Intelligent library consultation method and system

Abstract

The invention belongs to the technical field of book informatization management, and particularly relates to an intelligent library consultation method and system, and the method comprises the steps: S1, determining corresponding core keywords and extension keywords according to the content of each book, and forming a core keyword set and an extension keyword set corresponding to the book; s2, determining a retrieval keyword according to a user demand; and S3, outputting a retrieval result by comparing the retrieval keywords, the core key and the expanded keywords. The intelligent library consultation method and system adopt a mode of firstly determining the core keyword set and the expanded keyword set corresponding to the book and retrieving the keywords, so that the retrieval process is more intelligent and simpler, and the retrieval efficiency is improved. And the retrieval result is more accurate.

Description

Technical Field

[0001] The invention belongs to the technical field of book information management, and in particular relates to an intelligent library consultation method and system. Background Art

[0002] With the progress of society, all walks of life are beginning to develop in the direction of intelligence. More and more libraries are also beginning to adopt information-based and intelligent management and operation methods, which not only improves the efficiency of book management, but also facilitates book classification, retrieval, borrowing and other tasks. In particular, the introduction of intelligent management systems not only makes book borrowing and returning efficient and convenient, but also gives libraries powerful query and retrieval functions, allowing readers to easily travel through the ocean of knowledge.

[0003] However, with the increasing abundance of book resources and the surge in the number of books, a new challenge has quietly emerged: with so many types of books, how to choose the right book has become a problem for users. Especially for university libraries, there are so many professional books that readers find it difficult to choose, which has become a problem faced by many readers, especially university teachers and students.

[0004] Currently, most library consultation systems have query and retrieval functions, through which users can learn about the library's collection and borrow books. This traditional method of using users' independent retrieval to find books that meet borrowing needs is highly dependent on the user's search fields and retrieval technology. This one-way "people looking for books" retrieval method cannot accurately recommend books that users actually need, resulting in low efficiency and insufficient accuracy in book retrieval.

[0005] Especially in recent years, with the development of economy, people's emphasis on education and the popularization of books, many primary and secondary schools have also established libraries with huge collections of books. For primary and secondary school students, they lack basic search skills and cannot conduct effective searches. In fact, for most college students, without professional search knowledge and a lot of search practice training, they cannot conduct efficient searches.

[0006] Faced with a wide variety of books, these users who lack search experience and skills find it difficult to quickly obtain the target books. How to reduce the difficulty of book selection and improve the efficiency of book selection has become the key to upgrading library services.

[0007] Faced with this dilemma, we can't help but think: How can we quickly find the right book for ourselves among the vast number of books? The answer may lie in the intelligent recommendation function. Intelligent recommendation of books means taking readers as the center, actively recommending books to users according to readers' borrowing needs, and providing readers with more accurate information and knowledge services. At present, although the library management system is powerful, it seems to be unable to do so in terms of intelligent recommendation of books. This is undoubtedly a key breakthrough in upgrading library services.

[0008] In view of this, this application is hereby filed. Summary of the invention

[0009] The purpose of the present invention is to address the above technical problems and to provide an intelligent library consultation method and system that can actively and accurately recommend books to users based on their borrowing needs, thereby providing readers with more accurate information and knowledge services.

[0010] In view of this, the present invention provides an intelligent library consultation method, comprising the steps of:

[0011] S1, establishing the core keywords and extended keywords corresponding to each book according to its content, and forming a core keyword set and an extended keyword set corresponding to the book;

[0012] S2, establish search keywords based on user needs;

[0013] S3 outputs search results by comparing search keywords, core keywords and expanded keywords.

[0014] Furthermore, the step S1 includes:

[0015] S11, obtaining the electronic version file of the book;

[0016] S12, extracting keywords from the book title and using them as first-level keywords;

[0017] S13, extract keywords from the abstract, preface, and table of contents and use them as secondary keywords;

[0018] S14, extracting keywords from the text and using them as third-level keywords;

[0019] S15, determining core keywords based on primary, secondary and tertiary keywords;

[0020] S16, based on the frequency F of each core keyword in the text m and level factor L m Calculate the depth index α of each core keyword m ;

[0021] S17, according to the depth index α mSort the core keywords in the core keyword set by the size of to obtain a sorted core keyword set;

[0022] S18, performing synonym expansion on each core keyword to obtain an expanded keyword set of each core keyword.

