Library book automatic searching method and system
By automatically extracting book summary information and calculating label correlation, and assigning tags to books, the problem of difficulty in finding interdisciplinary books in the existing technology is solved, and more accurate and efficient book search is achieved.
Patent Information
- Application Number
- CN202510110995.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-23
- Publication Date
- 2025-06-03
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing book search system relies on preset classification systems and keyword matching, making it difficult to effectively find books with interdisciplinary or complex content, resulting in unsatisfactory search results.
By obtaining the summary information of the book, extracting features and calculating the correlation degree with the preset label, and automatically assigning tags to the book; at the same time, extracting features and calculating the correlation degree with the label according to user needs, and setting search conditions to find relevant books from the library.
It realizes accurate annotation and search of complex books in multiple fields, improves the relevance and accuracy of search results, and helps librarians better manage and maintain collection resources.
Smart Images
Figure CN120086389A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of library management, and more specifically, to a method and system for automatically searching for library books. Background Art
[0002] Existing book search systems mainly rely on a preset classification system (such as the Dewey Decimal Classification or the International Standard Book Number ISBN) and keyword matching technology to achieve book retrieval. However, this search method has certain limitations: on the one hand, it requires users to have a certain understanding of the library's classification rules and needs to accurately input keywords to obtain satisfactory results; on the other hand, for some interdisciplinary or books with more complex content and involving multiple fields, their classification and keywords may not comprehensively and accurately reflect the actual content of the books, resulting in less than ideal search results.
[0003] In order to improve the user experience and optimize the resource utilization efficiency, it is crucial to develop a method that can automatically and intelligently help users find the books they need. Summary of the Invention
[0004] To solve the above technical problems, the present application is proposed to provide a method and system for automatically searching for library books, aiming to automatically and intelligently help users find the books they need.
[0005] In a first aspect, the present invention provides a method for automatically searching for library books, including: taking any book in the library as a target book, obtaining the abstract information of the target book; extracting a first feature from the abstract information of the target book; calculating the correlation degree between the first feature and a plurality of preset tags; using the tags with a correlation degree higher than a preset threshold with the first feature as the tags of the target book; obtaining the user's need to search for a book; extracting a second feature from the user's need to search for a book; calculating the correlation degree between the second feature and the plurality of tags; setting search conditions according to the tags with a correlation degree higher than the threshold of the second feature, and searching for books in the library that have all the tags in the search conditions and pushing them to the user.
[0006] Optionally, in the foregoing method for automatically searching for library books, obtaining the abstract information of the target book includes: when the abstract information is not recorded in the target book, extracting sentences from the first paragraph of each chapter of the target book; evaluating the importance of the extracted multiple sentences according to a preset rule; combining the sentences with an importance higher than a preset level as the abstract information of the target book.
[0007] Optionally, in the aforementioned automatic library book search method, when evaluating the importance of multiple extracted sentences according to a preset rule, it includes: taking any one of the multiple sentences as a target sentence, detecting the position p(s) of the target sentence in the target book, where s represents the target sentence; calculating the vector v of the target sentence s and the vectors v of the other sentences in the multiple sentences s' , where s' represents any other sentence in the multiple sentences; calculating the importance of the target sentence where Z is the set of the multiple sentences, num(Z) is the number of the multiple sentences, p mid is the middle position of the book, the exp function is the exponential function with the real number e as the base, and || represents taking the absolute value.
[0008] Optionally, in the aforementioned automatic library book search method, when calculating the correlation degree between the first feature and a preset multiple labels, it includes: taking any one of the multiple labels as a candidate label, calculating the vector v of the first feature D and the vector v of the candidate label t , where t represents the candidate label and D represents the first feature; calculating the inverse document frequency S TF-IDF (t, D) of the first feature and the candidate label; calculating the correlation degree between the first feature and the candidate label where α is a weight factor, the exp function is the exponential function with the real number e as the base, and |||| represents the norm; calculating the correlation degree between the first feature and the candidate label according to the correlation degree R(t, D) between the first feature and the candidate label.
