Library resource intelligent recommendation method and system based on natural language processing

Abstract

The invention belongs to the technical field of natural language processing, and particularly relates to a library resource intelligent recommendation method and system based on natural language processing, and the method comprises the steps: obtaining historical reading information of a user; performing LDA subject analysis on each library resource in the historical reading list to obtain a subject category list; obtaining theme category preference information according to the theme category list and the reading duration information; obtaining first average reading duration information according to the theme category list and the reading duration information; obtaining recommendation values of all library resources according to the theme category preference information and the first average reading duration information, and obtaining a recommendation list of the user according to the recommendation values of all library resources; detecting feedback information of a user on the recommendation list; and updating the recommendation list according to the feedback information. The method can improve user experience and recommendation quality, and solves the problem of repeated and fixed recommendation of library resources.

Description

Technical Field

[0001] The present application belongs to the field of natural language processing technology, and in particular, to a method and system for intelligently recommending library resources based on natural language processing. Background Art

[0002] Natural language processing is an interdisciplinary field of computer science, artificial intelligence and linguistics, dedicated to achieving the ability for people and computers to communicate effectively through natural language. With the rapid development of artificial intelligence, communication between people and computers is becoming more and more convenient and efficient. Using natural language processing technology to let computers recommend natural language information such as music and books to users is a common method for platforms that include large amounts of data.

[0003] In existing library resource recommendation methods, the subject classification of library resources relies on the subjective judgment of authors or readers, and there is a lack of steps to detect user feedback on the recommendation results. Fixed library resources are often repeatedly recommended based on the user's historical reading records. Summary of the invention

[0004] The embodiments of the present application provide a method and system for intelligent recommendation of library resources based on natural language processing, which can solve the problem that the subject classification of library resources depends on the subjective judgment of authors or readers, lacks the step of detecting user feedback for the recommendation results, and often repeatedly recommends fixed library resources based on the user's historical reading records.

[0005] In a first aspect, an embodiment of the present application provides a library resource intelligent recommendation method based on natural language processing, wherein the library resource intelligent recommendation method based on natural language processing is applied to a library resource intelligent recommendation device, and the method comprises: Acquire the user's historical reading information; wherein the historical reading information includes a historical reading list and reading time information, and the reading time information includes the user's reading time for each library resource in the historical reading list; Performing LDA topic analysis on each library resource in the historical reading list to obtain the subject category distribution of each library resource in the historical reading list, and integrating the subject category distribution of each library resource to obtain a subject category list; wherein the subject category list refers to a list recording the subject category distribution of each library resource in the historical reading list, and the subject category distribution of the library resources in the subject category list corresponds one-to-one to the library resources in the historical reading list; Obtaining subject category preference information according to the subject category list and the reading time information; wherein the subject category preference information is used to reflect the user's preference for library resources of various subject categories; Obtaining first average reading time information according to the subject category list and the reading time information; wherein the first average reading time information refers to information recording the average reading time of users for library resources of various subject categories, so as to reflect the reading time preference of users for library resources of different subject categories; According to the subject category preference information and the first average reading time information, the recommendation values ​​of all library resources are obtained, and according to the recommendation values ​​of all library resources, a user's recommendation list is obtained; wherein the recommendation value is used to reflect the quantitative value of the recommendation degree of a library resource; Detecting feedback information of the user on the recommendation list; wherein the feedback information is used to reflect the active feedback operation and passive feedback operation of the user on the recommendation list; The recommendation list is updated according to the feedback information.

[0006] The above technical solutions in the embodiments of the present application have at least the following technical effects: The method for intelligent recommendation of library resources based on natural language processing provided by the present application first obtains the user's historical reading information, and the historical reading information includes a historical reading list and reading time information; secondly, each library resource in the historical reading list is input into the LDA topic model to obtain the subject category distribution of each library resource in the historical reading list, and the subject categories of each library resource are integrated to obtain a subject category list. This step allows the computer to determine the subject category of library resources through natural language processing technology, which can solve the problem that the subject classification of library resources depends on the subjective judgment of the author or reader; then, the subject category preference information is obtained according to the subject category list and the reading time information. This step uses the user's equivalent reading time for different subject categories to represent the user's subject category preference, which is conducive to intuitively and quickly obtaining the user's subject category preference; then, the first average reading time information is obtained according to the subject category list and the reading time information. This step The first step is to obtain the average reading time of users for library resources of different subject categories to indicate the reading time preference of users for different subject categories, which is conducive to improving the quality of personalized recommendations; then, according to the subject category preference information and the first average reading time information, the recommendation values ​​of all library resources are obtained, and the user's recommendation list is obtained according to the recommendation values ​​of all library resources. After determining the user's subject category preference and reading time preference, this step obtains the recommendation values ​​of all library resources according to the user's preferences, and determines the recommendation list based on the recommendation values ​​of all library resources, which is conducive to standardizing the recommendation standards and improving the quality of personalized recommendations; then, the user's feedback information on the recommendation list is detected. This step can reflect the user's feedback on the recommendation list and improve the user's recommendation experience; finally, the recommendation list is updated according to the feedback information. This step updates the recommendation list according to the user's feedback on the recommendation list, which can solve the problem of repeated recommendation of fixed library resources. This method determines the subject category distribution of library resources through the LDA topic model, which solves the problem that the subject classification of library resources depends on the subjective judgment of authors or readers. After determining the recommendation list according to the user's subject category preference and reading time preference, the user's feedback on the recommendation list is obtained, and the recommendation list is updated according to the user's feedback on the recommendation list. This can improve user experience and recommendation quality and solve the problem of repeatedly recommending fixed library resources based on the user's historical reading records.

