Book classification method and equipment

Abstract

The embodiment of the present application discloses a book grading method and device, which belongs to the field of data processing technology. The method includes: obtaining book features of a target book, the book features include multiple basic features and multiple complex features of the target book. Obtaining a book grade classification model, the book grade classification model includes multiple book grades and book features corresponding to the multiple book grades, the multiple book grades and the book features corresponding to the multiple book grades are obtained by clustering the book features of each sample book in multiple sample books. Based on the book features of the target book and the book features corresponding to each book grade in the multiple book grades, determine the correlation between the target book and each book grade in the multiple book grades, and use the book grade with the largest correlation as the book grade of the target book. The embodiment of the present application breaks through the concept of artificially dividing book grades at a subjective level, making book grading more flexible.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of data processing technology, and in particular to a book grading method and device. Background Art

[0002] The reading abilities of students at different stages of growth are different, and the books that students read need to meet their current reading abilities. Therefore, books need to be graded to obtain multiple book grades so that students can choose books that meet their reading abilities at the corresponding stage from the books corresponding to these multiple book grades.

[0003] The existing book classification system usually relies on expert experience, and experts manually set classification rules to classify books, resulting in the concept of book classification remaining at the subjective analysis level and the classification lacking flexibility. Summary of the invention

[0004] The embodiment of the present application provides a book grading method and device, which can solve the problem of lack of flexibility in grading in related technologies. The technical solution is as follows:

[0005] On the one hand, a book grading method is provided, the method comprising:

[0006] Acquire book features of a target book, wherein the book features include a plurality of basic features and a plurality of complex features of the target book, wherein the basic features are features obtained by statistically analyzing the quantity information related to the words, phrases, and illustrations included in the target book, and the complex features are features obtained by analyzing the semantic information related to the words, phrases, and illustrations included in the target book;

[0007] Obtaining a book grade classification model, the book grade classification model comprising a plurality of book grades and book features respectively corresponding to the plurality of book grades, the book features corresponding to each of the plurality of book grades comprising a plurality of basic features and a plurality of complex features, the plurality of book grades and the book features respectively corresponding to the plurality of book grades being obtained by clustering the book features of each sample book in a plurality of sample books;

[0008] Based on the basic features and complex features in the book features of the target book and the basic features and complex features in the book features corresponding to each book level in the multiple book levels, the correlation between the target book and each book level in the multiple book levels is determined, and the book level with the greatest correlation is used as the book level of the target book.

[0009] Optionally, the obtaining of the book features of the target book includes:

[0010] Acquire text information and illustration information of the target book;

[0011] Based on the text information and illustration information of the target book, a plurality of basic features of the target book are determined, wherein the plurality of basic features include at least one or more of the total number of pages, the total number of words, the total number of terms, the number of terms per page, the number of first-frequency terms and the ratio of first-frequency terms, the number of second-frequency terms and the ratio of second-frequency terms, the total number of illustrations, and the average length of sentences, wherein the first-frequency terms are terms that appear in the target book at a frequency exceeding a first reference frequency, the second-frequency terms are terms that appear in the target book at a frequency lower than a second reference frequency, and the first reference frequency is greater than the second reference frequency;

[0012] Based on the text information and illustration information of the target book, and a semantic analysis model, multiple complex features of the target book are determined, and the multiple complex features include at least one or more of entity information, attribute information of entities, relationship information between entities, sentence complexity information, sentence pattern information, sentence thought depth information, and semantic information in illustrations.

[0013] Optionally, determining the degree of association between the target book and each of the multiple book grades based on the basic features and complex features in the book features of the target book and the basic features and complex features in the book features corresponding to each of the multiple book grades includes:

[0014] Preprocessing the book features of the target book and the book features corresponding to each of the multiple book grades to determine the standardized book features of the target book and the standardized book features corresponding to each of the multiple book grades;

[0015] Based on the standardized book features of the target book and the standardized book features corresponding to each book level in the multiple book levels, the Euclidean distance between the target book and each book level in the multiple book levels is determined, and the Euclidean distance is used as the association between the target book and the corresponding book level.

[0016] Optionally, preprocessing the book features of the target book and the book features corresponding to each book level in the multiple book levels to determine the standardized book features of the target book and the standardized book features corresponding to each book level in the multiple book levels includes:

[0017] The standardized book feature of the target book and the standardized book feature corresponding to each book level in the multiple book levels are determined by the following formula:

[0018]

[0019] Wherein, m indicates the number of the target books and the sum of the number of levels of the multiple book levels, i indicates the target books and the i-th element in the multiple book levels, the element is the target book or the book level, Indicates feature j in the book features of the i-th element, the Indicates the minimum value of feature j among the target book and the multiple book levels, Indicates the maximum value of feature j among the target book and the multiple book levels.

[0020] Optionally, the method further comprises:

[0021] Obtaining book features of each sample book among a plurality of sample books, wherein the book features of each sample book include a plurality of basic features and a plurality of complex features of the corresponding sample book;

[0022] Selecting K sample books from the plurality of sample books, and using the K sample books as K cluster centers, where K is a positive integer;

[0023] For any sample book among the multiple sample books except the cluster center, based on the book features of the sample book and the book features of each of the K cluster centers, determine the association degree between the sample book and each of the K cluster centers, classify the sample book into the book set corresponding to the cluster center with the largest association degree, and obtain the book set corresponding to each of the K cluster centers;

[0024] For a book set corresponding to any cluster center among the K cluster centers, based on the book features of the sample books in the book set, the cluster center of the book set is updated to obtain the updated K cluster centers;

[0025] Determine the correlation between the K cluster centers after the update and the K cluster centers before the update. If the correlation is lower than the reference correlation, then based on the K cluster centers after the update, return to execute for any sample book among the multiple sample books except the cluster center, based on the book features of the sample book and the book features of each cluster center among the K cluster centers, determine the correlation between the sample book and each cluster center among the K cluster centers, until the correlation between the K cluster centers after the most recent update and the K cluster centers updated before the most recent update exceeds the reference correlation, then determine K book grades based on the book features of each cluster center among the K cluster centers after the most recent update, wherein the book features corresponding to each book grade in the K book grades are the book features of one cluster center.

[0026] Optionally, determining K book levels based on the book features of each of the K cluster centers after the most recent update includes:

[0027] Sorting the K cluster centers according to the characteristic values ​​of one or more complex characteristics in the book characteristics of each cluster center in the K cluster centers after the most recent update, wherein the characteristic values ​​of the one or more complex characteristics of the cluster centers with higher rankings are greater than the characteristic values ​​of the one or more complex characteristics of the cluster centers with lower rankings;

[0028] Assigning an initial book grade to each of the K cluster centers after sorting, and using the book features of each cluster center as the book features of the corresponding initial book grade, wherein the initial book grade corresponding to the cluster center with a higher sorting is greater than the initial book grade corresponding to the cluster center with a lower sorting;

[0029] Display K initial book levels and book features of each of the K initial book levels;

[0030] In response to the grade modification instruction, the book grade corresponding to any initial book grade among the K initial book grades is modified, and the modified K initial book grades are used as the K book grades.

[0031] Optionally, before selecting K sample books from the plurality of sample books, the method further includes:

[0032] In response to the grade configuration instruction, the number of book grades to be configured is obtained, and the number is used as K.

[0033] In another aspect, a computer device is provided, the computer device comprising a processor, the processor being configured to:

[0034] Acquire book features of a target book, wherein the book features include a plurality of basic features and a plurality of complex features of the target book, wherein the basic features are features obtained by statistically analyzing the quantity information related to the words, phrases, and illustrations included in the target book, and the complex features are features obtained by analyzing the semantic information related to the words, phrases, and illustrations included in the target book;

[0035] Obtaining a book grade classification model, the book grade classification model comprising a plurality of book grades and book features respectively corresponding to the plurality of book grades, the book features corresponding to each of the plurality of book grades comprising a plurality of basic features and a plurality of complex features, the plurality of book grades and the book features respectively corresponding to the plurality of book grades being obtained by clustering the book features of each sample book in a plurality of sample books;

[0036] Based on the basic features and complex features in the book features of the target book and the basic features and complex features in the book features corresponding to each book level in the multiple book levels, the correlation between the target book and each book level in the multiple book levels is determined, and the book level with the greatest correlation is used as the book level of the target book.