[0009] Optionally, in the aforementioned automatic library book search method, before calculating the correlation degree between the first feature and the candidate label according to the correlation degree R(t, D) between the first feature and the candidate label, the automatic library book search method further includes: counting the number of occurrences count(t, D) of the candidate label in the first feature; counting the maximum value max(count(w, D)) of the number of occurrences of all words in the first feature, where w represents any word in the first feature and the max function is used to take the maximum value; calculating the semantic approximation degree S BERT (t, D) between the first feature and the candidate label based on a preset BERT model; calculating the local uniqueness degree of the first feature relative to the candidate label Calculate the association degree between the first feature and the candidate label according to the degree of relevance R(t, D) between the first feature and the candidate label, including: according to the degree of relevance R(t, D) between the first feature and the candidate label and the degree of local distinctiveness U 1 (t, D) of the first feature relative to the candidate label, calculate the association degree between the first feature and the candidate label.
[0010] Optionally, in the above-mentioned automatic library book search method, before calculating the association degree between the first feature and the candidate label according to the degree of relevance R(t, D) between the first feature and the candidate label, the automatic library book search method further includes: counting the number of books num total (t) with the candidate label among all the books in the library; counting the number of books num dom (t) with the candidate label among the books in a specific field of the library, where dom represents any field in the specific field; calculating the degree of global distinctiveness of the first feature relative to the candidate label, where N is the number of all books in the library, and DOM represents all fields in the specific field; calculating the association degree between the first feature and the candidate label according to the degree of relevance R(t, D) between the first feature and the candidate label, including: according to the degree of relevance R(t, D) between the first feature and the candidate label and the degree of global distinctiveness U 2 (t, D) of the first feature relative to the candidate label, calculate the association degree between the first feature and the candidate label.
[0011] Optionally, in the above-mentioned automatic library book search method, before extracting the second feature from the user's book search requirement, the automatic library book search method further includes: obtaining the user's identity information; adding the user's identity information to the user's book search requirement.
[0012] Optionally, in the above-mentioned automatic library book search method, searching for books with all the labels in the search condition from all the books in the library and pushing them to the user includes: counting the number of labels of each book found; pushing the books found to the user in ascending order of the number of labels.
[0013] Second aspect, the present invention provides an automatic library book search system, including: an abstract acquisition module, which takes any book in the library as the target book and acquires the abstract information of the target book; a first feature extraction module, which extracts the first feature from the abstract information of the target book; a first association calculation module, which calculates the association degree between the first feature and a plurality of preset tags; a tag setting module, which takes the tags with an association degree higher than the preset threshold with the first feature as the tags of the target book; a requirement acquisition module, which acquires the requirement of the user to search for a book; a second feature extraction module, which extracts the second feature from the requirement of the user to search for a book; a second association calculation module, which calculates the association degree between the second feature and the plurality of tags; a search module, which sets search conditions according to the tags with an association degree higher than the threshold of the second feature, searches for books having all the tags in the search conditions from all the books in the library, and pushes them to the user.
[0014] One or more of the above technical solutions of the present invention have at least one or more of the following beneficial effects:
[0015] According to the technical solution of the present invention, for complex books involving multiple fields, multiple relevant tags can be assigned to them, making such books easier to be found by readers with different professional backgrounds. By intelligently annotating and classifying all the books in the library, it can help librarians better manage and maintain the collection resources, and at the same time facilitate the statistical analysis of hot topics and trends, providing data support for procurement decisions. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other objects, features and advantages of the present application will become more obvious. The accompanying drawings are used to provide a further understanding of the embodiments of the present application, and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation to the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.