[0023] Furthermore, the step S15 includes:

[0024] S151, obtaining first, second and third level keywords, and storing them in a hierarchical keyword set, forming a hierarchical keyword set including keywords at each level;

[0025] S152, obtaining the nth keyword in the hierarchical keyword set, where the initial value of n is 1;

[0026] S153, the nth keyword obtained in step S152 is compared with the elements in the primary keyword set consisting of primary keywords, the secondary keyword set consisting of secondary keywords, and the tertiary keyword set consisting of tertiary keywords respectively:

[0027] If the primary keyword set, the secondary keyword set, and the tertiary keyword set all include the current keyword, the current keyword is determined to be a core keyword, and the current keyword is stored in the core keyword set, and then step S156 is continued;

[0028] If two of the first-level keyword set, the second-level keyword set, and the third-level keyword set contain the current keyword, then step S154 needs to be continued;

[0029] If only one of the primary keyword set, the secondary keyword set, and the tertiary keyword set contains the current keyword, then step S155 needs to be continued;

[0030] S154, obtaining the frequency P of the current keyword appearing in the book text n , and the obtained frequency P n The frequency threshold P of the book 阈1 For comparison:

[0031] If P n ≥P 阈1 , then the current keyword is determined to be a core keyword, and the current keyword is stored in the core keyword set, and then step S156 is continued;

[0032] If P n <P 阈1 , it is determined that the current keyword is not a core keyword, and then the step S156 is continued;

[0033] S155, obtaining the frequency P of the current keyword appearing in the book text n , and the obtained frequency P n The frequency threshold P of the book 阈2 For comparison, P 阈2 >P 阈1 :

[0034] If P n ≥P 阈2 , then the current keyword is determined to be a core keyword, and the current keyword is stored in the core keyword set, and then step S156 is continued;

[0035] If P n <P 阈2 , it is determined that the current keyword is not a core keyword, and then the step S156 is continued;

[0036] S156, increase the value of n by 1, and then compare the value of n with the classification keyword threshold N of the book. 阈 For comparison:

[0037] If n≤N 阈 , then step S152 needs to be executed again to determine the next round of core keywords;

[0038] If n>N 阈 , the core keywords are determined.

[0039] Furthermore, the P 阈2 >1.5*P 阈1 .

[0040] Furthermore, the step S16 includes:

[0041] S161, obtain the frequency F of each core keyword in the text m and level factor L m ;

[0042] S162, calculate the depth index α of the core keyword m , α m =F m *L m *K m , where K m It is the correction factor of the core keywords.

[0043] Furthermore, in step S2: the user inputs the demand in the form of voice or text.

[0044] Furthermore, the step S3 includes:

[0045] S31, comparing the search keyword with the core keyword set and the extended keyword set of each book in turn, and determining whether the search keyword appears in the core keyword set and the extended keyword set of the core keyword:

[0046] If all search keywords appear in the core keyword set, the current book is regarded as a first-level preliminary screening book, and then step S32 is continued;

[0047] If all search keywords appear in the union of the core keyword set and the expanded keyword set, the current book is taken as a secondary preliminary screening book, and then step S33 is continued;

[0048] If all search keywords do not appear simultaneously in the union of the core keyword set and the expanded keyword set, it is determined that the current book does not match the target book;

[0049] S32, calculating the comprehensive relevance index θ of the current book, and storing the comprehensive relevance index θ of the current book in a first-level comprehensive relevance index set;

[0050] S33, calculating the comprehensive relevance index θ of the current book, and correcting the calculated comprehensive relevance index θ according to the difference between the search keyword and the matching keyword, and then storing the corrected comprehensive relevance index θ in the secondary comprehensive relevance index set;

[0051] S34, merging the first-level comprehensive correlation index set and the second-level comprehensive correlation index set, and sorting them according to the size of the comprehensive correlation index to form a final comprehensive correlation index set;

[0052] S35, multiple books ranked high in comprehensive relevance index are pushed to the user as search results.