[0037] Optionally, the processor is configured to:

[0038] Acquire text information and illustration information of the target book;

[0039] Based on the text information and illustration information of the target book, a plurality of basic features of the target book are determined, wherein the plurality of basic features include at least one or more of the total number of pages, the total number of words, the total number of terms, the number of terms per page, the number of first-frequency terms and the ratio of first-frequency terms, the number of second-frequency terms and the ratio of second-frequency terms, the total number of illustrations, and the average length of sentences, wherein the first-frequency terms are terms that appear in the target book at a frequency exceeding a first reference frequency, the second-frequency terms are terms that appear in the target book at a frequency lower than a second reference frequency, and the first reference frequency is greater than the second reference frequency;

[0040] Based on the text information and illustration information of the target book, and a semantic analysis model, multiple complex features of the target book are determined, and the multiple complex features include at least one or more of entity information, attribute information of entities, relationship information between entities, sentence complexity information, sentence pattern information, sentence thought depth information, and semantic information in illustrations.

[0041] Optionally, the processor is configured to:

[0042] Preprocessing the book features of the target book and the book features corresponding to each of the multiple book grades to determine the standardized book features of the target book and the standardized book features corresponding to each of the multiple book grades;

[0043] Based on the standardized book features of the target book and the standardized book features corresponding to each book level in the multiple book levels, the Euclidean distance between the target book and each book level in the multiple book levels is determined, and the Euclidean distance is used as the association between the target book and the corresponding book level.

[0044] Optionally, the processor is configured to:

[0045] The standardized book feature of the target book and the standardized book feature corresponding to each book level in the multiple book levels are determined by the following formula:

[0046]

[0047] Wherein, m indicates the number of the target books and the sum of the number of levels of the multiple book levels, i indicates the target books and the i-th element in the multiple book levels, the element is the target book or the book level, Indicates feature j in the book features of the i-th element, the Indicates the minimum value of feature j among the target book and the multiple book levels, Indicates the maximum value of feature j among the target book and the multiple book levels.

[0048] Optionally, the processor is configured to:

[0049] Obtaining book features of each sample book among a plurality of sample books, wherein the book features of each sample book include a plurality of basic features and a plurality of complex features of the corresponding sample book;

[0050] Selecting K sample books from the plurality of sample books, and using the K sample books as K cluster centers, where K is a positive integer;

[0051] For any sample book among the multiple sample books except the cluster center, based on the book features of the sample book and the book features of each of the K cluster centers, determine the association degree between the sample book and each of the K cluster centers, classify the sample book into the book set corresponding to the cluster center with the largest association degree, and obtain the book set corresponding to each of the K cluster centers;

[0052] For a book set corresponding to any cluster center among the K cluster centers, based on the book features of the sample books in the book set, the cluster center of the book set is updated to obtain the updated K cluster centers;

[0053] Determine the correlation between the K cluster centers after the update and the K cluster centers before the update. If the correlation is lower than the reference correlation, then based on the K cluster centers after the update, return to execute for any sample book among the multiple sample books except the cluster center, based on the book features of the sample book and the book features of each cluster center among the K cluster centers, determine the correlation between the sample book and each cluster center among the K cluster centers, until the correlation between the K cluster centers after the most recent update and the K cluster centers updated before the most recent update exceeds the reference correlation, then determine K book grades based on the book features of each cluster center among the K cluster centers after the most recent update, wherein the book features corresponding to each book grade in the K book grades are the book features of one cluster center.

[0054] Optionally, the processor is configured to:

[0055] Sorting the K cluster centers according to the characteristic values ​​of one or more complex characteristics in the book characteristics of each cluster center in the K cluster centers after the most recent update, wherein the characteristic values ​​of the one or more complex characteristics of the cluster centers with higher rankings are greater than the characteristic values ​​of the one or more complex characteristics of the cluster centers with lower rankings;

[0056] Assigning an initial book grade to each of the K cluster centers after sorting, and using the book features of each cluster center as the book features of the corresponding initial book grade, wherein the initial book grade corresponding to the cluster center with a higher sorting is greater than the initial book grade corresponding to the cluster center with a lower sorting;

[0057] Display K initial book levels and book features of each of the K initial book levels;

[0058] In response to the grade modification instruction, the book grade corresponding to any initial book grade among the K initial book grades is modified, and the modified K initial book grades are used as the K book grades.

[0059] Optionally, the processor is configured to:

[0060] In response to the grade configuration instruction, the number of book grades to be configured is obtained, and the number is used as K.

[0061] On the other hand, a computer-readable storage medium is provided, wherein a computer program is stored in the storage medium, and when the computer program is executed by a processor, the steps of the above-mentioned book grading method are implemented.

[0062] On the other hand, a computer program product comprising instructions is provided, which, when executed on a computer, enables the computer to execute the steps of the above-mentioned book rating method.

[0063] The technical solution provided by the embodiments of the present application can at least bring the following beneficial effects:

[0064] In an embodiment of the present application, a book feature and a book grade classification model of a target book are obtained, and the book feature includes multiple basic features and multiple complex features of the target book. Since the basic feature is a feature obtained by statistically analyzing the quantitative information related to the words, phrases, and illustrations included in the target book, and the complex feature is a feature obtained by analyzing the semantic information related to the words, phrases, and illustrations included in the target book, the features included in the multiple basic features and the multiple complex features are richer, so that the final divided book grade is more accurate. In addition, an embodiment of the present application obtains a book grade classification model, and the book grade classification model includes multiple book grades and book features corresponding to the multiple book grades, respectively. The multiple book grades and book features corresponding to the multiple book grades are obtained by clustering the book features of each sample book in multiple sample books. Therefore, when determining multiple book grades in the book grade classification model, the book grades can be flexibly divided based on multiple sample books based on actual needs. Furthermore, based on the rich book features in the target book and the book features corresponding to each book level in the multiple book levels, the correlation between the target book and each book level in the multiple book levels is determined, and the book level with the greatest correlation is used as the book level of the target book, thereby achieving the effect of correctly classifying the target book into the corresponding book level. The method of classifying book levels based on book features in the embodiment of the present application breaks through the concept of artificially classifying book levels at a subjective level, making book classification more flexible. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0066] Figure 1 is a schematic diagram of a system framework provided in an embodiment of the present application;

[0067] Figure 2 is a flow chart of a book grading method provided in an embodiment of the present application;

[0068] Figure 3 This is a flow chart of book feature extraction provided by an embodiment of the present application;

[0069] Figure 4 is a schematic diagram of a book grading effect provided in an embodiment of the present application;

[0070] Figure 5 is a flow chart for determining a book grade provided by an embodiment of the present application;

[0071] Figure 6 is a schematic diagram of the structure of a terminal provided in an embodiment of the present application;

[0072] Figure 7 It is a structural diagram of a server provided in an embodiment of the present application. DETAILED DESCRIPTION

[0073] In order to make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the implementation methods of the present application will be further described in detail below in conjunction with the accompanying drawings.

[0074] Before explaining in detail the book grading method provided in the embodiment of the present application, the application scenario and system architecture provided in the embodiment of the present application are first introduced.

[0075] The concept of graded reading of books originated in English-speaking developed countries in Europe and America. At present, the grading systems commonly used internationally include Lexile and DRA (Developmental Reading Assessment).

[0076] Lexile is a scientific and reliable book rating system certified and recognized by the National Institute of Human Health and Research. It measures the difficulty of books from four dimensions: vocabulary, syntax, semantics, and structure. It uses numbers plus the letter "L" as a measurement scale to measure the difficulty of books, that is, the Lexile index. The range of the Lexile index is 0L to 1700L. The smaller the number, the lower the difficulty of the book, and vice versa. Parents can choose suitable books for their students based on the Lexile index of the book.