[0017] Figure 1 It is a flowchart of an automatic library book search method according to an embodiment of the present application;
[0018] Figure 2 It is a partial flowchart of an automatic library book search method according to an embodiment of the present application;
[0019] Figure 3 It is another partial flowchart of an automatic library book search method according to an embodiment of the present application;
[0020] Figure 4 It is yet another partial flowchart of an automatic library book search method according to an embodiment of the present application;
[0021] Figure 5 Another partial flowchart of the automatic library book search method according to an embodiment of the present application;
[0022] Figure 6 It is a block diagram of an automatic library book search system according to an embodiment of the present application. Detailed implementation manners
[0023] Some implementation manners of the present invention will be described below with reference to the accompanying drawings. Those skilled in the art should understand that these implementation manners are only used to explain the technical principle of the present invention and are not intended to limit the protection scope of the present invention.
[0024] As Figure 1 shown, in an embodiment of the present invention, an automatic library book search method is provided, including:
[0025] Step S110: Take any book in the library as the target book and obtain the abstract information of the target book.
[0026] Step S120: Extract the first feature from the abstract information of the target book.
[0027] Step S130: Calculate the correlation degree between the first feature and a plurality of preset tags.
[0028] Step S140: Use the tags with a correlation degree higher than the preset threshold with the first feature as the tags of the target book.
[0029] In this embodiment, by extracting the first feature from the abstract information of the target book and calculating the correlation degree between these features and the preset tags, it is ensured that each book can be accurately labeled. This content-based tag assignment method can more accurately reflect the actual theme and content of the book, thereby improving the relevance and accuracy of the search results.
[0030] Step S150: Obtain the user's requirement for searching for a book.
[0031] Step S160: Extract the second feature from the user's requirement for searching for a book.
[0032] Step S170: Calculate the correlation degree between the second feature and a plurality of tags.
[0033] Step S180: Set search conditions according to the tags with a correlation degree higher than the threshold of the second feature, search for books with all the tags in the search conditions from all the books in the library, and push them to the user.
[0034] Traditional search methods usually rely on fixed classification systems or keyword matching, which may affect the search efficiency due to inaccurate user input. In this embodiment, by intelligently analyzing the user's needs and setting reasonable search conditions, invalid searches are reduced, the search speed is accelerated, and users can find the required book materials faster.
[0035] According to the technical solution of this embodiment, for complex books involving multiple fields, multiple relevant tags can be assigned to them, making such books easier to be found by readers with different professional backgrounds. By intelligently annotating and classifying all the books in the library, it can help librarians better manage and maintain the collection resources. At the same time, it is also convenient for statistical analysis of hot topics and trends, providing data support for procurement decisions.
[0036] As Figure 2 shown, in an embodiment of the present invention, an automatic search method for library books is provided. Compared with the foregoing embodiments, the automatic search method for library books in this embodiment, step S110 includes:
[0037] Step S210, when the abstract information is not recorded in the target book, extract sentences from the first paragraph of each chapter of the target book.
[0038] In this embodiment, when the target book lacks pre-prepared abstract information, it can automatically extract sentences from the first paragraph of each chapter, evaluate the importance of these sentences according to preset rules, and finally combine the sentences with importance higher than the preset level as the abstract information of the book. This process requires no manual intervention, greatly improving the processing efficiency.
[0039] Step S220, evaluate the importance of the extracted multiple sentences according to preset rules.
[0040] In this embodiment, the ways to evaluate the importance of sentences include but are not limited to the following:
[0041] (1) Take any sentence among the multiple sentences as the target sentence, and detect the position p(s) of the target sentence in the target book, where s represents the target sentence.
[0042] (2) Calculate the vector v of the target sentence s and the vectors v of the other sentences among the multiple sentences s' , where s' represents any other sentence among the multiple sentences.
[0043] (3) Calculate the importance of the target sentence where, Z is the set of multiple sentences, num(Z) is the number of multiple sentences, p mid is the middle position of the book, the exp function is the exponential function with the real number e as the base, and || represents taking the absolute value.