[0007] In a second aspect, the present application embodiment provides a library resource recommendation system, including: An acquisition unit, used to acquire the user's historical reading information; wherein the historical reading information includes a historical reading list and reading time information, and the reading time information includes the user's reading time for each library resource in the historical reading list; A subject category judgment unit is used to perform LDA subject analysis on each library resource in the historical reading list to obtain the subject category distribution of each library resource in the historical reading list, and integrate the subject category distribution of each library resource to obtain a subject category list; wherein, the subject category list refers to a list recording the subject category distribution of each library resource in the historical reading list, and the subject category distribution of the library resources in the subject category list corresponds one-to-one with the library resources in the historical reading list; An analysis unit, used to obtain subject category preference information according to the subject category list and the reading time information; wherein the subject category preference information is used to reflect the user's preference for library resources of various subject categories; The analysis unit is further used to obtain first average reading time information according to the subject category list and the reading time information; wherein the first average reading time information refers to information recording the average reading time of users for library resources of various subject categories, so as to reflect the reading time preference of users for library resources of different subject categories; A recommendation unit, used to obtain the recommendation value of all library resources according to the subject category preference information and the first average reading time information, and obtain the user's recommendation list according to the recommendation value of all library resources; wherein the recommendation value is used to reflect the quantitative value of the recommendation degree of a library resource; A feedback unit, used to detect feedback information of the user on the recommendation list; wherein the feedback information is used to reflect the active feedback operation and passive feedback operation of the user on the recommendation list; An updating unit is used to update the recommendation list according to the feedback information.

[0008] In a third aspect, an embodiment of the present application provides a library resource intelligent recommendation device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements any of the methods described in the first aspect when executing the computer program.

[0009] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described in any one of the first aspects above is implemented.

[0010] In a fifth aspect, an embodiment of the present application provides a computer program product. When the computer program product runs on a library resource intelligent recommendation device, the library resource intelligent recommendation device executes the library resource intelligent recommendation method based on natural language processing as described in any one of the above-mentioned first aspects.

[0011] It can be understood that the beneficial effects of the second to fifth aspects mentioned above can be found in the relevant description of the first aspect mentioned above, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0013] Figure 1 It is a flowchart of a method for intelligently recommending library resources based on natural language processing provided by an embodiment of the present application; Figure 2 It is a functional shape diagram of the normal distribution density function of step S523 of the library resource intelligent recommendation method based on natural language processing provided in one embodiment of the present application; Figure 3 It is a structural diagram of a library resource intelligent recommendation system based on natural language processing provided in an embodiment of the present application; Figure 4 It is a structural diagram of the server and terminal device provided in the embodiment of the present application. DETAILED DESCRIPTION

[0014] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0015] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0016] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0017] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when" or "uponce" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is detected" or "in response to detecting [described condition or event]", depending on the context.

[0018] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0019] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0020] In related technologies, the subject classification of library resources is often determined based on the author's own classification or readers' feedback, which is greatly influenced by subjective factors. It is inevitable that there will be multiple standards when defining subject classification. In addition, there is a lack of steps to detect user feedback for recommendation results, making it difficult to dynamically update recommendation results. The recommendation results can only be changed when the user changes his historical reading records, resulting in repeated recommendations of fixed library resources based on the user's historical reading records.

[0021] To solve the above problems, an embodiment of the present application provides a method for intelligent recommendation of library resources based on natural language processing. In the method for intelligent recommendation of library resources based on natural language processing, first, the user's historical reading information is obtained, and the historical reading information includes a historical reading list and reading time information; secondly, each library resource in the historical reading list is input into the LDA topic model to obtain the subject category distribution of each library resource in the historical reading list, and the subject categories of each library resource are integrated to obtain a subject category list. This step allows the computer to determine the subject category of library resources through natural language processing technology, which can solve the problem that the subject classification of library resources depends on the subjective judgment of the author or reader; then, the subject category preference information is obtained based on the subject category list and the reading time information. This step uses the user's equivalent reading time for different subject categories to represent the user's subject category preference, which is conducive to intuitively and quickly obtaining the user's subject category preference; then, the first average reading time information is obtained based on the subject category list and the reading time information. This step obtains The average reading time of users for library resources of different subject categories is used to indicate the reading time preference of users for different subject categories, which is conducive to improving the quality of personalized recommendations; then, based on the subject category preference information and the first average reading time information, the recommendation values ​​of all library resources are obtained, and the user's recommendation list is obtained based on the recommendation values ​​of all library resources. After determining the user's subject category preference and reading time preference, this step obtains the recommendation values ​​of all library resources based on the user's preferences, and determines the recommendation list based on the recommendation values ​​of all library resources, which is conducive to standardizing the recommendation standards and improving the quality of personalized recommendations; then, the user's feedback information on the recommendation list is detected. This step can reflect the user's feedback on the recommendation list and improve the user's recommendation experience; finally, the recommendation list is updated according to the feedback information. This step updates the recommendation list according to the user's feedback on the recommendation list, which can solve the problem of repeated recommendation of fixed library resources. This method determines the subject category distribution of library resources through the LDA topic model, which solves the problem that the subject classification of library resources depends on the subjective judgment of authors or readers. After determining the recommendation list according to the user's subject category preference and reading time preference, the user's feedback on the recommendation list is obtained, and the recommendation list is updated according to the user's feedback on the recommendation list. This can improve user experience and recommendation quality and solve the problem of repeatedly recommending fixed library resources based on the user's historical reading records.

[0022] The library resource intelligent recommendation method based on natural language processing provided in the embodiment of the present application can be applied to a library resource intelligent recommendation device. In this case, the library resource intelligent recommendation device is the execution entity of the library resource intelligent recommendation method based on natural language processing provided in the embodiment of the present application. The embodiment of the present application does not impose any restrictions on the specific type of the library resource intelligent recommendation device.

[0023] For example, the library resource intelligent recommendation device may include a server connected to a terminal device, which may be a smart phone or a computer. The server may obtain the user's historical reading information, input the library resources into the LDA topic model to obtain the topic category distribution, detect the user's feedback operation on the terminal device, send the library resources and information related to the library resources to the terminal device, and send a personalized recommendation list to the terminal device.

[0024] A server can be a desktop computer, a PC, a cloud server, a large local server, etc.

[0025] In order to better understand the intelligent recommendation method for library resources based on natural language processing provided in the embodiment of the present application, the specific implementation process of the intelligent recommendation method for library resources based on natural language processing provided in the embodiment of the present application is exemplarily introduced below.

[0026] Figure 1 A schematic flow chart of a method for intelligently recommending library resources based on natural language processing provided in an embodiment of the present application is shown. The method for intelligently recommending library resources based on natural language processing includes: S100, obtaining the user's historical reading information, wherein the historical reading information includes a historical reading list and reading time information, and the reading time information includes the user's reading time for each library resource in the historical reading list.

[0027] It can be understood that when a user logs in to his account on the library resource recommendation system, the server finds the user's historical reading information, which includes the user's historical reading list and the reading time information corresponding to the library resources in the historical reading list.