[0077] The DRA grading system is a test standard used to measure students' reading ability. Teachers or parents can find books of appropriate level based on the test results to improve students' reading ability. The test content focuses on students' reading integration, reading fluency, comprehension and continuity / concentration, etc. The score range is 1-80. The higher the score, the stronger the student's reading ability. Parents can also choose books for their students based on their DRA test results.

[0078] However, since Chinese is richer than English in terms of sentence complexity, emotional complexity, and depth of thought, the existing English book rating system is difficult to directly apply to the rating of Chinese books.

[0079] At the end of the last century, Chinese graded reading was first put into practice in Taiwan and Hong Kong, China. Since the 21st century, Chinese graded reading has gradually entered the field of vision of the book publishing industry, education experts and scholars. The "2020 China K12 Student "Graded Reading" White Paper" provides new ideas for students to choose books. The white paper shows that 99% of parents will help students develop reading habits through parent-child reading, thereby improving their reading level. However, parents generally seem to be somewhat helpless in terms of how to choose books suitable for their students and how to improve their reading effects. In line with this trend, relevant departments have released the 2020 Chinese Primary School Students Graded Reading List to provide guidance for students to read books.

[0080] However, the way relevant departments recommend books to students is for experts from relevant departments to directly grade books for students of different grades based on their experience. The grade of the books is the grade of the students. The book grading lacks flexibility and there is no detailed basis for book classification. When faced with new books, parents cannot grade them and therefore cannot know whether the new books are suitable for students.

[0081] Based on the above problems, the embodiment of the present application proposes a book grading method. The book grading method provided by the embodiment of the present application does not rely on the existing book grading data when grading books, and intelligently divides the book grades. And the book grade of the book is determined based on the book features of the book. The embodiment of the present application can divide the book grades more finely or coarsely according to actual needs.

[0082] Please refer to Figure 1 , Figure 1 1 is a schematic diagram of a system architecture according to an exemplary embodiment. The system architecture 100 includes at least one terminal 101 ( Figure 1 ), a book rating classification server 102, a basic Chinese library 103, a preset model server 104, and a router 105.

[0083] The terminal 101 is connected to the router 105 for communication, and the book classification server 102 is connected to the basic Chinese library 103, the preset model server 104, and the router 105 for communication. The communication connection can be a wired or wireless connection, which is not limited in the present embodiment.

[0084] The terminal 101 is used for uploading the target book on the one hand, and for checking the ratings of the target book and other books on the other hand.

[0085] The book grade classification server 102 is used to obtain a book grade classification model on the one hand, and to determine the grade of a target book on the other hand.

[0086] In some embodiments, the electronic book data loading module, the basic feature statistics module, the complex feature detection module, and the book feature expression module are deployed in the book grade classification server 102. The book grade classification server 102 implements the grade classification of the target book based on these four modules. The detailed functions of these four modules will be described in subsequent embodiments and will not be elaborated here.

[0087] The basic Chinese database 103 includes information such as Chinese character database, word database, word frequency, and part of speech, etc. Thus, the book grade classification server 102 can call this information in the basic Chinese database 103 to extract basic features of the target book and other books.

[0088] The semantic analysis model is deployed in the preset model server 104. In this way, the book level classification server 102 can call the semantic analysis model in the preset model server 104 to extract the complex features of the target book and other books. The detailed functions of the semantic analysis model will be described in the subsequent embodiments and will not be elaborated here.

[0089] The router 105 is used to connect the terminal 101 and the book rating classification server 102 .

[0090] The terminal 101 may be any electronic product that can interact with a user through one or more methods such as a keyboard, a touch pad, a touch screen, a remote controller, voice interaction, or a handwriting device. For example, a PC (Personal Computer), a mobile phone, a smart phone, a PDA (Personal Digital Assistant), a Pocket PC (PPC), a tablet computer, a smart TV, etc. In the embodiment of the present application, the terminal 101 may be a smart learning terminal such as a tablet computer.

[0091] The book rating classification server 102 may be a single server, or a server cluster consisting of multiple servers, or a cloud computing service center.

[0092] Basic Chinese Library 103 is a platform for reading and sharing Chinese documents.

[0093] The preset model server 104 may be a single server, or a server cluster consisting of multiple servers, or a cloud computing service center.

[0094] Router 105 is an interconnection device in a computer network.

[0095] The models and modules involved in the above-mentioned server are all software modules, which can realize corresponding functions through the method provided in the embodiment of the present application.

[0096] In addition, those skilled in the art should understand that the above-mentioned terminal 101, book grade classification server 102, basic Chinese library 103, and preset model server 104 are only examples, and other existing or future terminals or basic Chinese libraries or servers, if applicable to the embodiments of the present application, should also be included in the protection scope of the embodiments of the present application and are included here by reference.

[0097] It should be noted that Figure 1 The system architecture shown is described by taking the terminal 101, the book grade classification server 102, the basic Chinese library 103, and the preset model server 104 as independent devices. Optionally, the method provided in the embodiment of the present application can also be applied to a centralized terminal device, which has the functions of the above-mentioned terminal 101 and the book grade classification server 102, the basic Chinese library 103, and the preset model server 104, and the embodiment of the present application does not limit this.

[0098] Next, the book grading method provided in the embodiment of the present application is explained in detail.

[0099] Figure 2 This is a flowchart of a book grading method provided in an embodiment of the present application, which is applied to a book grading server. Figure 2 , the method comprises the following steps.

[0100] Step 201: Obtain book features of the target book, which include multiple basic features and multiple complex features of the target book. The basic features are features obtained based on statistics of quantitative information related to words, phrases, and illustrations included in the target book, and the complex features are features obtained based on semantic information analysis related to words, phrases, and illustrations included in the target book.

[0101] The target book may be one book or multiple books, which is not limited in this implementation application example.

[0102] Before obtaining the book features of the target book, the book classification server first needs to obtain the target book. There are two ways to obtain the target book.

[0103] In a possible implementation, the terminal has an electronic version of a target book that the user wants to classify. At this time, the user can directly upload the electronic version of the target book to the book classification server through the terminal.

[0104] In another possible implementation, since the basic Chinese library stores the electronic versions of a large number of Chinese books, when the target book that the user wants to classify is a paper book, the terminal can send a search request to the basic Chinese library, and the search request carries information that can represent the target book. When the basic Chinese library receives the search request, it searches the basic Chinese library for the electronic version of the target book according to the information representing the target book carried in the search request, and sends the retrieved electronic version of the target book to the book grade classification server. Among them, the information that can represent the target book can be, for example, the title of the target book or the book number of the target book, and of course, it can also be other information.

[0105] After obtaining the target book, the book grade classification server further obtains the book features of the target book through step 201. In some embodiments, the implementation process of step 201 can be implemented through the following steps.

[0106] Step 1: Obtain the text information and illustration information of the target book.

[0107] Step 2: Based on the text information and illustration information of the target book, multiple basic features of the target book are determined. The multiple basic features include at least one or more of the total number of pages, the total number of single words, the total number of words, the number of words per page, the number of first-frequency words and the proportion of first-frequency words, the number of second-frequency words and the proportion of second-frequency words, the total number of illustrations and the average length of sentences. The first-frequency words are words that appear in the target book more frequently than a first reference frequency, the second-frequency words are words that appear in the target book less frequently than a second reference frequency, and the first reference frequency is greater than the second reference frequency.

[0108] Step three: Based on the text information and illustration information of the target book and the semantic analysis model, determine multiple complex features of the target book, where the multiple complex features include at least one or more of entity information, attribute information of entities, relationship information between entities, sentence complexity information, sentence pattern information, sentence thought depth information, and semantic information in illustrations.

[0109] Among them, Figure 3As shown, the book grade classification server includes four modules: an electronic book data loading module, a basic feature statistics module, a complex feature detection module, and a book feature expression module. Among them, the electronic book data loading module is used to obtain the text information and illustration information of the target book. The basic feature statistics module is used to determine multiple basic features of the target book. The complex feature detection module is used to determine multiple complex features of the target book. The book feature expression module is used to express the book feature data (multiple basic features and multiple complex features) obtained above. For example, the multiple basic features and multiple complex features of the target book are expressed using a book feature set.