[0044] In this embodiment, by evaluating the importance of multiple extracted sentences, it is ensured that the generated abstract is an effective summary of the book content. In particular, by considering the position of the sentence, its similarity to other sentences, and the distance from the middle position of the book to calculate its importance, this method can more accurately capture the core ideas and key points of the book.
[0045] Step S230: Combine the sentences with importance higher than the preset level as the abstract information of the target book.
[0046] According to the technical solution of this embodiment, not only the problem that some books lack abstract information is solved, but also the quality and relevance of the abstract are optimized through intelligent algorithms, greatly improving the performance and service ability of the library book automatic search system.
[0047] As Figure 3 shown, in an embodiment of the present invention, a method for automatically searching for library books is provided. Compared with the foregoing embodiments, in the method for automatically searching for library books in this embodiment, step S130 includes:
[0048] Step S310: Take any one of the multiple tags as a candidate tag, and calculate the vector v of the first feature D and the vector v of the candidate tag t , where t represents the candidate tag and D represents the first feature.
[0049] Step S320: Calculate the inverse document frequency S of the first feature and the candidate tag TF-IDF (t, D).
[0050] In this embodiment, by calculating the difference between the first feature vector and the candidate tag vector and combining the inverse document frequency to measure the degree of correlation, this method can more accurately evaluate the association degree between the book content and the tag. This not only improves the accuracy of tag assignment but also ensures that each tag can effectively represent the core theme of the book.
[0051] Step S330: Calculate the degree of correlation between the first feature and the candidate tag where α is a weight factor, the exp function is an exponential function with the real number e as the base, and |||| represents the norm.
[0052] In this embodiment, the above formula is used to calculate the degree of correlation, which takes into account the Euclidean distance between the feature vectors and the importance of the tags. This method can filter out irrelevant tags, making the finally set search conditions more in line with the actual needs of users, thereby improving the relevance and accuracy of the search results.
[0053] Step S340: Calculate the association degree between the first feature and the candidate label according to the correlation degree R(t, D) between the first feature and the candidate label.
[0054] According to the technical solution of this embodiment, by comparing the vector representation of the label with the first feature of the book content, the semantic connection between the two can be better captured, rather than just based on surface text matching. This deep understanding helps to identify content that is semantically similar although the literal expressions are different, providing more comprehensive and accurate recommendations for users.
[0055] As Figure 4 shown, in an embodiment of the present invention, a method for automatically searching for library books is provided. Compared with the foregoing embodiments, before step S340 of the method for automatically searching for library books in this embodiment, the method for automatically searching for library books further includes:
[0056] Step S410: Count the number of occurrences count(t, D) of the candidate label in the first feature.
[0057] Step S420: Count the maximum value max(count(w, D)) of the number of occurrences of all words in the first feature, where w represents any word in the first feature, and the max function is used to take the maximum value.
[0058] In this embodiment, by counting the number of occurrences of the candidate label in the first feature and the maximum number of occurrences of all words in the first feature, the representativeness of each label in the book content can be more finely quantified. This helps to more accurately identify those labels that have a key impact on the book theme.
[0059] Step S430: Calculate the semantic similarity degree S BERT (t, D) between the first feature and the candidate label based on a preset BERT model.
[0060] In this embodiment, the semantic similarity degree between the first feature and the candidate label is calculated based on a preset BERT model. This method goes beyond traditional lexical matching methods and delves into the semantic level at the sentence level for comparison. In this way, not only can the surface text similarity be captured, but also the implicit relationships in the context can be understood, thus providing more accurate label assignment.
[0061] Step S440: Calculate the local uniqueness degree of the first feature relative to the candidate label
[0062] In this embodiment, by calculating the degree of local distinctiveness, the distinctiveness and importance of a tag relative to other words can be measured. This evaluation method takes into account the relative scarcity and particularity of the tag in the context of the entire document, so that tags that are semantically crucial but have a low frequency of occurrence will not be overlooked, improving the quality of tag assignment.