[0028] Exemplarily, the historical reading list may be all the reading records of the user, or the reading records of the user in a recent period of time, or the most recent N library resources in all the reading records of the user, where N may be any positive integer.

[0029] With this setting, personalized recommendation results can be obtained according to different users.

[0030] S200, performing LDA topic analysis on each library resource in the historical reading list to obtain the subject category distribution of each library resource in the historical reading list, and integrating the subject category distribution of each library resource to obtain a subject category list. The subject category list refers to a list recording the subject category distribution of each library resource in the historical reading list, and the subject category distribution of the library resources in the subject category list corresponds one-to-one with the library resources in the historical reading list.

[0031] It is understandable that the LDA topic model can be used for LDA topic analysis. The LDA topic model can be but not limited to the LDA model or the PLSA model. The LDA topic model can output the topic category distribution of the document. For example, the topic category distribution output of a document is (science × 0.5, automobile × 0.3, Mercedes-Benz × 0.1, other × 0.1), then the document is a document introducing Mercedes-Benz and involving the field of science. The topic category list records the topic category distribution of each library resource in the historical reading list, and the topics in the topic category distribution that are less than the preset threshold can be screened out (such as classifying the topic category with a topic category distribution less than 0.1 as others).

[0032] The LDA topic model adopts the LDA model. First, text data from multiple fields are obtained as a corpus. Secondly, the dictionary of the LDA model is obtained (the dictionary contains at least all the words in the corpus). Then, the corpus is preprocessed by word segmentation, stop words and punctuation removal, and the word frequency in the corpus is counted to obtain a sparse matrix of "article-word frequency". The hyperparameters (number of topic categories, topic category names, word distribution words corresponding to each topic category) are preset. The sparse matrix of "article-word frequency" is used as input to train the LDA topic model. The perplexity is used to judge the performance of the LDA topic model. When the perplexity meets the standard, the model training is completed. The trained LDA topic model can output the topic category distribution of a text.

[0033] This setting can standardize the subject categories of library resources and prevent them from being influenced by the subjective opinions of authors or readers.

[0034] Optionally, the intelligent recommendation method for library resources based on natural language processing also includes: S210, classify each library resource in the historical reading list into multiple levels; wherein the multi-level classification refers to classifying the subject categories of the library resources according to the subdivision levels of different subject categories.

[0035] It can be understood that if the LDA topic model can obtain the distribution of the three subject categories of education, military, and science, education, military, and science belong to the first-level category, but these three subject categories can still be subdivided into second-level categories. For example, education can be subdivided into primary education, secondary education, university education, etc., and primary education can be further subdivided into primary mathematics education, primary Chinese education, etc. Therefore, multiple sub-LDA topic models can be set to perform second-level classification of subject categories, and then analogize to multi-level classification.

[0036] The difference between the sub-LDA topic model and the LDA topic model is that the sub-LDA topic model presets different hyperparameters (the number of topic categories, the names of topic categories, the word distribution words corresponding to each topic category), and selects text resources of secondary or multi-level category topic categories as the corpus for training.

[0037] Such a setting can subdivide the subject categories and improve the personalization of the recommendation results.

[0038] S220, detecting the recommendation level set by the user, wherein the recommendation level is used to reflect the subdivision level of the subject category that the user accepts the recommendation.

[0039] It is understandable that users can set different recommendation levels by themselves. Different recommendation levels will be recommended and calculated using different subject category levels. The higher the recommendation level, the broader the subject category classification, such as the first-level categories such as science and education. The lower the recommendation level, the more detailed the subject category classification in the subject category list, such as the second-level categories such as secondary education and university education.

[0040] This setting enables different levels of recommendation personalization based on the user's wishes.

[0041] S230, obtaining a list of subject categories according to the recommendation levels.

[0042] It can be understood that after the user sets different recommendation levels, the classification level of the subject category is determined according to the recommendation level, and then the corresponding subject category of each library resource in the historical reading list is obtained, and then the subject category list is obtained one by one according to the subject category of each library resource.

[0043] Exemplarily, the recommendation levels may correspond one-to-one to the classification levels of the subject categories.

[0044] This setting enables different levels of recommendation personalization based on the user's wishes.

[0045] S300, obtaining subject category preference information according to the subject category list and the reading time information, wherein the subject category preference information is used to reflect the user's preference for library resources of various subject categories.

[0046] It can be understood that the user's subject category preference can be represented by the user's reading time for library resources of various subject categories. For example, the user's historical reading list includes three library resources, and the subject categories of the three library resources are (science×1.0), (transportation×1.0), and (science×1.0), and the reading time is 1 day, 3 days, and 2 days respectively. The user's reading time for the science theme and the reading time for the transportation theme are both 3 days, which means that the user has an equal degree of preference for the two themes of science and transportation. Therefore, according to the user's historical reading subject categories in the subject category list and the one-to-one corresponding reading time in the reading time information, the user's reading time for each subject category can be obtained and represented as subject category preference information. The subject category preference information can be represented as a one-dimensional array with fixed subject category sorting.

[0047] Such a setting can effectively represent the user's subject category preference.

[0048] In a possible implementation, in step S300, obtaining subject category preference information according to the subject category list and reading time information includes: S310, calculating the distribution data of each subject category in the subject category list and the corresponding reading time in the reading time information to obtain the total reading time of the same category of various subject categories. The total reading time of the same category refers to the sum of the equivalent reading time of the user for the same subject category, and the equivalent reading time refers to the reading time of library resources distributed according to the subject category distribution, and distributed in equal proportion to the reading time of different subject categories.

[0049] It can be understood that by adding the equivalent reading time of the same subject category in the subject category list and the reading time information, the total reading time of the user for the same subject category can be obtained. Then, by analogy to various subject categories, the total reading time of the same category in various subject categories can be obtained.

[0050] Exemplarily, suppose there are N theme category distribution data in the theme category list, then there are N corresponding F(N) reading time data in the reading time information, and the N theme category distribution data include three theme category distribution data of a, b, and c for the A1 theme category, among which the theme category distribution of a is (A1×0.5, others×0.5), the theme category distribution of b is (A1×0.3, others×0.7), and the theme category distribution of c is (A1×0.7, others×0.3). Then, the equivalent reading time of a (0.5×F(a)), the equivalent reading time of b (0.3×F(b)), and the equivalent reading time of c (0.7×F(c)) are added to obtain the total reading time of the same category of the A1 theme category, and so on to all theme categories such as A2, A3, A4... to obtain the total reading time of the same category of various theme categories.