[0053] Further, the step S32 includes:

[0054] S321, determining the value of the calibration factor β1 of the matching keyword according to the ranking of the matching keyword in the core keyword set;

[0055] S322, based on the calibration factor β1 and the depth index α of the matching keyword m Calculate the comprehensive relevance index θ of the current book;

[0056] S323, storing the comprehensive relevance index θ of the current book into the first-level comprehensive relevance index set.

[0057] Furthermore, the step S33 includes:

[0058] S331, determining the value of the calibration factor β2 of the matching keyword according to the ranking of the core keyword corresponding to the matching keyword in the core keyword set;

[0059] S332, based on the calibration factor β2 and the depth index α of the matching keyword m Calculate the comprehensive relevance index θ of the current book;

[0060] S333, correcting the calculated comprehensive correlation index θ according to the difference between the search keyword and the matching keyword;

[0061] S334, storing the corrected comprehensive correlation index θ into the secondary comprehensive correlation index set.

[0062] An intelligent library consultation system includes a computer-readable storage medium storing a computer program and a processor. When the computer program is read by the processor, the intelligent library consultation method can be implemented.

[0063] The beneficial effect of the present invention is that the intelligent library consultation method and system of the present invention first establish the core keyword set and the extended keyword set corresponding to the book, and the method of searching for keywords, which is essentially a retrieval method that brings the books to be searched in the library and the user's search needs close to, finds, and matches each other. This is a new "two-way" way of finding books. Compared with the traditional "people looking for books" method, it can make the retrieval process more intelligent and simple, and the retrieval results more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 It is an overall flow chart of the intelligent library consultation method of the present invention;

[0065] Figure 2 It is a flowchart of determining the core keyword set and the expanded keyword set in step S1 of the present invention;

[0066] Figure 3 It is a flowchart of determining core keywords according to the primary, secondary and tertiary keywords in step S15 of the present invention;

[0067] Figure 4 is the depth index α in step S16 of the present invention m Flow chart of the calculation process;

[0068] Figure 5 It is a flow chart of outputting search results by comparing search keywords, core keywords and expanded keywords in step S3 of the present invention. DETAILED DESCRIPTION

[0069] The following will be combined with the drawings in the embodiments of the present application to clearly describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments in the present application belong to the scope of protection of this application.

[0070] In the description of this application, it should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to this application. Technologies, methods and devices known to ordinary technicians in the relevant field may not be discussed in detail, but where appropriate, the technologies, methods and devices should be considered as part of the authorization specification.

[0071] It should be noted that, in the present application, the terms "comprise", "include" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises one..." does not exclude the presence of other identical elements in the process, method, article or device including the element. In addition, it should be noted that the scope of the method and device in the embodiment of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in reverse order according to the functions involved, for example, the described method may be performed in an order different from that described, and various steps may also be added, omitted, or combined. In addition, the features described with reference to certain examples may be combined in other examples.

[0072] like Figures 1 to 5 As shown, an intelligent library consultation method comprises the steps of:

[0073] S1, establishing the core keywords and extended keywords corresponding to each book according to its content, and forming a core keyword set and an extended keyword set corresponding to the book;

[0074] S2, establish search keywords based on user needs;

[0075] S3 outputs search results by comparing search keywords, core keywords and expanded keywords.

[0076] In the intelligent library consultation method of the present invention, before outputting the search results according to the user's needs, the core keywords and extended keywords of each book are established in advance according to the content of each book, and these core keywords and extended keywords are used as elements of the set to form a core keyword set corresponding to the book one by one and an extended keyword set corresponding to each core keyword one by one. Through this step, the main content of the book is extracted, identified, summarized and simplified, so that in the comparison process in the subsequent step S3, the search results can be quickly and accurately output by comparing the consistency between the search keywords, core keywords and extended keywords.