[0063] In this embodiment, step S340 may specifically include: calculating the association degree between the first feature and the candidate tag according to the correlation degree R(t, D) between the first feature and the candidate tag and the degree of local distinctiveness U 1 (t, D) of the first feature relative to the candidate tag.
[0064] According to the technical solution of this embodiment, by introducing advanced semantic analysis technology and unique tag evaluation indicators, the intelligent level and service quality of the library book automatic search system are significantly improved, providing users with a more efficient, accurate and personalized search experience.
[0065] As Figure 5 shown, in an embodiment of the present invention, a method for automatically searching for library books is provided. Compared with the foregoing embodiments, in the method for automatically searching for library books in this embodiment, before step S340, the method for automatically searching for library books further includes:
[0066] Step S510, counting the number num total (t) of books with candidate tags among all the books in the library.
[0067] Step S520, counting the number num dom (t) of books with candidate tags among the books in a specific field of the library, where dom represents any field in the specific field.
[0068] In this embodiment, by counting the number of books with candidate tags among all the books in the library and the number of books with candidate tags among the books in a specific field, the importance of each tag can be evaluated within a wider range. This not only considers the content characteristics of individual books but also takes into account the distribution of the entire collection of resources, thus providing a more comprehensive and objective basis for tag assignment.
[0069] Step S530, calculating the degree of global distinctiveness of the first feature relative to the candidate tag where N is the number of all books in the library, and DOM represents all fields in the specific field.
[0070] In this embodiment, when calculating the global uniqueness degree, considering the different usage frequencies of the same label in books in different fields, it is possible to identify those labels that are particularly important in certain fields but relatively common in other fields. This mechanism helps to cross traditional disciplinary boundaries, provides more accurate and diverse label assignments for complex books involving multiple fields, and promotes the effective discovery of interdisciplinary resources.
[0071] In this embodiment, step S340 may specifically include: calculating the association degree between the first feature and the candidate label according to the correlation degree R(t, D) between the first feature and the candidate label and the global uniqueness degree U 2 (t, D) of the first feature with respect to the candidate label.
[0072] According to the technical solution of this embodiment, by introducing the consideration of the global uniqueness degree, the intelligent level of label assignment and book search is significantly improved, which not only enhances the user experience, but also brings a qualitative leap to library management and knowledge dissemination. This method ensures that the labels can not only accurately describe the book content, but also reflect its value in a larger range, providing users with richer and more meaningful information resources.
[0073] In one embodiment of the present invention, a method for automatically searching library books is provided. Compared with the foregoing embodiments, in the method for automatically searching library books of this embodiment, before step S160, the method for automatically searching library books further includes:
[0074] Obtain the user's identity information.
[0075] Add the user's identity information to the user's requirement for searching books.
[0076] In this embodiment, searching in combination with the user's identity information can filter out books that are not relevant to the user, ensuring that the pushed books not only meet the user's search conditions, but also match their personal characteristics. This precise positioning greatly improves the relevance and practicality of the search results.
[0077] Step S180 includes:
[0078] Count the number of labels of each book obtained by the search.
[0079] Push the books obtained by the search to the user in ascending order of the number of labels.
[0080] According to the technical solution of this embodiment, pushing the found books sorted by the number of tags from low to high helps reduce the problem of information overload when users face too many choices. Books with fewer tags usually mean that their themes are more concentrated, which is easier for users who are new to the field to understand and digest; while books with more tags may cover more extensive content or involve multiple related fields, and are suitable for advanced learning or research.
[0081] As Figure 6 shown, in an embodiment of the present invention, an automatic library book search system is provided, including:
[0082] An abstract acquisition module 610, taking any book in the library as the target book and acquiring the abstract information of the target book.
[0083] A first feature extraction module 620 extracts the first feature from the abstract information of the target book.
[0084] A first correlation calculation module 630 calculates the correlation degree between the first feature and a plurality of preset tags.
[0085] A tag setting module 640 uses the tags whose correlation degree with the first feature is higher than the preset threshold as the tags of the target book.