[0110] In step 1, the text information of the target book is all the texts in the target book, and the illustration information is all the illustrations in the target book.

[0111] After obtaining all the text and all the illustrations of the target book, the book grade classification server can upload all the text and all the illustrations of the target book to the basic feature statistics module and the complex feature detection module respectively, so as to obtain multiple basic features and multiple complex features of the target book according to step two and step three respectively.

[0112] It should be noted that in step 2, when determining the first frequency words and the second frequency words in multiple basic features of the target book, the basic feature statistics module in the book grade classification server can call the Chinese character library, thesaurus and other information included in the basic Chinese library to obtain the first frequency words and the second frequency words in the target book, and then determine the final number and proportion of the first frequency words, as well as the number and proportion of the second frequency words based on the total number of words in the target book.

[0113] Among them, in step 2, the first frequency words are words that appear more frequently in the target book, also called high-frequency words. The frequency of appearance of the high-frequency words in the target book exceeds the first reference frequency. The first reference frequency indicates that the number of occurrences of the first frequency words accounts for a large proportion of the total number of words. The first reference frequency can be set in advance, such as 80%, and the embodiment of the present application is not limited to this.

[0114] The second frequency words are words that appear less frequently in the target book, also known as low-frequency words. The low-frequency words appear less frequently in the target book than the second reference frequency. The second reference frequency indicates that the number of occurrences of the second frequency words accounts for a smaller proportion of the total number of words. The second reference frequency can be set in advance, such as 20%, and the embodiment of the present application is not limited to this.

[0115] It should be noted that, in order to reflect the difference between high-frequency words and low-frequency words, the first reference frequency and the second reference frequency are different, and the first reference frequency must be greater than the second reference frequency.

[0116] The average sentence length is the average number of words included in all sentences in the target book.

[0117] After obtaining the text information and illustration information of the target book, the basic feature statistics module counts the total number of pages, the total number of single words, the total number of words, the number of words per page, the number of first-frequency words and the proportion of first-frequency words, the number of second-frequency words and the proportion of second-frequency words, the total number of illustrations and the average length of sentences of the target book based on all the text and all the illustrations of the target book to obtain the basic features of the target book.

[0118] For example, Figure 4 As shown, the total number of pages of the target book "The Crow and the Fox" is 2, the total number of words is 334, the total number of words is 12, the number of first-frequency words (that is, the number of high-frequency words) is 10, the number of second-frequency words (that is, the number of low-frequency words) is 2, the total number of illustrations is 2, and the average sentence length is 21, that is, the average number of words included in all sentences in "The Crow and the Fox" is 21. Among them, the number of words per page, the proportion of first-frequency words, and the proportion of second-frequency words are not shown.

[0119] It should be noted that, in step three, the complex feature detection module in the book grade classification server calls the semantic analysis model in the preset server to determine multiple complex features of the target book.

[0120] In step three, the semantic analysis model is a model for analyzing the semantic information of the target book, and the semantic analysis model includes an entity extraction model, an attribute extraction model, a relationship extraction model, a sentence complexity detection model, a sentence discrimination model, a sentiment discrimination model, and an image summary extraction model.

[0121] Among them, the entity extraction model is used to extract the entity information of the target book, the attribute extraction model is used to extract the attribute information of the entity, the relationship extraction model is used to extract the relationship information between entities, the sentence complexity detection model is used to extract the complexity information of the sentence, the sentence discrimination model is used to extract the sentence information, the sentiment discrimination model is used to extract the thought depth information of the sentence, and the image summary extraction model is used to extract the semantic information in the illustration.

[0122] Among them, the entity information of the target book is the relevant information of the entities appearing in the target book, and the relevant information may be, for example, the names of the entities appearing and / or the number of entities appearing. Among them, the entity refers to a specific thing. The attribute information of the entity is the attribute information of the entity appearing in the target book, and the attribute information may be, for example, the attributes of the entities appearing and / or the number of entity attributes appearing. The relationship information between entities is the relationship information of the entities appearing in the target book, and the relationship information may be, for example, the relationship between the entities appearing and / or the number of relationships between the entities appearing. The complexity information of the sentences is the complexity of the sentences in the target book. The sentence information is the form and quantity information of the sentences in the target book. The thought depth information of the sentences is the degree of emotion expressed by the sentences in the target book. The semantic information in the illustrations is the information expressed by the illustrations in the target book.

[0123] For example, the entity information in the target book "The Little Pony Crossing the River" is the pony and the river, the attribute information of the entity is that the color of the pony is black, and the relationship information between the entities is that the pony crossed the river.

[0124] like Figure 4 As shown, the entity information of the target book "The Crow and the Fox" is 16, which means that 16 entities appear in the target book. The attribute information of the entity is 7, which means that 7 attributes of the entity appear in the target book. The relationship information between entities is 15, which means that 15 relationships between entities appear in the target book. The sentence complexity is 1. The sentence information is: 12 declarative sentences, 0 judgment sentences, 3 interrogative sentences, and 0 elliptical sentences. Among them, the sentence thought depth information and the semantic information in the illustration are not shown.

[0125] Each model included in the above-mentioned semantic analysis model has been trained with small-scale samples in advance, so it can be used to extract entity information, entity attribute information, relationship information between entities, sentence complexity information, sentence structure information, sentence thought depth information and semantic information in illustrations in the target book.

[0126] Since the implementation process of training the above models is the same, any model can be trained according to the following method to obtain each trained model. Next, taking the sentence complexity detection model as an example, the process of training the sentence complexity detection model is explained.

[0127] The implementation process of the above-mentioned training sentence complexity detection model can be: obtaining multiple training samples and labels for each training sample in the multiple training samples, wherein each training sample is a book, and the label of each training sample is the sentence complexity of the corresponding book. The label of each training sample is to manually calibrate the complexity of all sentences in the training sample, and then take the average value of the complexity of all sentences to obtain the final sentence complexity of the training sample. That is, the label of each training sample is manually calibrated. Then, the multiple training samples and the labels for each training sample in the multiple training samples are input into the initialized neural network model, and the initialized neural network model can obtain the above-mentioned sentence complexity detection model after learning. After this training process, the sentence complexity detection model can be used to determine the sentence complexity information in the target book.

[0128] When training other models, you only need to replace the labels of the training samples with the feature information to be extracted by the corresponding model. For example, when training an entity extraction model, the labels of the training samples are the entity information in the corresponding book.

[0129] After obtaining a plurality of basic features and a plurality of complex features of the target book respectively, the feature data in the target book can be expressed according to the book feature expression module.

[0130] For example, the target book set X can be set to [x (1) ,x (2) ,…,x (m) ], where m represents the number of target books, and the book features of the target books are expressed as where x (i) represents the book features of the i-th target book, and n represents the number of features.

[0131] For example, if there are 5 target books, then the target book set is X = [x (1) ,x (2) ,x (3) ,x (4) ,x (5) ], if the number of book features is 10, then the book feature expression of the first target book is

[0132] Step 202: Obtain a book grade classification model, which includes multiple book grades and book features corresponding to the multiple book grades, respectively. The book features corresponding to each of the multiple book grades include multiple basic features and multiple complex features. The multiple book grades and the book features corresponding to the multiple book grades are obtained by clustering the book features of each sample book in the multiple sample books.

[0133] Among them, the multiple book grades in the book grade classification model and the book features corresponding to the multiple book grades are pre-trained using multiple sample books in the basic Chinese library.

[0134] The training process can be as follows: for multiple sample books, first determine the book features of each sample book, combine the book features of each sample book into a feature set, and then pre-process the book features of each sample book based on the feature set to obtain standardized book features of the sample books. Finally, train the sample books according to the K-means clustering method to obtain multiple book levels and book features corresponding to the multiple book levels.

[0135] In order to standardize the book features of the sample books in order according to the book features of each sample book in the preprocessing stage, the book features of each sample book can be combined into a feature set.

[0136] The implementation process of obtaining the book features of each sample book in the multiple sample books may refer to the above step 201, which will not be described in detail here.

[0137] The above implementation method of preprocessing the book features of each sample book can be: using min-max standardization to ensure that the new feature set is dimensionless on [0,1]. (i) Represents the standardized book features of the i-th sample book.