[0051] With such a setting, the total reading time of users for library resources of various subject categories can be accurately obtained.

[0052] S320, normalize the total reading time of the same category of various subject categories to obtain subject category preference information.

[0053] It can be understood that the total reading time of the same category is normalized. For example, the user's historical reading list includes three library resources. The subject categories of the three library resources are (science×1.0), (transportation×1.0), and (science×1.0). The reading time is 1 day, 3 days, and 2 days respectively. The user's reading time for science topics and transportation topics are both 3 days. After normalization, the preferences for science and transportation topics in the subject category preference information are both 0.5.

[0054] Such a setting can effectively and normatively represent the user's preference for different topic categories.

[0055] S400, obtaining first average reading time information according to the subject category list and the reading time information, wherein the first average reading time information refers to information recording the average reading time of users for library resources of various subject categories, so as to reflect the reading time preference of users for library resources of different subject categories.

[0056] It can be understood that in addition to the personalization of subject categories, the personalized recommendation of users also includes the personalization of reading time. The personalization of reading time can recommend library resources with appropriate text volume to users, so that the user's reading experience will not be fragmented. Personalization of reading time requires setting various targeted reading time preferences for users according to various subject categories, that is, each subject category of the user has a reading time preference. The reading time preference can be expressed by the user's average reading time for the corresponding subject category, and the user's average reading time for the corresponding subject category can be obtained through the subject category list and reading time information. The average reading time of users for a subject category = the total time of the equivalent reading time of users for this subject category / the number of readings of users for this subject category.

[0057] This setting improves the personalization and quality of recommendation results.

[0058] In a possible implementation, in step S400, obtaining first average reading time information according to the subject category list and the reading time information includes: S410, according to the subject category list and reading time information, obtain the total reading time of the same category of various subject categories and the total amount of the same category of various subject categories; wherein the total amount of the same category refers to the sum of the same subject category distribution coefficients in all subject category distribution data in the subject category list.

[0059] It can be understood that the total amount of the same category is similar to the total reading time of the same category. All the subject category distribution data in the subject category list are classified according to different subject categories. For example, if a subject category distribution data is (science×0.5, transportation×0.5), then 0.5 is added to the science subject category and 0.5 is added to the transportation subject category. This is deduced to all the subject category distribution data in the subject category list to obtain the total amount of the same category of various subject categories and obtain the total reading time of the same category of various subject categories in accordance with step S310.

[0060] Such a setting is conducive to improving the accuracy of personalization of reading time.

[0061] S420: Obtain first average reading time information according to the total reading time of the same category and the total amount of the same category in various subject categories.

[0062] It can be understood that the total reading time of the same category in a subject category is divided by the total number of the same category in the corresponding subject category to obtain the average reading time of the subject category, and the same is applied to various subject categories to obtain the first average reading time information.

[0063] Such an arrangement improves the accuracy of the first average reading time information obtained.

[0064] S500, according to the subject category preference information and the first average reading time information, obtain the recommendation value of all library resources, and obtain the user's recommendation list according to the recommendation value of all library resources. The recommendation value is used to reflect the quantitative value of the recommendation degree of a library resource.

[0065] It can be understood that the subject category preference information reflects the user's subject category preference. After the LDA topic model determines the subject category distribution of all library resources, the subject category distribution of library resources is compared with the user's subject category preference to obtain the subject category recommendation value of the library resources; the first average reading time information reflects the user's reading time preference. The text volume of library resources can be converted into reading time at a fixed ratio, and then the user's reading time preference and the reading time of library resources are compared to obtain the reading time recommendation value of the library resources. The subject category recommendation value and the reading time recommendation value can be directly added to obtain the recommended value of the library resources, or a coefficient can be added and then added to obtain the recommended value of the library resources. This can be extended to all library resources to obtain the recommended values ​​of all library resources, and then sorted from large to small according to the recommended values. The recommended list can be a fixed number list with the largest recommendation value, or a dynamically changing list with an increase in the number of library resources as user demand increases.

[0066] This setting improves the quality of the recommendation list.

[0067] In a possible implementation, in step S500, the recommendation values ​​of all library resources are obtained according to the subject category preference information and the first average reading time information, including: S510, obtaining a first recommendation value for each library resource according to the subject category preference information, wherein the first recommendation value is used to reflect the subject category recommendation degree of a library resource.

[0068] It can be understood that the subject category preference information is matched with the subject category of any selected library resource to obtain the first recommendation value. The matching can be performed by multiplying and then adding the distribution coefficients of similar subject categories (ie, dot product).

[0069] Exemplarily, the user's subject category preference information is expressed as (education×a1, sports×a2, science×a3) (a1, a2, a3 are all positive numbers), and the subject category of the selected library resources is (education×b1, sports×b2, science×b3) (b1+b2+b3=1), then the first recommendation value = a1×b1+a2×b2+a2×b2, and the first recommendation value is at least equal to 0.

[0070] Such a configuration can improve the accuracy of the first recommended value.

[0071] In a possible implementation, in step S510, obtaining a first recommendation value for each library resource according to the subject category preference information includes: S511, quantify and normalize the preference degree of each topic category in the topic category preference information of the user to obtain the topic category preference value of each topic category, wherein the topic category preference value is used to reflect the normalized quantified value of the user's preference degree for a topic category.

[0072] It can be understood that the degree of preference for each subject category in the subject category preference information of the user is expressed in terms of reading time. Therefore, the subject category preference degrees are quantified to unify the units of the subject category preference degrees, and then normalized to remove the units, retaining only the proportional relationship between the subject category preferences, thereby obtaining the subject category preference values ​​of each subject category.

[0073] For example, the user's subject category preference information is (education×1.25 days, sports×50h, transportation×20h), then unified quantization is first performed = (education×30h, sports×50h, transportation×20h), and then normalization is performed = (education×0.3, sports×0.5, transportation×0.2), then 0.3, 0.5, and 0.2 are the subject category preference values ​​of each subject.

[0074] This setting makes the topic category preference information more standardized and accurate.

[0075] S512, randomly select a library resource, and determine whether the library resource intelligent recommendation device has recorded the subject category distribution of the selected library resource. If not, perform LDA subject analysis on the selected library resource to obtain the subject category distribution of the selected library resource.