[0086] In this embodiment, by extracting the first feature from the abstract information of the target book and calculating the correlation degree between these features and the preset tags, it is ensured that each book can be accurately marked. This content-based tag assignment method can more accurately reflect the actual theme and content of the book, thereby improving the relevance and accuracy of the search results.
[0087] A demand acquisition module 650 acquires the user's demand for searching for books.
[0088] A second feature extraction module 660 extracts the second feature from the user's demand for searching for books.
[0089] A second correlation calculation module 670 calculates the correlation degree between the second feature and a plurality of tags.
[0090] A search module 680 sets search conditions according to the tags whose correlation degree of the second feature is higher than the threshold, searches for books in the library that have all the tags in the search conditions from all the books in the library, and pushes them to the user.
[0091] Traditional search methods usually rely on fixed classification systems or keyword matching, which may affect the search efficiency due to inaccurate user input. In this embodiment, by intelligently analyzing the user's needs and setting reasonable search conditions, invalid searches are reduced, the search speed is accelerated, and users can find the required book materials faster.
[0092] According to the technical solution of this embodiment, for complex books involving multiple fields, multiple relevant tags can be assigned to them, making such books easier to be found by readers with different professional backgrounds. By intelligently annotating and classifying all the books in the library, it can help librarians better manage and maintain the collection resources. At the same time, it is also convenient for statistical analysis of hot topics and trends, providing data support for procurement decisions.
[0093] The basic principles of this application have been described above in conjunction with specific embodiments. However, it should be noted that the advantages, benefits, effects, etc. mentioned in this application are only examples and not limitations. It cannot be considered that these advantages, benefits, effects, etc. are essential for each embodiment of this application. In addition, the above-disclosed specific details are only for the purpose of illustration and easy understanding, rather than limitations. These details do not limit this application to necessarily adopt the above specific details to implement.
[0094] The block diagrams of the devices, apparatuses, equipment, and systems involved in this application are only illustrative examples and do not intend to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, equipment, and systems can be connected, arranged, and configured in any way. Words such as "including", "comprising", "having", etc. are open-ended terms, meaning "including but not limited to", and can be used interchangeably with each other. The word "or" and "and" used here refer to the word "and / or", and can be used interchangeably with each other, unless the context clearly indicates otherwise. The word "such as" used here refers to the phrase "such as but not limited to", and can be used interchangeably with each other.
[0095] It should also be noted that in the devices, equipment, and methods of this application, each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent solutions of this application.
[0096] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this application. Various modifications to these aspects are very obvious to those skilled in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of this application. Therefore, this application is not intended to be limited to the aspects shown herein, but rather to the broadest scope consistent with the principles and novel features disclosed herein.
[0097] The above description has been given for purposes of illustration and description. In addition, this description does not intend to limit the embodiments of this application to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions, and sub-combinations thereof.
Claims
1. Automatic search method for library books, including: Taking any book in the library as a target book, obtaining summary information of the target book; extracting a first feature from the summary information of the target book; Calculating the correlation between the first feature and a plurality of preset tags; Using a label whose correlation with the first feature is higher than a preset threshold as a label of the target book; Obtain users' demand for searching books; Extracting a second feature from the user's demand for searching for books; Calculating the degree of association between the second feature and the multiple tags; Search conditions are set according to the tags whose relevance of the second feature is higher than the threshold, and books with all the tags in the search conditions are searched from all the books in the library, and pushed to the user.
2. The automatic library book search method according to claim 1, wherein: Get summary information of the target book, including: When the summary information is not recorded in the target book, extracting sentences from the first paragraph of each chapter of the target book; Evaluate the importance of multiple extracted sentences according to preset rules; The sentences whose importance is higher than a preset level are combined as summary information of the target book.