[0138] Among them, the standardized book characteristics of the sample books can be determined by the following formula:

[0139]

[0140] Where m indicates the total number of books in the sample book, i indicates the i-th element in the sample book, Indicates feature j in the book features of the i-th element, min indicates the minimum value of feature j in the sample books, and max indicates the maximum value of feature j in the sample books.

[0141] For example, if the book feature is the number of pages, the maximum number of pages in the m sample books is 50, and the minimum number is 10. If the number of pages in the first sample book is 20, then the standardized number of pages in the first sample book is:

[0142] After obtaining the standardized book features of the sample books, the process of determining multiple book grades and the book features corresponding to the multiple book grades can be divided into the following steps.

[0143] Step 1: Select K sample books from multiple sample books, and use the K sample books as K cluster centers, where K is a positive integer.

[0144] Step 2: For any sample book except the cluster center among the multiple sample books, determine the correlation between the sample book and each of the K cluster centers based on the book features of the sample book and the book features of each of the K cluster centers, and classify the sample book into the book set corresponding to the cluster center with the largest correlation, so as to obtain the book set corresponding to each of the K cluster centers.

[0145] Step 3: For the book set corresponding to any cluster center among the K cluster centers, based on the book features of the sample books in the book set, update the cluster center of the book set to obtain the updated K cluster centers.

[0146] Step 4: Determine the correlation between the updated K cluster centers and the K cluster centers before the update. If the correlation is lower than the reference correlation, then based on the updated K cluster centers, return to execute for any sample book except the cluster center among the multiple sample books, based on the book features of the sample book and the book features of each cluster center in the K cluster centers, determine the correlation between the sample book and each cluster center in the K cluster centers, until the correlation between the most recently updated K cluster centers and the K cluster centers updated before the most recently updated one exceeds the reference correlation, then based on the book features of each cluster center in the most recently updated K cluster centers, determine K book grades, wherein the book features corresponding to each book grade in the K book grades are the book features of a cluster center.

[0147] In addition, the staff at the book grade classification server can independently determine the number of book grades according to actual needs, thereby achieving a finer or coarser granularity of book grade division. In this scenario, in some embodiments, before selecting K sample books in step 1, the number of book grades to be configured can also be obtained in response to the grade configuration instruction, and the number of book grades to be configured is used as the K value. The grade configuration instruction indicates the number of book grades to be configured.

[0148] For example, before executing step 1, a selection option for the number of book grades will be generated on the book grade classification server interface. At this time, the staff at the book grade classification server can select the number of book grades based on this option. In this way, after detecting the staff's operation, the book grade classification server will use the number of book grades selected by the staff as the K value.

[0149] For example, 9 book levels can be set according to the student grade, or two book levels can be set for each grade, so that 18 book levels can be set. Of course, book levels can also be set according to other conditions, and the value of K is not limited in the embodiment of the present application.

[0150] After the K value is determined, K sample books can be selected from multiple sample books as described in step 1, and the K sample books are respectively used as K cluster centers. Each cluster center includes the standardized book features of the corresponding book.

[0151] After determining K cluster centers, according to the description in step 2, a book set corresponding to each of the K cluster centers is obtained.

[0152] Among them, the implementation process of determining the correlation between the sample book and each of the K cluster centers can be: determining the Euclidean distance between the sample book and each of the K cluster centers, and taking the Euclidean distance as the correlation between the sample book and the corresponding cluster center. The smaller the Euclidean distance, the greater the correlation between the sample book and the corresponding cluster center.

[0153] For example, if two sample images are selected as cluster centers, the standardized book feature of one cluster center is expressed as x' (1) =(0.25,0.5,0.1), the standardized book feature of the other cluster center is expressed as x' (2) =(0.5,0.5,0.4), the standardized book feature of a sample book among other sample books is expressed as x' (3) =(0.5,0.1,0.2). Then the sample book x' (3) With cluster center x' (1) , x' (2) The Euclidean distances between them are

[0154] Due to d 23 Less than d 13 , represents the sample book x' (3) With cluster center x' (2) The correlation is greater, then the sample book x' (3) Classified into cluster center x' (2) The corresponding book collection.

[0155] After obtaining the book set corresponding to each of the K cluster centers according to step 2 above, as described in step 3, the cluster center of the book set is updated to obtain the updated K cluster centers.

[0156] For example, since the sample book x' (3)The standardized book feature expression is x' (3) =(0.5,0.1,0.2), cluster center x' (2) The standardized book feature expression is x' (2) =(0.5,0.5,0.4), the sample book x' (3) Classified into cluster center x' (2) In the corresponding book collection, the sample book x' (3) and cluster center x' (2) Take the mean of the standardized book features to get the updated cluster center x' (2) The standardized book features are expressed as

[0157] After obtaining the updated cluster center, the final K book grades and the standardized book features corresponding to each book grade can be determined based on the description of step 4.

[0158] In order to make the final K cluster centers standardized book features more accurate, the K cluster centers will be iteratively updated, that is, repeating the above steps 2 and 3, so that the final updated K cluster centers standardized book features reach a stable state, and the K cluster centers standardized book features that reach a stable state are used as the final K cluster centers standardized book features. There are two ways to determine that the K cluster centers standardized book features reach a stable state.

[0159] In one possible implementation, by repeating steps 2 and 3 above, the change between the standardized book features of the K cluster centers finally updated and the standardized book features of the K cluster centers before the update is made as small as possible, until the change value reaches a preset threshold, that is, the iteration reaches a stable state.

[0160] Among them, the threshold value can be set in advance, that is, the reference correlation degree can be set in advance. For example, the threshold value can be set to 0.001. That is, if the change value between the updated K cluster center standardized book features and the K cluster center standardized book features before the update reaches 0.001, the K cluster center standardized book features obtained by the most recent update will be used as the final K cluster center standardized book features.

[0161] In another possible implementation, by repeating the above steps 2 and 3, after reaching a certain number of iterations, the final K cluster center standardized book features are obtained.

[0162] The number of iterations may also be set in advance. For example, the number of iterations may be set to 1000, that is, after repeatedly iterating and updating the K cluster centers 1000 times, the standardized book features of the K cluster centers obtained by the 1000th update are used as the standardized book features of the K cluster centers finally obtained.

[0163] After finally obtaining the standardized book features of K cluster centers, K book grades can be determined based on the standardized book features of each of the K cluster centers. There are two ways to determine the K book grades.

[0164] In a possible implementation, the K cluster centers can be sorted according to the characteristic values ​​of one or more complex characteristics in the book characteristics of each cluster center in the K cluster centers after the most recent update, wherein the characteristic values ​​of one or more complex characteristics of the cluster centers with higher ranking are greater than the characteristic values ​​of one or more complex characteristics of the cluster centers with lower ranking. An initial book grade is configured for each of the sorted K cluster centers, and the book characteristics of each cluster center are used as the book characteristics of the corresponding initial book grade, wherein the initial book grade corresponding to the cluster center with higher ranking is greater than the initial book grade corresponding to the cluster center with lower ranking. The K initial book grades and the book characteristics of each initial book grade in the K initial book grades are displayed. In response to a grade modification instruction, the book grade corresponding to any initial book grade in the K initial book grades is modified, and the modified K initial book grades are used as the K book grades.

[0165] When the K cluster centers are sorted according to complex features, one or more features in the complex features can be selected, and then the feature average value of each cluster center can be calculated respectively.

[0166] For example, if the sentence complexity information in the complex feature is selected, cluster center 1 includes 10 sample books, and the sentence complexity of these 10 sample books is expressed as x = (2, 3, 1, 4, 6, 8, 9, 5, 4, 2), then the feature average value of cluster center 1 is The sentence complexity of the 10 sample books included in cluster center 2 is expressed as x = (1, 4, 3, 7, 6, 9, 2, 2, 5, 2), then the feature average value of cluster center 2 is Therefore, based on the feature average of the complex feature, the order of sorting cluster centers 1 and 2 is: cluster center 1, cluster center 2.