[0076] It can be understood that after arbitrarily selecting a library resource, it is first determined whether the subject category distribution of the selected library resource is recorded in the server of the library resource intelligent recommendation device. If the subject category distribution of the selected library resource is not recorded, the selected library resource can be input into the trained LDA topic model to obtain the subject category distribution of the selected library resource. The classification level of the subject category can be determined according to the recommendation level set by the user, and the subject category distribution of the selected library resource can be recorded in the server.

[0077] With this setting, the subject category of an arbitrarily selected library resource can be quickly obtained.

[0078] S513, obtaining the subject category preference value corresponding to the subject category distribution of the selected library resources.

[0079] It can be understood that after selecting a library resource and obtaining the subject category distribution of the selected library resource, the corresponding subject category preference value of the subject category distribution of the selected library resource is obtained. For example, if the subject category distribution of the selected library resource is (education×0.8, sports×0.2), the subject category preference value of the education subject and the subject category preference value of the sports subject are obtained.

[0080] With such settings, the corresponding topic category preference value can be quickly obtained.

[0081] S514, obtaining a first recommendation value of the selected library resource according to the subject category preference value corresponding to the subject category distribution of the selected library resource.

[0082] It can be understood that the subject category distribution of the selected library resources and the corresponding subject category preference values ​​are matched. The matching method can be to multiply and add the distribution coefficients of the same subject categories (i.e., dot product) to obtain the first recommendation value of the selected library resources. For example, the subject category distribution of the selected library resources is (education×0.8, sports×0.2), the subject category preference value of the education subject is 0.3, and the subject category preference value of the sports subject is 0.5, then the first recommendation value = 0.3×0.8+0.2×0.5=0.34. The maximum value of the first recommendation value is 1, and the minimum value is 0.

[0083] With such a setting, the range of the first recommended value is between 0 and 1, thereby improving the standardization of the first recommended value.

[0084] S520, based on the first average reading time information, obtain recommended text volume information, and then determine a second recommendation value for each library resource based on the recommended text volume information. The recommended text volume information is used to reflect the recommended volume of the library resource of each subject category of the user by the library resource intelligent recommendation device, and the second recommendation value is used to reflect the degree of recommendation of the volume of the library resource.

[0085] It can be understood that the first average reading time information includes the average reading time of users for various subject categories. The average reading time of various subject categories is converted into text volume according to a fixed ratio. After obtaining the recommended text volume of library resources in each subject category, a library resource is arbitrarily selected and the text volume and subject category of the library resource are obtained. The text volume of the library resource is compared with the recommended text volume of the library resources of the corresponding subject category to determine the second recommendation value of the arbitrarily selected library resource.

[0086] Such an arrangement improves the accuracy of the second recommended value.

[0087] In a possible implementation, in step S520, determining the second recommendation value of each library resource according to the recommended text quantity information includes: S521, obtaining the text volume and subject category distribution of an arbitrarily selected library resource.

[0088] It can be understood that after arbitrarily selecting a library resource, the text volume and subject category distribution of the selected library resource can be obtained through the server.

[0089] This setting is the first step to obtain the second recommended value.

[0090] S522, obtaining the amount of recommended text corresponding to the selected library resource according to the recommended text amount information and the subject category distribution of the selected library resource.

[0091] It can be understood that the recommended text volume information includes the recommended text volume of various subject categories. The subject category distribution of the selected library resources is multiplied by the recommended text volume of the corresponding subject category and then added (i.e., the dot product) to obtain the recommended text volume corresponding to the selected library resources. For example, the subject category distribution of the selected library resources is (education×0.8, sports×0.2), the recommended text volume for education themes is 3 million words, and the recommended text volume for sports themes is 5 million words. Then the recommended text volume corresponding to the selected library resources is 0.8×3 million+0.2×5 million=3.4 million words.

[0092] With such a setting, the amount of recommended text corresponding to the subject category of the selected library resources can be accurately obtained.

[0093] S523, obtaining a second recommendation value of the selected library resource according to the text volume of the selected library resource and the recommended text volume corresponding to the selected library resource.

[0094] It is understandable that the text volume of the selected library resource and the corresponding recommended text volume can be subtracted to obtain the difference in text volume, and the difference in text volume can be converted into the second recommended value according to a fixed ratio, for example, the difference in text volume can be converted into the second recommended value at a ratio of 10 million:1 (the difference in text volume exceeding 10 million is also regarded as 10 million), or a bell-shaped function (a function with a high middle and low sides) can be set, such as the density function of the normal distribution. , , the function shape is Figure 2 As shown, the horizontal axis is the difference in text volume, and the vertical axis is the second recommended value. At this time, the maximum value of the second recommended value is 1 and the minimum value is 0.

[0095] With such configuration, the second recommended value can be obtained in detail.

[0096] S530: Obtain recommendation values ​​of all library resources according to the first recommendation values ​​and the second recommendation values ​​of all library resources.

[0097] It can be understood that after determining the first recommended value and the second recommended value of any library resource, the first recommended value and the second recommended value can be added to obtain the recommended value of the library resource, or i1×first recommended value+i2×second recommended value=recommended value (i1, i2 are positive numbers) can be used to obtain the recommended value of the library resource, and analogously to all library resources to obtain the recommended values ​​of all library resources. With this setting, you can get detailed recommendation values ​​for all library resources.

[0098] S600: Detecting feedback information of the user on the recommendation list, wherein the feedback information is used to reflect the active feedback operation and passive feedback operation of the user on the recommendation list.

[0099] It is understandable that the user's feedback information on the recommendation list can be detected by the terminal device. The feedback information can be instructions entered by the user on the terminal device, or can be related settings set by the user about the recommendation list, such as the display mode of the recommendation list, etc.

[0100] Such a setting can improve the quality of recommendation results.

[0101] Optionally, the feedback information includes search keywords, search conditions, a list of uninterested users and a list of browsed users. The intelligent recommendation method for library resources based on natural language processing also includes: S610, detecting search information input by the user, wherein the search information includes search words and search conditions input by the user, the search words refer to words or sentences input by the user, and the search conditions refer to the limiting conditions of the search process selected by the user.

[0102] It can be understood that the search information input by the user is detected by the terminal device, and the search information includes search terms and search conditions. The search term can be a keyword, such as "black hole", or a natural language sentence, such as "What is a black hole?", "How are black holes formed?", and the search conditions can be a limitation on the author's name, a limitation on the writing date, a limitation on the subject category, etc., but are not limited to these.