[0077] The traditional, one-way "people looking for books" method is mostly carried out by comparing the search keywords established by the user with all the text in the entire content or partial content (such as title, abstract) of all books in the database. This method requires a large amount of data to be compared and is prone to a large amount of search noise. Compared with the traditional search mode, in the intelligent library consultation method described in the present invention, the main content of the book and the user's needs are first analyzed and simplified through steps S1 and S2 respectively, and the core keyword set and extended keyword set corresponding to the book, as well as the search keywords that reflect the user's needs, are obtained. Then, the search results are output by comparing the search keywords, core key words and extended keywords. In this way, the amount of data to be compared can be greatly reduced, a large amount of search noise can be eliminated, and the search process is simpler and faster; the search results are more accurate.

[0078] At the same time, in the process of comparing books with user needs, the present invention first establishes the core keyword set and the extended keyword set corresponding to the book, and the method of searching keywords is essentially a retrieval method that brings the searched books in the library and the user's search needs close to, finds, and matches each other. This is a new "two-way" way of finding books. Compared with the traditional "people looking for books" method, it can make the retrieval process more intelligent and simple, and the retrieval results more accurate.

[0079] Furthermore, the step S1 includes:

[0080] S11, obtaining the electronic version file of the book;

[0081] S12, extracting keywords from the book title and using them as first-level keywords;

[0082] S13, extract keywords from the abstract, preface, and table of contents and use them as secondary keywords;

[0083] S14, extracting keywords from the text and using them as third-level keywords;

[0084] S15, determining core keywords based on primary, secondary and tertiary keywords;

[0085] S16, based on the frequency F of each core keyword in the text m and level factor L m Calculate the depth index α of each core keyword m ;

[0086] S17, according to the depth index α m Sort the core keywords in the core keyword set by the size of to obtain a sorted core keyword set;

[0087] S18, performing synonym expansion on each core keyword to obtain an expanded keyword set of each core keyword.

[0088] In step S12, when the book lacks one or more of the abstract, preface and table of contents, one or more of the abstract, preface and table of contents may be selected to extract secondary keywords according to actual conditions.

[0089] In step S16, the value of m is 1, 2, 3, ..., m max , where m max is the total number of core keywords. Generally, the value of m corresponds to the core keywords one by one. Thus, the word frequency F m and level factor L m It will also correspond to each core keyword one by one.

[0090] As some examples of the present invention, in steps S12, S13, and S14, the establishment of primary keywords, secondary keywords, and tertiary keywords can be carried out by TF-IDF (Term Frequency-Inverse Document Frequency) keyword extraction method, word frequency statistics method, deep learning-based clustering algorithm and model training method, etc. These methods have been clearly disclosed in the prior art and will not be repeated here.

[0091] Furthermore, in step S15, the process of determining the core keywords according to the primary, secondary and tertiary keywords includes:

[0092] S151, obtaining first, second and third level keywords, and storing them in a hierarchical keyword set, forming a hierarchical keyword set including keywords at each level;

[0093] S152, obtaining the nth keyword in the hierarchical keyword set, where the initial value of n is 1;

[0094] S153, the nth keyword obtained in step S152 is compared with the elements in the primary keyword set consisting of primary keywords, the secondary keyword set consisting of secondary keywords, and the tertiary keyword set consisting of tertiary keywords respectively:

[0095] If the primary keyword set, the secondary keyword set, and the tertiary keyword set all include the current keyword, the current keyword is determined to be a core keyword, and the current keyword is stored in the core keyword set, and then step S156 is continued;

[0096] If two of the first-level keyword set, the second-level keyword set, and the third-level keyword set contain the current keyword, then step S154 needs to be continued;

[0097] If only one of the primary keyword set, the secondary keyword set, and the tertiary keyword set contains the current keyword, then step S155 needs to be continued;

[0098] S154, obtaining the frequency P of the current keyword appearing in the book text n , and the obtained frequency P n The frequency threshold P of the book 阈1 For comparison:

[0099] If P n ≥P 阈1 , then the current keyword is determined to be a core keyword, and the current keyword is stored in the core keyword set, and then step S156 is continued;

[0100] If P n <P 阈1 , it is determined that the current keyword is not a core keyword, and then the step S156 is continued;