[0167] After the sorting order of the cluster centers is obtained, the initial book levels are configured for the K cluster centers in turn according to the sorting order. For example, if the sorting order of cluster centers 1 and 2, the initial book levels of cluster centers 1 and 2 are: book level 2, book level 1, respectively.

[0168] After obtaining K initial book grades, the K initial book grades and the standardized book features of each of the K initial book grades can be displayed on the book grade classification server interface, and there is also a modification option next to each initial book grade, which is used by experts to modify the initial book grades. Among them, the experts at the book grade classification server can modify the initialized K book grades according to the standardized book features of each initial book grade to obtain the final K book grades.

[0169] In another possible implementation, the K cluster centers can be sorted according to the characteristic values ​​of one or more complex characteristics in the book characteristics of each cluster center in the K cluster centers after the most recent update, wherein the characteristic values ​​of one or more complex characteristics of the cluster centers with higher ranking are greater than the characteristic values ​​of one or more complex characteristics of the cluster centers with lower ranking. A book grade is respectively configured for the sorted K cluster centers, and the book characteristics of each cluster center are used as the book characteristics of the corresponding book grade, wherein the book grade corresponding to the cluster center with higher ranking is greater than the book grade corresponding to the cluster center with lower ranking.

[0170] That is to say, after obtaining K initial book levels and the book features of each of the K initial book levels in the above implementation method, the K initial book levels and the book features of each of the K initial book levels can also be directly used as the final K book levels and the book features of each of the K book levels.

[0171] Step 203: Based on the basic features and complex features in the book features of the target book and the basic features and complex features in the book features corresponding to each book level in the multiple book levels, determine the correlation between the target book and each book level in the multiple book levels, and use the book level with the greatest correlation as the book level of the target book.

[0172] That is to say, after obtaining the book features of the target book, based on the correlation between the book features of the target book and the book features of each book level in the K book levels, the book level with the greatest correlation can be used as the book level of the target book. There are two ways to determine the correlation between the book features of the target book and each book level in the K book levels.

[0173] In one possible implementation, Figure 5As shown, the book features of the target book and the book features corresponding to each book level in the multiple book levels are preprocessed to determine the standardized book features of the target book and the standardized book features corresponding to each book level in the multiple book levels. Based on the standardized book features of the target book and the standardized book features corresponding to each book level in the multiple book levels, the Euclidean distance between the target book and each book level in the multiple book levels is determined, and the Euclidean distance is used as the correlation between the target book and the corresponding book level.

[0174] Specifically, the description of preprocessing the target book and determining the Euclidean distance between the target book and each of the K book levels can refer to the description of the sample book in the above step 2, which will not be repeated here.

[0175] After the Euclidean distance between the target book and each book grade is obtained, the book grade corresponding to the minimum Euclidean distance is used as the book grade of the target book.

[0176] In another possible implementation, after preprocessing the target book, the similarity between the target book and any book level can be calculated by using a cosine similarity calculation method, and the similarity can be used as the correlation between the target book and the corresponding book level.

[0177] Among them, the formula for calculating cosine similarity is as follows:

[0178]

[0179] Among them, X i represents the standardized book feature value of the target book, Y i represents the standardized book feature value of each book level in K book levels, and n represents the number of book features.

[0180] After obtaining the similarity between the target book and each book level, the book level corresponding to the maximum similarity value is used as the book level of the target book.

[0181] In addition, after obtaining the book grade of the target book, the book grade classification server Figure 4As shown, the book grade of the target book and the standardized book features of the target book can be sent to the terminal. After receiving the book grade of the target book and the standardized book features of the target book, the terminal generates a book grade evaluation interface based on the book grade of the target book and the standardized book features of the target book, and displays the book grade evaluation interface on the terminal. The book grade evaluation interface displays the standardized book features of the target book and the book grade of the target book. In this way, parents can understand the book grade of the book in real time, and determine whether the target book is suitable for students to read based on the book grade of the target book.

[0182] Optionally, the book grade evaluation interface can also display the student's reading history, which includes information such as the books the student has read and how many books are included in each grade.

[0183] In the embodiment of the present application, without relying on the existing book grade data, the core features of the book are extracted by extracting multiple basic features and multiple complex features of the target book. In addition, the value of the book grade K depends on the actual needs, and the intelligent division of the book grade is realized. The personnel at the book grade classification server can divide the book grade in a more fine-grained or coarse-grained manner. And based on the correlation between the target book and each book grade, the book grade of the target book is accurately determined. In addition, after the book grade of the target book is obtained, the book grade of the target book will be displayed in real time on the terminal, so that parents can understand the book grade of the target book in time and determine whether the target book is suitable for students to read.

[0184] The embodiment of the present application provides a book grading device, which can be implemented by software, hardware or a combination of both to become part or all of a terminal device, and the terminal device can be Figure 6 The terminal device shown in the figure comprises: a first acquisition module, a second acquisition module and a first determination module.

[0185] A first acquisition module is used to acquire book features of a target book, wherein the book features include a plurality of basic features and a plurality of complex features of the target book, wherein the basic features are features obtained by statistically analyzing the quantity information related to the words, phrases, and illustrations included in the target book, and the complex features are features obtained by analyzing the semantic information related to the words, phrases, and illustrations included in the target book;

[0186] A second acquisition module is used to acquire a book grade classification model, wherein the book grade classification model includes multiple book grades and book features corresponding to the multiple book grades, wherein the book features corresponding to each of the multiple book grades include multiple basic features and multiple complex features, and the multiple book grades and the book features corresponding to the multiple book grades are obtained by clustering the book features of each sample book in the multiple sample books;

[0187] The first determination module is used to determine the correlation between the target book and each book level in multiple book levels based on the basic features and complex features in the book features of the target book and the basic features and complex features in the book features corresponding to each book level in multiple book levels, and to take the book level with the greatest correlation as the book level of the target book.

[0188] Optionally, the first acquisition module is specifically used to:

[0189] Obtain text information and illustration information of the target book;

[0190] Based on the text information and illustration information of the target book, multiple basic features of the target book are determined, the multiple basic features at least including one or more of the total number of pages, the total number of single words, the total number of words, the number of words per page, the number of first-frequency words and the ratio of first-frequency words, the number of second-frequency words and the ratio of second-frequency words, the total number of illustrations, and the average length of sentences, the first-frequency words are words that appear in the target book with a frequency exceeding a first reference frequency, the second-frequency words are words that appear in the target book with a frequency lower than a second reference frequency, and the first reference frequency is greater than the second reference frequency;

[0191] Based on the text information and illustration information of the target book, and the semantic analysis model, multiple complex features of the target book are determined. The multiple complex features include at least one or more of entity information, attribute information of entities, relationship information between entities, sentence complexity information, sentence pattern information, sentence thought depth information, and semantic information in illustrations.

[0192] Optionally, the first determining module includes:

[0193] A first determining unit is used to pre-process the book features of the target book and the book features corresponding to each book level in the multiple book levels, and determine the standardized book features of the target book and the standardized book features corresponding to each book level in the multiple book levels;

[0194] The second determination unit is used to determine the Euclidean distance between the target book and each book level in the multiple book levels based on the standardized book features of the target book and the standardized book features corresponding to each book level in the multiple book levels, and use the Euclidean distance as the correlation between the target book and the corresponding book level.

[0195] Optionally, the first determining unit is specifically configured to:

[0196] The standardized book features of the target book and the standardized book features corresponding to each book level in the multiple book levels are determined by the following formula:

[0197]

[0198] Wherein, m indicates the number of target books and the sum of the number of levels of multiple book levels, i indicates the target book and the i-th element in multiple book levels, and the element is the target book or the book level. Indicates feature j in the book features of the i-th element, Indicates the minimum value of feature j among the target book and multiple book levels, Indicates the maximum value of feature j among the target book and multiple book levels.