[0103] With such a setting, the search information input by the user can be quickly obtained and the user's active feedback operation can be detected.

[0104] S620, obtaining a search keyword according to the search phrase, wherein the search keyword refers to a keyword of the search phrase.

[0105] It can be understood that you can use keyword extraction algorithms to extract search keywords, such as TF-IDF, TextRank and other keyword extraction algorithms, you can also use an external large language model to extract search keywords, you can also use the trained LDA topic model to extract the topic categories of the search terms, and use the topic categories as search keywords.

[0106] This setting can simplify the search terms and obtain the search keywords.

[0107] S630: When the user browses the recommendation list, detecting the browsing operation of the user, wherein the browsing operation refers to the operation input by the user when browsing the recommendation list.

[0108] It can be understood that the browsing operation of the user is detected by the terminal device, and the browsing operation may be a screen swiping operation, or an operation of adding to a not interested list, and so on.

[0109] With this arrangement, the user's passive feedback operation and active feedback operation are detected.

[0110] S640, obtaining a list of not interested input by the user according to the browsing operation, wherein the list of not interested is used to reflect the library resources that the user is not interested in.

[0111] It can be understood that the operations of no interest input by the user during the browsing operation are obtained, and the library resources of the operations of no interest input by the user can be included in the no-interest list.

[0112] With this configuration, the list of items of no interest input by the user can be quickly obtained.

[0113] S650: According to the browsing operation, the library resources browsed by the user in the recommended list are added to the browsed list, wherein the browsed list is used to reflect the recommended library resources that the user has browsed and paid attention to.

[0114] It can be understood that from the browsing operation, the library resources browsed by the user are marked, and the marked library resources are included in the browsed list. A time threshold can be added to the mark, and the library resources whose display time on the terminal device exceeds the time threshold are included in the browsed list.

[0115] With this setting, you can get the browsed list accurately.

[0116] S700: Update the recommendation list according to the feedback information.

[0117] It can be understood that updating the recommendation list based on feedback information can be that the feedback information changes the display method of the recommendation list for the user, then the number of names in the recommendation list is updated accordingly, or the user enters a "not interested" operation, then the uninterested library resources in the recommendation list are deleted, and the same number of library resources are added to the recommendation list, and the adding method is to take the library resources with the largest recommendation value that are not included in the recommendation list.

[0118] Exemplarily, the feedback information is that the user changes the display mode of the recommendation list, resulting in a screen displaying 2 library resources instead of 4 library resources, and correspondingly the number of lists in the recommendation list is reduced to 1 / 2 of the original.

[0119] This setting improves the quality of the recommendation list.

[0120] In a possible implementation, in step S700, updating the recommendation list according to the feedback information includes: S710, based on the search keywords in the feedback information, add search keywords to the recommendation list. The search keywords are used to add selection conditions to the recommendation list, and the selection conditions refer to the conditions that only library resources containing the search keywords can be selected into the recommendation list.

[0121] It can be understood that it is possible to detect whether the library resources contain the search keywords, mark all library resources that contain the search keywords, and obtain a recommendation list from the marked library resources. The recommendation list can be obtained in descending order of recommendation values, or in descending order of the number of search keywords in the library resources, or the recommendation list can be obtained based on both the recommendation value and the number of search keywords.

[0122] This configuration lists a feasible way to update the recommendation list based on feedback information.

[0123] S720, based on the search condition in the feedback information, add a search condition to the recommendation list, wherein the search condition is used to add a selection condition to the recommendation list.

[0124] It can be understood that the search conditions can be limitations on the author's name, the writing date, the subject category, etc. The search conditions are used as selection conditions for the recommendation list, and library resources that do not meet the selection conditions cannot be included in the recommendation list.

[0125] This configuration lists a feasible way to update the recommendation list based on feedback information.

[0126] S730: based on the uninterested list in the feedback information, delete the uninterested library resources in the recommended list, wherein the uninterested library resources refer to the library resources in the uninterested list.

[0127] It can be understood that after the user inputs the list of no interest, the same items in the recommended list and the list of no interest can be matched and found, and the same items in the recommended list can be deleted to update the recommended list.

[0128] This configuration lists a feasible way to update the recommendation list based on feedback information.

[0129] S740, based on the browsed list in the feedback information, reduce the recommendation value of the browsed library resources in the recommendation list by a first fixed value, and clear the browsed list to update the recommendation list. The browsed library resources refer to the library resources in the browsed list.

[0130] It can be understood that the recommendation value of the library resources browsed by the user is reduced by a first fixed value, the specific size of which depends on the specific situation, and the browsed list is cleared at the same time, so that when the user browses the same library resource in the recommendation list multiple times, the recommendation value of the library resource can be reduced multiple times.

[0131] Such a setting lists a feasible way to update the recommendation list based on feedback information, which is conducive to keeping the recommendation list fresh.

[0132] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0133] Corresponding to the intelligent recommendation method for library resources based on natural language processing described in the above embodiment, the embodiment of the present application also provides an intelligent recommendation system for library resources based on natural language processing, and each unit of the system can implement each step of the intelligent recommendation method for library resources based on natural language processing. Figure 3 A structural block diagram of a library resource intelligent recommendation system based on natural language processing provided in an embodiment of the present application is shown. For ease of explanation, only the parts related to the embodiment of the present application are shown.

[0134] Reference Figure 3 , the system comprises: An acquisition unit, used to acquire the user's historical reading information; wherein the historical reading information includes a historical reading list and reading time information, and the reading time information includes the user's reading time for each library resource in the historical reading list; A subject category judgment unit is used to perform LDA subject analysis on each library resource in the historical reading list to obtain the subject category distribution of each library resource in the historical reading list, and integrate the subject category distribution of each library resource to obtain a subject category list; wherein, the subject category list refers to a list recording the subject category distribution of each library resource in the historical reading list, and the subject category distribution of the library resources in the subject category list corresponds one-to-one with the library resources in the historical reading list; An analysis unit, used to obtain subject category preference information according to the subject category list and the reading time information; wherein the subject category preference information is used to reflect the user's preference for library resources of various subject categories; The analysis unit is further used to obtain first average reading time information according to the subject category list and the reading time information; wherein the first average reading time information refers to information recording the average reading time of users for library resources of various subject categories, so as to reflect the reading time preference of users for library resources of different subject categories; A recommendation unit, used to obtain the recommendation value of all library resources according to the subject category preference information and the first average reading time information, and obtain the user's recommendation list according to the recommendation value of all library resources; wherein the recommendation value is used to reflect the quantitative value of the recommendation degree of a library resource; A feedback unit, used to detect feedback information of the user on the recommendation list; wherein the feedback information is used to reflect the active feedback operation and passive feedback operation of the user on the recommendation list; An updating unit is used to update the recommendation list according to the feedback information.