[0101] S155, obtaining the frequency P of the current keyword appearing in the book text n , and the obtained frequency P n The frequency threshold P of the book 阈2 For comparison, P 阈2 >P 阈1 :

[0102] If P n ≥P 阈2 , then the current keyword is determined to be a core keyword, and the current keyword is stored in the core keyword set, and then step S156 is continued;

[0103] If P n <P 阈2 , it is determined that the current keyword is not a core keyword, and then the step S156 is continued;

[0104] S156, increase the value of n by 1, and then compare the value of n with the classification keyword threshold N of the book. 阈 For comparison:

[0105] If n≤N 阈 , then step S152 needs to be executed again to determine the next round of core keywords;

[0106] If n>N 阈 , the core keywords are determined.

[0107] Among them, the P 阈1 and P 阈2 It can be determined based on experience or experiments. 阈1 and P 阈2 The value of is related to the total number of words in the book. For example, for books with a total number of words between 30,000 and 50,000, the value of P 阈1 and P 阈2 The value of can be 50 to 100; for books with a total word count of 60,000 to 100,000, the P 阈1 and P 阈2 The value can be 100 to 200, etc.

[0108] In addition, the N 阈 It is the total number of first-level keywords, second-level keywords, and third-level keywords.

[0109] Preferably, the P 阈2 >1.5*P 阈1 .

[0110] In the process of determining the core keywords given in step S15, factors such as the position of the keywords in the book and the frequency of occurrence are combined, so that the selection of core keywords can have the advantages of being comprehensive and accurate, and can provide a good foundation for the subsequent comparison process.

[0111] Furthermore, the step S16 includes:

[0112] S161, obtain the frequency F of each core keyword in the text m and level factor L m ;

[0113] S162, calculate the depth index α of the core keyword m , α m =F m *L m *K m .

[0114] Among them, the frequency of the core keywords in the text is F m The number of times the core keyword appears in the text;

[0115] Level factor L m It reflects the influence of the position of the core keyword in the first-level keyword set, the second-level keyword set, and the third-level keyword set on the depth index. m =L1+L2+L3, when L1, L2, L3 are non-zero values, L1>L2>L3, the values ​​of L1, L2, L3 can be pre-set and stored in the system.

[0116] As an example of the present invention, when the core keyword appears in the first-level keyword set, L1=0.5~1.0, otherwise L1=0; when the core keyword appears in the second-level keyword set, L2=0.3~0.6, otherwise L2=0; when the core keyword appears in the third-level keyword set, L3=0.1~0.3, otherwise L3=0.

[0117] K m is the correction coefficient of the core keyword. Usually, K m The value is negatively correlated with the total number of words in the book, that is, the larger the total number of words in the book, the higher the K m The smaller the value of K is, the smaller the total number of words in the book is, m The larger the value of .

[0118] Preferably, K m The value of can be the reciprocal of the total number of words in the book.

[0119] As some examples of the present invention, in step S18: the method of performing synonym expansion on each core keyword can be: using synonym tools, such as Thesaurus.com or synonym dictionary websites, etc. to find synonyms; using semantic networks or knowledge graphs, such as the semantic relationship between words provided by WordNet to find synonyms; using corpus tools, such as AntConc, WordSmith, etc., to perform word frequency statistics and vocabulary collocation analysis to find synonyms related to keywords; using word embedding technology, such as Word2Vec, GloVe, etc., to map words to high-dimensional vector space, and find synonyms by calculating the similarity between vectors; using semantic similarity calculation tools, such as BERT, RoBERTa and other pre-trained language models, to evaluate the semantic similarity between different words to determine synonyms. The above technology has been clearly disclosed in the prior art and will not be repeated here.

[0120] As some examples of the present invention, in step S2: the user may input the demand in voice, text or other methods.

[0121] As some examples of the present invention, in step S2: the method of establishing the search keywords according to user needs can be: a clustering algorithm and a model training method based on deep learning, etc.