[0199] Optionally, the device further comprises:

[0200] A third acquisition module is used to acquire the book features of each sample book among the multiple sample books, where the book features of each sample book include multiple basic features and multiple complex features of the corresponding sample book;

[0201] A selection module is used to select K sample books from a plurality of sample books, and use the K sample books as K cluster centers, where K is a positive integer;

[0202] A classification module is used for determining the association degree between the sample book and each of the K cluster centers for any sample book except the cluster center among the multiple sample books based on the book features of the sample book and the book features of each of the K cluster centers, classifying the sample book into the book set corresponding to the cluster center with the largest association degree, and obtaining the book set corresponding to each of the K cluster centers;

[0203] An updating module, for updating the clustering center of the book set corresponding to any clustering center among the K clustering centers based on the book features of the sample books in the book set, to obtain the updated K clustering centers;

[0204] The second determination module is used to determine the correlation between the K cluster centers after the update and the K cluster centers before the update. If the correlation is lower than the reference correlation, based on the K cluster centers after the update, return to execute for any sample book except the cluster center among the multiple sample books, based on the book features of the sample book and the book features of each cluster center in the K cluster centers, determine the correlation between the sample book and each cluster center in the K cluster centers, until the correlation between the K cluster centers after the most recent update and the K cluster centers updated before the most recent update exceeds the reference correlation, then based on the book features of each cluster center in the K cluster centers after the most recent update, determine K book grades, wherein the book features corresponding to each book grade in the K book grades are the book features of a cluster center.

[0205] Optionally, the second determining module is specifically configured to:

[0206] According to the characteristic value of one or more complex features in the book features of each of the K cluster centers after the most recent update, the K cluster centers are sorted, wherein the characteristic value of one or more complex features of the cluster center with a higher ranking is greater than the characteristic value of one or more complex features of the cluster center with a lower ranking;

[0207] Assign an initial book level to each of the K cluster centers after sorting, and use the book features of each cluster center as the book features of the corresponding initial book level, where the initial book level corresponding to the cluster center with a higher sorting level is greater than the initial book level corresponding to the cluster center with a lower sorting level;

[0208] Display K initial book levels and book features of each of the K initial book levels;

[0209] In response to the grade modification instruction, the book grade corresponding to any initial book grade among the K initial book grades is modified, and the modified K initial book grades are used as the K book grades.

[0210] Optionally, the device further comprises:

[0211] The fourth acquisition module is used to obtain the number of book grades to be configured in response to the grade configuration instruction, and use the number as K.

[0212] In the embodiment of the present application, without relying on the existing book grade data, the core features of the book are extracted by extracting multiple basic features and multiple complex features of the target book. In addition, the value of the book grade K depends on the actual needs, and the intelligent division of the book grade is realized. The personnel at the book grade classification server can divide the book grade in a more fine-grained or coarse-grained manner. And based on the correlation between the target book and each book grade, the book grade of the target book is accurately determined. In addition, after the book grade of the target book is obtained, the book grade of the target book will be displayed in real time on the terminal, so that parents can understand the book grade of the target book in time and determine whether the target book is suitable for students to read.

[0213] It should be noted that: the book grading device provided in the above embodiment only uses the division of the above functional modules as an example to illustrate when grading books. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the book grading device provided in the above embodiment and the book grading method embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0214] Figure 6 6 is a block diagram of a terminal 600 provided in an embodiment of the present application. Generally, the terminal 600 includes: a processor 601 and a memory 602 .

[0215] The processor 601 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 601 may be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). The processor 601 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 601 may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 601 may also include an AI (Artificial Intelligence) processor, which is used to process computing operations related to machine learning.

[0216] The memory 602 may include one or more computer-readable storage media, which may be non-transitory. The memory 602 may also include a high-speed random access memory, and a non-volatile memory, such as one or more disk storage devices, flash memory storage devices. In some embodiments, the non-transitory computer-readable storage medium in the memory 602 is used to store at least one instruction, which is used to be executed by the processor 601 to implement the book grading method provided in the method embodiment of the present application.

[0217] In some embodiments, the terminal 600 may further optionally include: a peripheral device interface 603 and at least one peripheral device. The processor 601, the memory 602 and the peripheral device interface 603 may be connected via a bus or a signal line. Each peripheral device may be connected to the peripheral device interface 603 via a bus, a signal line or a circuit board. Specifically, the peripheral device includes: at least one of a radio frequency circuit 604, a touch display screen 605, a camera 606, an audio circuit 607, a positioning component 608 and a power supply 609.

[0218] Those skilled in the art will understand that Figure 6 The structure shown in the figure does not constitute a limitation on the terminal 600, and the terminal 600 may include more or less components than those shown in the figure, or combine some components, or adopt a different component arrangement.

[0219] Figure 7 700 is a schematic diagram of the structure of a server provided in an embodiment of the present application. The server 700 includes a central processing unit (CPU) 701, a system memory 704 including a random access memory (RAM) 702 and a read-only memory (ROM) 703, and a system bus 705 connecting the system memory 704 and the central processing unit 701. The server 700 also includes a basic input / output system (I / O system) 706 that helps transmit information between various devices in the computer, and a large-capacity storage device 707 for storing an operating system 713, an application program 714, and other program modules 715.

[0220] The basic input / output system 706 includes a display 708 for displaying information and an input device 709 such as a mouse and a keyboard for user inputting information. The display 708 and the input device 709 are connected to the central processing unit 701 through an input / output controller 710 connected to the system bus 705. The basic input / output system 706 may also include an input / output controller 710 for receiving and processing inputs from a plurality of other devices such as a keyboard, a mouse, or an electronic stylus. Similarly, the input / output controller 710 also provides output to a display screen, a printer, or other types of output devices.

[0221] The mass storage device 707 is connected to the central processing unit 701 through a mass storage controller (not shown) connected to the system bus 705. The mass storage device 707 and its associated computer-readable media provide non-volatile storage for the server 700. That is, the mass storage device 707 may include a computer-readable medium (not shown) such as a hard disk or a CD-ROM drive.

[0222] Without loss of generality, computer readable media may include computer storage media and communication media. Computer storage media include volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information such as computer readable instructions, data structures, program modules or other data. Computer storage media include RAM, ROM, EPROM, EEPROM, flash memory or other solid-state storage technologies, CD-ROM, DVD or other optical storage, cassettes, magnetic tapes, disk storage or other magnetic storage devices. Of course, those skilled in the art will appreciate that computer storage media are not limited to the above. The above-mentioned system memory 704 and mass storage device 707 can be collectively referred to as memory.

[0223] According to various embodiments of the present application, the server 700 can also be connected to a remote computer on the network through a network such as the Internet. That is, the server 700 can be connected to the network 712 through the network interface unit 711 connected to the system bus 705, or the network interface unit 711 can be used to connect to other types of networks or remote computer systems (not shown).

[0224] The above-mentioned memory also includes one or more programs, and the one or more programs are stored in the memory and configured to be executed by the CPU.

[0225] In some embodiments, a computer-readable storage medium is also provided, in which a computer program is stored, and when the computer program is executed by a processor, the steps of the book grading method in the above embodiment are implemented. For example, the computer-readable storage medium can be a ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage device, etc.

[0226] It is worth noting that the computer-readable storage medium mentioned in the embodiments of the present application may be a non-volatile storage medium, in other words, a non-transitory storage medium.

[0227] It should be understood that all or part of the steps to implement the above embodiments can be implemented by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. The computer instructions can be stored in the above-mentioned computer readable storage medium.

[0228] That is, in some embodiments, a computer program product comprising instructions is also provided, which, when executed on a computer, enables the computer to execute the steps of the above-mentioned book grading method.

[0229] It should be understood that the "at least one" mentioned herein refers to one or more, and "multiple" refers to two or more. In the description of the embodiments of the present application, unless otherwise specified, " / " means or, for example, A / B can mean A or B; "and / or" in this article is only a description of the association relationship of the associated objects, indicating that there can be three relationships, for example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, in order to facilitate the clear description of the technical solutions of the embodiments of the present application, in the embodiments of the present application, the words "first", "second" and the like are used to distinguish the same items or similar items with basically the same functions and effects. Those skilled in the art can understand that the words "first", "second" and the like do not limit the quantity and execution order, and the words "first", "second" and the like do not limit them to be necessarily different.

