehensive recommendation value of the e-book, and its formula is:

[0081] Comprehensive recommendation value = matching index * weight 1 + recommendation index * weight 2.

[0082] In this embodiment, in S5, calculate the comprehensive recommendation values of all e-books that the current user may be interested in, sort them in descending order, and display the e-book with the highest comprehensive recommendation score as the recommendation result to the user.

[0083] As Figure 2 shown, an optimization system for an e-book recommendation algorithm based on a tag system disclosed by the present invention includes:

[0084] A data collection module: used to collect the viewing data of users on the e-book application platform and the metadata of relevant e-books;

[0085] A data processing module: used to preprocess the collected user viewing data and e-book metadata. The preprocessing process includes removing noise and redundant information, data cleaning, and normalization operations;

[0086] A user profile construction module: used to statistically analyze the behavior data of users and generate a user profile based on the tag system for each user;

[0087] A book recommendation index calculation module: used to calculate the e-book recommendation index based on the tag system;

[0088] A matching index calculation module: used to calculate the matching index between the user and multiple e-books;

[0089] A comprehensive recommendation value calculation module: used to calculate the comprehensive recommendation value of the e-book according to the matching index and the book recommendation index;

[0090] A display recommendation result module: used to sort the comprehensive recommendation values of multiple e-books, and display the e-book with the highest comprehensive recommendation score as the recommendation result to the user.

[0091] In this embodiment, assume that the tags of an e-book include "science fiction" and "novel", and the weights of the two tags are q1 and q2 respectively, and q1 + q2 = 1;

[0092] It is known that the dimensions for measuring the viewing situation of e-books include the number of readings and the number of readers. Among them, the number of readings of the e-book is A, and the number of readers is B;

[0093] The average number of readings of high-quality e-books under the tag "science fiction" is K1, and the average number of readers is K2;

[0094] The average number of readings of high-quality e-books under the tag "novel" is S1, and the average number of readers is S2.

[0095] The recommendation indexes of the two tags of the e-book are respectively:

[0096] Tag "science fiction" recommendation index = (A - K1) / K1 + (B - K2) / K2,

[0097] The recommended index of the label "novel" = (A - S1) / S1 + (B - S2) / S2,

[0098] The recommended index of the e - book is as follows:

[0099] Recommended index = recommended index of the label "science fiction" * q1 + recommended index of the label "novel" * q2;

[0100] The calculated value is denoted as T, which is the recommended index of the e - book.

[0101] In the embodiment, assume that the portrait labels of the current user are "novel", "history" and "society", and the labels of a specific e - book include two, namely "science fiction" and "novel". The calculation formula of the matching index between the current user and the specific e - book is:

[0102] Matching index = weight of the "novel" label = q2.

[0103] Assume that it is desired to recommend books with both high matching degree and high quality to users. Then set both weight 1 and weight 2 to 0.5. Then the comprehensive recommendation value = q2 * 0.5+T * 0.5. In other embodiments, the values of weight 1 and weight 2 can be adaptively adjusted according to the recommendation purpose.

[0104] This embodiment realizes more accurate and personalized recommendation by constructing a label system and deeply mining user interests and e - book features.

[0105] This embodiment comprehensively considers the matching degree between the user and the e - book and the popularity of the e - book, improving the accuracy and satisfaction of the recommendation results.

[0106] The optimization method of the e - book recommendation algorithm based on the label system disclosed in this embodiment can be applied to the e - book recommendation scenario of the e - book platform, providing users with a more convenient and efficient information acquisition experience.

[0107] In this specification, each embodiment is described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. The same or similar parts among the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0108] The foregoing description of the disclosed embodiments enables those skilled in the art to practice or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An optimization method for an e - book recommendation algorithm based on a tag system, characterized in that, It includes the following steps: S1. Collect the reading data of users from the e-book application platform and perform preprocessing; S2. Build a user profile based on the preprocessed user reading data, and build a book recommendation index under the tag system based on the metadata of the e-books; S3. Calculate the matching index between the current user and the recommended books based on the user profile and the book recommendation index; S4. Calculate the comprehensive recommendation value of the e-books based on the matching index combined with the book recommendation index in the e-book dimension; S5. Display the recommendation results to the current user based on the comprehensive recommendation value of the e-books.

2. The optimization method of the e-book recommendation algorithm based on the tag system according to claim 1, characterized in that In S1, the reading data of the current user includes the behavior data of the current user on the e-book application platform and the e-book metadata corresponding to the behavior data; Among them, the behavior data includes the user's search records, browsing records, reading duration, reading frequency, likes, comments and shares; the e-book metadata includes the author, publication year, category, introduction and rating.

3. The optimization method of the e-book recommendation algorithm based on the tag system according to claim 1, characterized in that, In S1, the preprocessing process includes: S11. Perform preliminary processing on the collected reading data of the current user to remove noise and redundant information; S12. Perform cleaning and normalization operations on the reading data of the current user after removing noise and redundant information; Among them, the noise and redundant information include invalid clicks and repeated comments; the cleaning and normalization operations include converting unstructured text data into structured feature vectors for subsequent analysis and processing.

4. The optimization method of the e-book recommendation algorithm based on the tag system according to claim 1, characterized in that, In S2, the process of building a user profile based on the preprocessed user reading data includes: S21. Statistically analyze the behavior data of the current user; among them, the statistical content includes the reading duration, number of likes, number of comments and reading frequency of the current user for various types of e-books; S22. Cluster the reading characteristics of the user based on the statistical and analysis results; ​ 5. The optimization method of the e - book recommendation algorithm based on the tag system according to claim 1, characterized in that, ​ ​ ​ ​ 6. The optimization method of the e-book recommendation algorithm based on the tag system according to claim 5, characterized in that ​ ​ ​ ​ ​ 7. The optimization method of the e-book recommendation algorithm based on the tag system according to claim 1, wherein ​ The calculation formula for the matching index between the current user and a specific e-book is as follows: Matching index = ∑(user profile label ∩ e-book label) * user profile label weight.

8. The optimization method of the e-book recommendation algorithm based on the tag system according to claim 1, characterized in that, In S4, the comprehensive recommendation value of the e-book is calculated by combining the matching index and the book recommendation index, and its calculation formula is: Comprehensive recommendation value = matching index * weight 1 + recommendation index * weight 2.

9. The optimization method of the e-book recommendation algorithm based on the tag system according to claim 1, characterized in that In S5, the comprehensive recommendation values of all e-books that the current user may be interested in are calculated and sorted in descending order, and the e-book with the highest comprehensive recommendation score is presented to the user as the recommendation result.

10. An optimization system for an e-book recommendation algorithm based on a tag system, characterized in that, It is applicable to the optimization method of the e-book recommendation algorithm based on the tag system as described in any one of claims 1-9, including: Data collection module: used to collect the reading data of users on the e-book application platform and the metadata of relevant e-books; Data processing module: used to preprocess the collected user reading data and e-book metadata, and the preprocessing process includes removing noise and redundant information, data cleaning and normalization operations; User profile construction module: used to statistically analyze the user's behavior data and generate a user profile based on the tag system for each user; Book recommendation index calculation module: used to calculate the e-book recommendation index based on the tag system; Matching index calculation module: used to calculate the matching index between the user and multiple e-books; Comprehensive recommendation value calculation module: used to calculate the comprehensive recommendation value of the e-book according to the matching index and the book recommendation index; Recommendation result display module: used to sort the comprehensive recommendation values of multiple e-books, and present the e-book with the highest comprehensive recommendation score to the user as the recommendation result.

Images