3. The automatic library book search method according to claim 2, wherein: The importance of multiple extracted sentences is evaluated according to preset rules, including: Taking any sentence among the plurality of sentences as a target sentence, detecting a position p(s) of the target sentence in the target book, where s represents the target sentence; Calculate the vector v of the target sentence s And the vectors v of other sentences in the plurality of sentences s' , where s' represents any other sentence in the plurality of sentences; Calculate the importance of the target sentence Where Z is the set of the plurality of sentences, num(Z) is the number of the plurality of sentences, and p mid is the middle position of the book, exp function is an exponential function with the real number e as the base, and || represents taking the absolute value.
4. The automatic library book search method according to claim 1, wherein: Calculating the correlation between the first feature and a plurality of preset tags includes: Take any label among the multiple labels as a candidate label and calculate the vector v of the first feature D and the vector v of the candidate label t , where t represents the candidate tag, and D represents the first feature; Calculate the inverse document frequency S of the first feature and the candidate tag TF-IDF (t,D); Calculate the correlation between the first feature and the candidate label Among them, α is the weight factor, exp function is the exponential function with real number e as the base, and |||| represents the norm; According to the correlation degree R(t,D) between the first feature and the candidate tag, the association degree between the first feature and the candidate tag is calculated.
5. The automatic library book search method according to claim 4, wherein: Before calculating the association degree between the first feature and the candidate tag according to the correlation degree R(t,D) between the first feature and the candidate tag, the automatic library book search method further includes: Count the number of occurrences of the candidate tag in the first feature count(t,D); Count the maximum value of the number of occurrences of all words in the first feature, max(count(w,D)), where w represents any word in the first feature, and the max function is used to obtain the maximum value; The semantic similarity S between the first feature and the candidate tag is calculated based on the preset BERT model. BERT (t,D); Calculate the local uniqueness of the first feature relative to the candidate label According to the correlation degree R(t,D) between the first feature and the candidate tag, calculating the association degree between the first feature and the candidate tag includes: The association degree between the first feature and the candidate tag is calculated according to the correlation degree R(t,D) between the first feature and the candidate tag and the local uniqueness degree U1(t,D) of the first feature relative to the candidate tag.
6. The automatic library book search method according to claim 4, wherein: Before calculating the association degree between the first feature and the candidate tag according to the correlation degree R(t,D) between the first feature and the candidate tag, the automatic library book search method further includes: Count the number of books with the candidate tag among all the books in the library total (t); Count the number of books in the specific field of the library with the candidate label num dom (t), wherein dom represents any one of the specific domains; Calculate the global uniqueness of the first feature relative to the candidate label Where N is the number of all books in the library, and DOM represents all fields in the specific field; According to the correlation degree R(t,D) between the first feature and the candidate tag, calculating the association degree between the first feature and the candidate tag includes: The association degree between the first feature and the candidate tag is calculated according to the correlation degree R(t,D) between the first feature and the candidate tag and the global uniqueness degree U2(t,D) of the first feature relative to the candidate tag.
7. The automatic library book search method according to claim 1, wherein: Before extracting the second feature from the user's demand for searching for books, the automatic library book search method further includes: Obtaining identity information of the user; The user's identity information is added to the user's need to search for books.
8. The automatic library book search method according to claim 1, wherein: Searching for books with all tags in the search condition from all books in the library and pushing them to the user includes: Count the number of tags for each book found; The found books are pushed to the user in the order of the number of tags from low to high.
9. Library book automatic search system, including: The abstract acquisition module takes any book in the library as a target book and acquires the abstract information of the target book; A first feature extraction module extracts a first feature from the abstract information of the target book; A first association calculation module, calculating the association degree between the first feature and a plurality of preset tags; A tag setting module, which uses a tag whose correlation with the first feature is higher than a preset threshold as a tag of the target book; Demand acquisition module, which obtains the user's demand for searching for books; A second feature extraction module extracts a second feature from the user's demand for searching for books; A second association calculation module, calculating the association degree between the second feature and the multiple tags; The search module sets search conditions according to the tags whose relevance of the second feature is higher than the threshold, searches for books with all the tags in the search conditions from all the books in the library, and pushes the books to the user.