[0230] The above-mentioned embodiments are provided for the present application and are not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A book grading method, characterized in that: The method comprises: Acquire text information and illustration information of a target book, wherein the target book is a book for which a user requests analysis of book grade; Based on the text information and illustration information of the target book, a plurality of basic features of the target book are determined, wherein the plurality of basic features include at least one or more of the total number of pages, the total number of words, the total number of terms, the number of terms per page, the number of first-frequency terms and the ratio of first-frequency terms, the number of second-frequency terms and the ratio of second-frequency terms, the total number of illustrations, and the average length of sentences, wherein the first-frequency terms are terms that appear in the target book at a frequency exceeding a first reference frequency, the second-frequency terms are terms that appear in the target book at a frequency lower than a second reference frequency, and the first reference frequency is greater than the second reference frequency; Based on the text information and illustration information of the target book, determining multiple complex features of the target book through a semantic analysis model, wherein the multiple complex features include at least one or more of entity information, attribute information of entities, relationship information between entities, sentence complexity information, sentence pattern information, sentence thought depth information, and semantic information in illustrations; Obtaining a book grade classification model, the book grade classification model comprising a plurality of book grades and book features respectively corresponding to the plurality of book grades, the book features corresponding to each of the plurality of book grades comprising a plurality of basic features and a plurality of complex features, the plurality of book grades and the book features respectively corresponding to the plurality of book grades are obtained by clustering the book features of each sample book in a plurality of sample books, and the plurality of book grades are determined after sorting the plurality of cluster centers according to the feature values ​​of one or more complex features in the book features of each cluster center in the plurality of cluster centers, wherein the feature values ​​of one or more complex features of the cluster center with a higher ranking are greater than the feature values ​​of one or more complex features of the cluster center with a lower ranking, and the book grade corresponding to the cluster center with a higher ranking is greater than the book grade corresponding to the cluster center with a lower ranking; The standardized book feature of the target book and the standardized book feature corresponding to each book level in the multiple book levels are determined by the following formula: Wherein, m indicates the number of the target books and the sum of the number of levels of the multiple book levels, i indicates the target books and the i-th element in the multiple book levels, the element is the target book or the book level, Indicates feature j in the book features of the i-th element, the Indicates the minimum value of feature j among the target book and the multiple book levels, Indicates the maximum value of feature j among the target book and the multiple book levels; Based on the standardized book features of the target book and the standardized book features corresponding to each book level in the multiple book levels, determine the Euclidean distance between the target book and each book level in the multiple book levels, use the Euclidean distance as the correlation between the target book and the corresponding book level, and use the book level with the largest correlation as the book level of the target book; The book grade of the target book and the standardized book features of the target book are sent to the terminal of the user, so that the terminal generates and displays a book grade evaluation interface based on the book grade of the target book and the standardized book features of the target book, and the book grade evaluation interface is used to display the book grade of the target book, the standardized book features of the target book and the student's reading history, so that the user can determine whether the target book is suitable for the student to read according to the book grade of the target book, and the reading history includes the number of books included in the multiple book grades among the books read by the student.

2. The method according to claim 1, characterized in that The method further comprises: Acquire a book feature of each sample book in the plurality of sample books, wherein the book feature of each sample book includes a plurality of basic features and a plurality of complex features of the corresponding sample book; Selecting K sample books from the plurality of sample books, and using the K sample books as K cluster centers, where K is a positive integer; For any sample book among the multiple sample books except the cluster center, based on the book features of the sample book and the book features of each of the K cluster centers, determine the association degree between the sample book and each of the K cluster centers, classify the sample book into the book set corresponding to the cluster center with the largest association degree, and obtain the book set corresponding to each of the K cluster centers; For a book set corresponding to any cluster center among the K cluster centers, based on the book features of the sample books in the book set, the cluster center of the book set is updated to obtain the updated K cluster centers; Determine the correlation between the K cluster centers after the update and the K cluster centers before the update. If the correlation is lower than the reference correlation, then based on the K cluster centers after the update, return to execute for any sample book among the multiple sample books except the cluster center, based on the book features of the sample book and the book features of each cluster center among the K cluster centers, determine the correlation between the sample book and each cluster center among the K cluster centers, until the correlation between the K cluster centers after the most recent update and the K cluster centers updated before the most recent update exceeds the reference correlation, then determine K book grades based on the book features of each cluster center among the K cluster centers after the most recent update, wherein the book features corresponding to each book grade in the K book grades are the book features of one cluster center.

3. The method according to claim 2, characterized in that The step of determining K book levels based on the book features of each of the K cluster centers after the most recent update includes: Sorting the K cluster centers according to the feature values ​​of one or more complex features in the book features of each cluster center in the K cluster centers after the most recent update; Assigning an initial book level to each of the K cluster centers after sorting, and using the book features of each cluster center as the book features of the corresponding initial book level; Display K initial book levels and book features of each of the K initial book levels; In response to the grade modification instruction, the book grade corresponding to any initial book grade among the K initial book grades is modified, and the modified K initial book grades are used as the K book grades.

4. The method according to claim 2, characterized in that Before selecting K sample books from the plurality of sample books, the method further includes: In response to the grade configuration instruction, the number of book grades to be configured is obtained, and the number is used as K.

5. A computer device, characterized in that: The computer device comprises a processor configured to: Acquire text information and illustration information of a target book, wherein the target book is a book for which a user requests analysis of book grade; Based on the text information and illustration information of the target book, a plurality of basic features of the target book are determined, wherein the plurality of basic features include at least one or more of the total number of pages, the total number of single words, the total number of words, the number of words per page, the number of first-frequency words and the ratio of first-frequency words, the number of second-frequency words and the ratio of second-frequency words, the total number of illustrations, and the average length of sentences, wherein the first-frequency words are words that appear in the target book at a frequency exceeding a first reference frequency, the second-frequency words are words that appear in the target book at a frequency lower than a second reference frequency, and the first reference frequency is greater than the second reference frequency; Based on the text information and illustration information of the target book, determining multiple complex features of the target book through a semantic analysis model, wherein the multiple complex features include at least one or more of entity information, attribute information of entities, relationship information between entities, sentence complexity information, sentence pattern information, sentence thought depth information, and semantic information in illustrations; Obtaining a book grade classification model, the book grade classification model comprising a plurality of book grades and book features respectively corresponding to the plurality of book grades, the book features corresponding to each of the plurality of book grades comprising a plurality of basic features and a plurality of complex features, the plurality of book grades and the book features respectively corresponding to the plurality of book grades are obtained by clustering the book features of each sample book in a plurality of sample books, and the plurality of book grades are determined after sorting the plurality of cluster centers according to the feature values ​​of one or more complex features in the book features of each cluster center in the plurality of cluster centers, wherein the feature values ​​of one or more complex features of the cluster center with a higher ranking are greater than the feature values ​​of one or more complex features of the cluster center with a lower ranking, and the book grade corresponding to the cluster center with a higher ranking is greater than the book grade corresponding to the cluster center with a lower ranking; The standardized book feature of the target book and the standardized book feature corresponding to each book level in the multiple book levels are determined by the following formula: Wherein, m indicates the number of the target books and the sum of the number of levels of the multiple book levels, i indicates the target books and the i-th element in the multiple book levels, the element is the target book or the book level, Indicates feature j in the book features of the i-th element, the Indicates the minimum value of feature j among the target book and the multiple book levels, Indicates the maximum value of feature j among the target book and the multiple book levels; Based on the standardized book features of the target book and the standardized book features corresponding to each book level in the multiple book levels, determine the Euclidean distance between the target book and each book level in the multiple book levels, use the Euclidean distance as the correlation between the target book and the corresponding book level, and use the book level with the largest correlation as the book level of the target book; The book grade of the target book and the standardized book features of the target book are sent to the terminal of the user, so that the terminal generates and displays a book grade evaluation interface based on the book grade of the target book and the standardized book features of the target book, and the book grade evaluation interface is used to display the book grade of the target book, the standardized book features of the target book and the student's reading history, so that the user can determine whether the target book is suitable for the student to read according to the book grade of the target book, and the reading history includes the number of books included in the multiple book grades among the books read by the student.

Images