[0122] Furthermore, the step S3 includes:

[0123] S31, comparing the search keyword with the core keyword set and the extended keyword set of each book in turn, and determining whether the search keyword appears in the core keyword set and the extended keyword set of the core keyword:

[0124] If all search keywords appear in the core keyword set, it means that the core keyword of the current book is the same as the search keyword, and the current book is taken as a first-level preliminary screening book, and then step S32 is continued;

[0125] If all search keywords appear in the union of the core keyword set and the extended keyword set, it means that the core keyword of the current book is equal to the search keyword, and the current book is taken as a secondary preliminary screening book, and then step S33 is continued;

[0126] If all the search keywords do not appear simultaneously in the union of the core keyword set and the extended keyword set, it means that the core keyword of the current book is different from the search keyword, and it is determined that the current book does not match the target book, and the screening of the current book ends;

[0127] S32, calculating the comprehensive relevance index θ of the current book, and storing the comprehensive relevance index θ of the current book in a first-level comprehensive relevance index set;

[0128] S33, calculating the comprehensive relevance index θ of the current book, and correcting the calculated comprehensive relevance index θ according to the difference between the search keyword and the matching keyword, and then storing the corrected comprehensive relevance index θ in the secondary comprehensive relevance index set;

[0129] S34, merging the first-level comprehensive correlation index set and the second-level comprehensive correlation index set, and sorting them according to the size of the comprehensive correlation index to form a final comprehensive correlation index set;

[0130] S35, multiple books ranked high in comprehensive relevance index are pushed to the user as search results.

[0131] It should be noted that the matching keyword in the present invention refers to the core keyword that is the same as the search keyword or an extended keyword of an equivalent core keyword. Specifically, when the search keyword is the same as the core keyword, the matching keyword is the core keyword; and when the search keyword is the same as the extended keyword of the core keyword, the matching keyword is the extended keyword corresponding to the core keyword.

[0132] Furthermore, in step S3, it is necessary to traverse each book in turn and screen the books to obtain a first-level comprehensive relevance index set and a second-level comprehensive relevance index set containing a plurality of elements.

[0133] Specifically, when the core keywords of the current book are different from the search keywords, it is determined that the current book does not match the target book. After the screening of the current book is ended, the screening and comparison of the next book can be carried out; and when the core keywords of the current book are the same or equal to the search keywords, the comprehensive relevance index θ of the current book can be calculated, and the comprehensive relevance index θ of the current book can be stored in the first-level comprehensive relevance index set or the second-level comprehensive relevance index set before entering the screening and comparison of the next book.

[0134] Furthermore, the step S32 includes:

[0135] S321, determining the value of the calibration factor β1 of the matching keyword according to the ranking of the matching keyword in the core keyword set. Generally, the higher the ranking, the larger the value of the calibration factor β1. 0<the value of the calibration factor β1<1. The value of β1 can be set by experiment or based on experience.

[0136] S322, based on the calibration factor β1 and the depth index α of the matching keyword m Calculate the comprehensive relevance index θ of the current book;

[0137] S323, storing the comprehensive relevance index θ of the current book into the first-level comprehensive relevance index set.

[0138] Furthermore, in step S322, when there is only one search keyword, there will be only one corresponding matching keyword. At this time, the comprehensive correlation index θ=β1*α m ; When there are multiple search keywords, there will also be multiple corresponding matching keywords. The search keywords and the corresponding matching keywords maintain a one-to-one correspondence. At this time, it is necessary to calculate the comprehensive relevance index θ of the matching keywords corresponding to each search keyword in turn, and then calculate the average value of each comprehensive relevance index θ, and use it as the comprehensive relevance index θ of the current book.

[0139] It can be understood that in step S32, the matching keyword is the core keyword, so the matching keyword has the same depth index α as its corresponding core keyword. m .

[0140] Furthermore, the step S33 includes:

[0141] S331, determining the value of the calibration factor β2 of the matching keyword according to the ranking of the core keyword corresponding to the matching keyword in the core keyword set; generally, the higher the ranking, the larger the value of the calibration factor β2, 0<the value of the calibration factor β2<1, and the value of β2 can be set by experiment or based on experience;

[0142] S332, based on the calibration factor β2 and the depth index α of the matching keyword m Calculate the comprehensive relevance index θ of the current book;

[0143] S333, correcting the calculated comprehensive correlation index θ according to the difference between the search keyword and the matching keyword;

[0144] S334, storing the corrected comprehensive correlation index θ into the secondary comprehensive correlation index set.