[0135] It should be noted that the information interaction, execution process and other contents between the above-mentioned units are based on the same concept as the method embodiment of the present application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0136] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the division of the above-mentioned functional units is used as an example for illustration. In practical applications, the above-mentioned functions can be assigned to different functional units as needed, that is, the internal structure of the system can be divided into different functional units to complete all or part of the functions described above. The functional units in the embodiment can be integrated into a processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0137] The present application also provides a library resource intelligent recommendation device, which may include a server connected to a terminal device for communication. Figure 4 This is a schematic diagram of the structure of a server and a terminal device provided in an embodiment of the present application. Figure 4 As shown, the server 4 of the library resource intelligent recommendation device of this embodiment includes: at least one processor 40 ( Figure 4 Only one is shown), at least one memory 41 ( Figure 4 Only one is shown in the figure) and a computer program 42 stored in the at least one memory 41 and executable on the at least one processor 40. When the processor 40 executes the computer program 42, the server 4 of the library resource intelligent recommendation device implements the steps in any of the above-mentioned library resource intelligent recommendation method embodiments based on natural language processing, or the server 4 of the library resource intelligent recommendation device implements the functions of each unit in the above-mentioned system embodiments.

[0138] Exemplarily, the computer program 42 may be divided into one or more units, which are stored in the memory 41 and executed by the processor 40 to complete the present application. The one or more units may be a series of computer program instruction segments capable of completing specific functions, which are used to describe the execution process of the computer program 42 in the server 4 of the library resource intelligent recommendation device.

[0139] The server 4 of the library resource intelligent recommendation device can be a desktop computer, a computer, a cloud server, a large local server, etc. The server 4 of the library resource intelligent recommendation device can include, but is not limited to, a processor 40 and a memory 41. Those skilled in the art can understand that Figure 4It is only an example of the server 4 of the library resource intelligent recommendation device and does not constitute a limitation on the server 4 of the library resource intelligent recommendation device. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components, for example, it may also include input and output devices, network access devices, buses, etc.

[0140] The processor 40 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0141] In some embodiments, the memory 41 may be an internal storage unit of the server 4 of the library resource intelligent recommendation device, such as a hard disk or memory of the server 4 of the library resource intelligent recommendation device. In other embodiments, the memory 41 may also be an external storage device of the server 4 of the library resource intelligent recommendation device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the server 4 of the library resource intelligent recommendation device. Further, the memory 41 may also include both an internal storage unit and an external storage device of the server 4 of the library resource intelligent recommendation device. The memory 41 is used to store an operating system, an application program, a boot loader (BootLoader), data, and other programs, such as the program code of the computer program, etc. The memory 41 may also be used to temporarily store data that has been output or is to be output.

[0142] An embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in any of the above method embodiments are implemented.

[0143] An embodiment of the present application provides a computer program product. When the computer program product runs on a library resource intelligent recommendation device, the library resource intelligent recommendation device implements the steps in any of the above-mentioned method embodiments.

[0144] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium can at least include: any entity or device that can carry the computer program code to the library resource intelligent recommendation device, recording medium, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, RandomAccess Memory), electric carrier signal, telecommunication signal and software distribution medium. For example, a USB flash drive, a mobile hard disk, a disk or an optical disk.

[0145] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0146] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0147] In the embodiments provided in the present application, it should be understood that the disclosed library resource intelligent recommendation method based on natural language processing, the library resource intelligent recommendation system based on natural language processing, and the library resource intelligent recommendation device can be implemented in other ways. For example, the above-described library resource intelligent recommendation method based on natural language processing, the library resource intelligent recommendation system based on natural language processing, and the library resource intelligent recommendation device embodiments are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0148] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0149] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A library resource intelligent recommendation method based on natural language processing, characterized in that: The library resource intelligent recommendation method based on natural language processing is applied to a library resource intelligent recommendation device, and the method comprises: Acquire the user's historical reading information; wherein the historical reading information includes a historical reading list and reading time information, and the reading time information includes the user's reading time for each library resource in the historical reading list; Performing LDA topic analysis on each library resource in the historical reading list to obtain the subject category distribution of each library resource in the historical reading list, and integrating the subject category distribution of each library resource to obtain a subject category list; wherein the subject category list refers to a list recording the subject category distribution of each library resource in the historical reading list, and the subject category distribution of the library resources in the subject category list corresponds one-to-one to the library resources in the historical reading list; Obtaining subject category preference information according to the subject category list and the reading time information; wherein the subject category preference information is used to reflect the user's preference for library resources of various subject categories; Obtaining first average reading time information according to the subject category list and the reading time information; wherein the first average reading time information refers to information recording the average reading time of users for library resources of various subject categories, so as to reflect the reading time preference of users for library resources of different subject categories; According to the subject category preference information and the first average reading time information, the recommendation values ​​of all library resources are obtained, and according to the recommendation values ​​of all library resources, a user's recommendation list is obtained; wherein the recommendation value is used to reflect the quantitative value of the recommendation degree of a library resource; Detecting feedback information of the user on the recommendation list; wherein the feedback information is used to reflect the active feedback operation and passive feedback operation of the user on the recommendation list; The recommendation list is updated according to the feedback information.

2. The method for intelligent recommendation of library resources based on natural language processing according to claim 1, characterized in that: The method further comprises: Classifying each library resource in the historical reading list into multiple levels; wherein the multiple levels of classification refer to classifying the subject categories of the library resources according to the subdivision levels of different subject categories; Detecting the recommendation level set by the user; wherein the recommendation level is used to reflect the subdivision level of the subject category that the user accepts the recommendation; The subject category list is obtained according to the recommendation level.