[0145] It should be noted that in step S331, when the matching keyword is a core keyword, the calibration factor β2 and the depth index α of the matching keyword are m Consistent with the corresponding core keyword; when the matching keyword is the extended keyword of the core keyword, the calibration factor β2 and depth index α of the matching keyword m It is consistent with the core keywords corresponding to the expanded keywords.

[0146] Similarly, in step S332, when there is only one search keyword, there will be only one corresponding matching keyword. At this time, the comprehensive correlation index θ=β2*α m ; When there are multiple search keywords, there will also be multiple corresponding matching keywords. At this time, it is necessary to calculate the comprehensive relevance index θ of the matching keywords corresponding to each search keyword in turn, and then calculate the average value of each comprehensive relevance index θ, and use it as the comprehensive relevance index θ of the current book.

[0147] Furthermore, in the step S333, the calculated comprehensive correlation index θ can be corrected according to the difference between the search keyword and the matching keyword by multiplying the comprehensive correlation index θ by a correction coefficient. Preferably, the value of the correction coefficient is 0.8 to 1. When the matching keyword of the search keyword is a core keyword, the value of the correction coefficient is 1. When the matching keyword of the search keyword is an extended keyword of the core keyword, the value of the correction coefficient is less than 1. The specific value can be determined by comparing the difference between the search keyword and the matching keyword.

[0148] Furthermore, in step S35, the top 5 to 50 books may be pushed to the user as search results.

[0149] Furthermore, when the search results are pushed to the users, the library collection information, such as quantity, borrowing status, storage location, etc., can also be delivered to the users.

[0150] In addition, the present invention also provides an intelligent library consultation system, which includes a computer-readable storage medium storing a computer program and a processor. When the computer program is read by the processor, the above-mentioned intelligent library consultation method can be implemented.

[0151] The embodiments of the present application are described above in conjunction with the accompanying drawings. In the absence of conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present application, ordinary technicians in this field can also make many forms without departing from the purpose of the present application and the scope of protection of the claims, all of which are within the protection of the present application.

Claims

1. An intelligent library consultation method, characterized in that: Includes steps: S1, establishing the core keywords and extended keywords corresponding to each book according to its content, and forming a core keyword set and an extended keyword set corresponding to the book; S2, establish search keywords based on user needs; S3 outputs search results by comparing search keywords, core keywords and expanded keywords.

2. The intelligent library consultation method according to claim 1, characterized in that: The step S1 comprises: S11, obtaining the electronic version file of the book; S12, extracting keywords from the book title and using them as first-level keywords; S13, extract keywords from the abstract, preface, and table of contents and use them as secondary keywords; S14, extracting keywords from the text and using them as third-level keywords; S15, determining core keywords based on primary, secondary and tertiary keywords; S16, based on the frequency F of each core keyword in the text m and level factor L m Calculate the depth index α of each core keyword m ; S17, according to the depth index α m sort the core keywords in the core keyword set by the size of to obtain a sorted core keyword set; S18, performing synonym expansion on each core keyword to obtain an expanded keyword set of each core keyword.

3. The intelligent library consultation method according to claim 2, characterized in that: The step S15 comprises: S151, obtaining first, second and third level keywords, and storing them in a hierarchical keyword set, forming a hierarchical keyword set including keywords at each level; S152, obtaining the nth keyword in the hierarchical keyword set, where the initial value of n is 1; S153, the nth keyword obtained in step S152 is compared with the elements in the primary keyword set consisting of primary keywords, the secondary keyword set consisting of secondary keywords, and the tertiary keyword set consisting of tertiary keywords respectively: If the primary keyword set, the secondary keyword set, and the tertiary keyword set all include the current keyword, the current keyword is determined to be a core keyword, and the current keyword is stored in the core keyword set, and then step S156 is continued; If two of the first-level keyword set, the second-level keyword set, and the third-level keyword set contain the current keyword, then step S154 needs to be continued; If only one of the primary keyword set, the secondary keyword set, and the tertiary keyword set contains the current keyword, then step S155 needs to be continued; S154, obtaining the frequency P of the current keyword appearing in the book text n , and the obtained frequency P n The frequency threshold P of the book 阈1 For comparison: If P n ≥P 阈1 , then the current keyword is determined to be a core keyword, and the current keyword is stored in the core keyword set, and then step S156 is continued; If P n <P 阈1 , it is determined that the current keyword is not a core keyword, and then the step S156 is continued; S155, obtaining the frequency P of the current keyword appearing in the book text n , and the obtained frequency P n The frequency threshold P of the book 阈2 For comparison, P 阈2 >P 阈1 : If P n ≥P 阈2 , then the current keyword is determined to be a core keyword, and the current keyword is stored in the core keyword set, and then step S156 is continued; If P n <P 阈2 , it is determined that the current keyword is not a core keyword, and then the step S156 is continued; S156, increase the value of n by 1, and then compare the value of n with the classification keyword threshold N of the book. 阈 For comparison: If n≤N 阈 , then step S152 needs to be executed again to determine the next round of core keywords; If n>N 阈 , the core keywords are determined.

4. The intelligent library consultation method according to claim 3, characterized in that: The P 阈2 >1.5*P 阈1 .

5. The intelligent library consultation method according to claim 2, characterized in that: The step S16 comprises: S161, obtain the frequency F of each core keyword in the text m and level factor L m ; S162, calculate the depth index α of the core keyword m , α m =F m *L m *K m , where K m It is the correction factor of the core keywords.

6. The intelligent library consultation method according to claim 1, characterized in that: In step S2: the user inputs the requirement in the form of voice or text.

7. The intelligent library consultation method according to claim 1, characterized in that: The step S3 comprises: S31, comparing the search keyword with the core keyword set and the extended keyword set of each book in turn, and determining whether the search keyword appears in the core keyword set and the extended keyword set of the core keyword: If all search keywords appear in the core keyword set, the current book is regarded as a first-level preliminary screening book, and then step S32 is continued; If all search keywords appear in the union of the core keyword set and the expanded keyword set, the current book is taken as a secondary preliminary screening book, and then step S33 is continued; If all search keywords do not appear simultaneously in the union of the core keyword set and the expanded keyword set, it is determined that the current book does not match the target book; S32, calculating the comprehensive relevance index θ of the current book, and storing the comprehensive relevance index θ of the current book in a first-level comprehensive relevance index set; S33, calculating the comprehensive relevance index θ of the current book, and correcting the calculated comprehensive relevance index θ according to the difference between the search keyword and the matching keyword, and then storing the corrected comprehensive relevance index θ in the secondary comprehensive relevance index set; S34, merging the first-level comprehensive correlation index set and the second-level comprehensive correlation index set, and sorting them according to the size of the comprehensive correlation index to form a final comprehensive correlation index set; S35, multiple books ranked high in comprehensive relevance index are pushed to the user as search results.

8. The intelligent library consultation method according to claim 7, characterized in that: The step S32 comprises: S321, determining the value of the calibration factor β1 of the matching keyword according to the ranking of the matching keyword in the core keyword set; S322, based on the calibration factor β1 and the depth index α of the matching keyword m Calculate the comprehensive relevance index θ of the current book; S323, storing the comprehensive relevance index θ of the current book into the first-level comprehensive relevance index set.

9. The intelligent library consultation method according to claim 7, characterized in that: The step S33 comprises: S331, determining the value of the calibration factor β2 of the matching keyword according to the ranking of the core keyword corresponding to the matching keyword in the core keyword set; S332, based on the calibration factor β2 and the depth index α of the matching keyword m Calculate the comprehensive relevance index θ of the current book; S333, correcting the calculated comprehensive correlation index θ according to the difference between the search keyword and the matching keyword; S334, storing the corrected comprehensive correlation index θ into the secondary comprehensive correlation index set.

10. An intelligent library consultation system, comprising a computer-readable storage medium storing a computer program and a processor, wherein when the computer program is read by the processor, the intelligent library consultation method according to any one of claims 1 to 9 can be implemented.

Images