3. The method for intelligent recommendation of library resources based on natural language processing as claimed in claim 1, characterized in that: The obtaining of the subject category preference information according to the subject category list and the reading time information includes: Calculate the distribution data of each subject category in the subject category list and the corresponding reading time in the reading time information to obtain the total reading time of the same category of various subject categories; wherein the total reading time of the same category refers to the sum of the equivalent reading time of users for the same subject category, and the equivalent reading time refers to the reading time of library resources distributed in equal proportion to different subject categories according to the subject category distribution; The total reading time of the same category of various subject categories is normalized to obtain the subject category preference information.

4. The method for intelligent recommendation of library resources based on natural language processing as claimed in claim 3, characterized in that: The obtaining first average reading time information according to the subject category list and the reading time information includes: According to the subject category list and the reading time information, the total reading time of the same category of various subject categories and the total amount of the same category of various subject categories are obtained; wherein the total amount of the same category refers to the sum of the same subject category distribution coefficients in all subject category distribution data in the subject category list; The first average reading time information is obtained according to the total reading time of the same category and the total amount of the same category of various subject categories.

5. The method for intelligent recommendation of library resources based on natural language processing as claimed in claim 1, characterized in that: The step of obtaining the recommendation values ​​of all library resources according to the subject category preference information and the first average reading time information includes: Obtaining a first recommendation value for each library resource according to the subject category preference information; wherein the first recommendation value is used to reflect the subject category recommendation degree of a library resource; According to the first average reading time information, the recommended text volume information is obtained, and then the second recommendation value of each library resource is determined according to the recommended text volume information; wherein the recommended text volume information is used to reflect the recommended volume of the library resource of each subject category of the user by the library resource intelligent recommendation device, and the second recommendation value is used to reflect the degree of recommendation of the text volume of a library resource; The recommendation value of all library resources is obtained according to the first recommendation value and the second recommendation value of all library resources.

6. The method for intelligent recommendation of library resources based on natural language processing as claimed in claim 5, characterized in that: The obtaining of a first recommendation value for each library resource according to the subject category preference information includes: Quantifying and normalizing the preference degree of each topic category in the topic category preference information of the user to obtain a topic category preference value of each topic category; wherein the topic category preference value is used to reflect the normalized quantified value of the user's preference degree for a topic category; Randomly select a library resource, determine whether the library resource intelligent recommendation device records the subject category distribution of the selected library resource, and if not, perform LDA subject analysis on the selected library resource to obtain the subject category distribution of the selected library resource; Obtaining the subject category preference value corresponding to the subject category distribution of the selected library resources; The first recommendation value of the selected library resource is obtained according to the subject category preference value corresponding to the subject category distribution of the selected library resource.

7. The method for intelligent recommendation of library resources based on natural language processing as claimed in claim 6, characterized in that: The determining the second recommendation value of each library resource according to the recommended text quantity information includes: Get the text volume and subject category distribution of an arbitrarily selected library resource; According to the recommended text quantity information and the subject category distribution of the selected library resources, obtaining the recommended text quantity corresponding to the selected library resources; The second recommendation value of the selected library resource is obtained according to the text volume of the selected library resource and the recommended text volume corresponding to the selected library resource.

8. The method for intelligent recommendation of library resources based on natural language processing as claimed in claim 1, characterized in that: The feedback information includes search keywords, search conditions, a list of people not interested and a list of people browsed, and the method further includes: Detecting search information input by a user; wherein the search information includes a search term and a search condition input by the user, wherein the search term refers to a word or sentence input by the user, and the search condition refers to a limiting condition of the search process selected by the user; Obtaining a search keyword according to the search phrase; wherein the search keyword refers to a keyword of the search phrase; When the user browses the recommendation list, detecting the user's browsing operation; wherein the browsing operation refers to the operation input by the user when browsing the recommendation list; Acquire a list of not interested input by the user according to the browsing operation; wherein the list of not interested is used to reflect the library resources that the user is not interested in; According to the browsing operation, the library resources browsed by the user in the recommended list are added to the browsed list; wherein the browsed list is used to reflect the recommended library resources that the user has browsed and paid attention to.

9. The method for intelligent recommendation of library resources based on natural language processing as claimed in claim 8, characterized in that: The updating of the recommendation list according to the feedback information includes: Based on the search keyword in the feedback information, the search keyword is added to the recommendation list; wherein the search keyword is used to add a selection condition to the recommendation list, and the selection condition refers to the condition that only library resources containing the search keyword can be selected into the recommendation list; Based on the search condition in the feedback information, adding the search condition to the recommendation list; wherein the search condition is used to add a selection condition to the recommendation list; Based on the uninterested list in the feedback information, deleting the uninterested library resources in the recommended list; wherein the uninterested library resources refer to the library resources in the uninterested list; Based on the browsed list in the feedback information, the recommendation value of the browsed library resources in the recommendation list is reduced by a first fixed value, and the browsed list is cleared to update the recommendation list; wherein the browsed library resources refer to the library resources in the browsed list.

10. A library resource recommendation system, comprising: An acquisition unit, used to acquire the user's historical reading information; wherein the historical reading information includes a historical reading list and reading time information, and the reading time information includes the user's reading time for each library resource in the historical reading list; A subject category judgment unit is used to perform LDA subject analysis on each library resource in the historical reading list to obtain the subject category distribution of each library resource in the historical reading list, and integrate the subject category distribution of each library resource to obtain a subject category list; wherein, the subject category list refers to a list recording the subject category distribution of each library resource in the historical reading list, and the subject category distribution of the library resources in the subject category list corresponds one-to-one with the library resources in the historical reading list; An analysis unit, used to obtain subject category preference information according to the subject category list and the reading time information; wherein the subject category preference information is used to reflect the user's preference for library resources of various subject categories; The analysis unit is further used to obtain first average reading time information according to the subject category list and the reading time information; wherein the first average reading time information refers to information recording the average reading time of users for library resources of various subject categories, so as to reflect the reading time preference of users for library resources of different subject categories; A recommendation unit, used to obtain the recommendation value of all library resources according to the subject category preference information and the first average reading time information, and obtain the user's recommendation list according to the recommendation value of all library resources; wherein the recommendation value is used to reflect the quantitative value of the recommendation degree of a library resource; A feedback unit, used to detect feedback information of the user on the recommendation list; wherein the feedback information is used to reflect the active feedback operation and passive feedback operation of the user on the recommendation list; An updating unit is used to update the recommendation list according to the feedback information.

Images