A method and system for evaluating and recommending graded reading of extracurricular books for primary schools

By constructing the corpus and multiple linear regression model of extracurricular reading books in primary schools, and combining with Pearson's correlation coefficient to improve the collaborative filtering algorithm, the problems of accuracy and high computing resources consumption in the grading recommendation of extracurricular books in primary schools are solved, and the accuracy of reading ability between books and primary school students is achieved Match and efficient recommendation.

CN114969564BActive Publication Date: 2024-12-03MINZU UNIVERSITY OF CHINA
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210623160.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-06-02
Publication Date
2024-12-03
Estimated Expiration
2042-06-02

AI Technical Summary

Technical Problem

The existing graded reading evaluation methods for extracurricular books in primary schools lack quantitative research and cannot effectively match the reading ability of books and primary school students, resulting in low recommendation accuracy and high computing resources, especially in massive information environments, which are difficult to obtain resources that meet user needs.

Method used

The corpus of extracurricular reading books for primary schools was constructed, and 59 features were extracted from five levels: pictures, Chinese characters, vocabulary, sentences and chapters were extracted, and the hierarchical reading evaluation model was fitted using a multivariate linear regression model, and combined with Pearson's correlation coefficient, the collaborative filtering recommendation algorithm was improved. Calculate user similarity and predict the applicable school stages of the book.

Benefits of technology

It achieves accurate matching of extracurricular books in primary schools with applicable school stages, improves the accuracy and efficiency of the recommendation system, reduces the consumption of computing resources, and can provide efficient recommendation services in a massive information environment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114969564B_ABST
    Figure CN114969564B_ABST
Patent Text Reader

Abstract

The present invention discloses a method and system for evaluating and recommending graded reading of extracurricular books for primary schools; wherein the method comprises the following steps: selecting books applicable to three stages of primary school from extracurricular books, and constructing a corpus of extracurricular reading books for primary schools; starting from the three comprehensive features of book diversity, book complexity, and book cohesion, analyzing and judging to obtain multiple features at five levels of pictures, Chinese characters, vocabulary, sentences, and chapters as candidate feature sets for the graded reading evaluation method for primary school extracurricular books; using data analysis software to screen out the optimal comprehensive features to construct the graded reading evaluation method for primary school extracurricular books; applying the graded reading evaluation method for primary school extracurricular books to predict the applicable stages of books to the process of calculating user similarity by a collaborative filtering recommendation algorithm, using the Pearson correlation coefficient calculation method to calculate the similarity between users, and then calculating the predicted scores of users for unrated books. The present invention optimizes the performance of the extracurricular book recommendation system.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of natural language processing, and mainly relates to a method for evaluating and recommending graded reading of extracurricular books for primary schools and a system for recommending graded reading of extracurricular books for primary schools. Background Art

[0002] Reading ability is a necessary and important ability. It is in a basic and core position in human development. Especially for primary school students, reading can promote the comprehensive development of their intelligence, morality and aesthetic level. Faced with the complex ocean of books, how to recommend extracurricular books that match the reading ability of students in their current stage has become a difficult problem facing teachers and parents. Graded reading can provide ideas for solving this problem. The purpose of graded reading is to adapt the reading ability of readers in their current stage to the difficulty of the text. The graded reading assessment system for English, such as the Lexile reading framework and the AZ grading method, has been relatively well-established and popularized after long-term practice and development. Domestic research on graded reading started late, and there is a lack of a mature reading system. Most studies remain at the level of qualitative recommendations, lacking quantitative research, and cannot truly solve the problem of recommending extracurricular books for primary schools.

[0003] The main task of the graded reading assessment system is to predict the difficulty of the text, so as to recommend reading materials that match their reading ability to specific readers, thereby improving readers' reading ability. Text readability refers to the degree and nature of the ease of reading and understanding of the text. It is an important part of graded reading. Its application is mainly concentrated in the fields of mother tongue teaching and teaching Chinese as a foreign language. Research on the readability of primary and secondary school texts is concentrated at the level of Chinese textbooks, while research on the readability of extracurricular reading texts in primary and secondary schools needs to be explored. Therefore, it is of great significance to conduct research on the evaluation and recommendation of graded reading of extracurricular books in primary schools, recommend extracurricular reading materials with text difficulty that match the stage of study for primary school students, and cultivate and improve their reading ability.

[0004] The research on Chinese text readability is mainly aimed at teaching and in-class research. The scale of the corpus used to construct it is relatively limited, and its reliability and validity have yet to be verified. Secondly, the evaluation criteria for the difficulty of Chinese texts proposed by Chinese text readability are inconsistent, and the selected text features are too many and complex to be easily generalized. The actual application performance has yet to be verified.

[0005] In the scenario of graded recommendation of extracurricular books for primary school students, the traditional collaborative filtering recommendation algorithm needs to traverse the rating data of all books and then calculate the user similarity when making recommendations. The data calculated in this way is not only inaccurate, but also consumes a lot of computing power as the data increases. In addition, in the face of the ever-increasing geometric growth of information, how to recommend and obtain resources that meet user needs from massive information resources has become an issue that needs to be focused on, especially in the field of extracurricular book recommendations for primary school students. At this stage, we have no shortage of excellent works and readers who love reading. The key is how to recommend a good book to the most suitable students in their grades so that their reading ability can be substantially improved. Summary of the invention

[0006] The purpose of the present invention is to solve the problem of difficulty in matching primary school extracurricular reading books with applicable school stages and to optimize the performance of an extracurricular book recommendation system.

[0007] To achieve the above object, on the one hand, the present invention provides a method for evaluating and recommending graded reading of extracurricular books for primary schools, the method comprising the following steps:

[0008] We selected books suitable for three primary school stages from extracurricular books and constructed a corpus of primary school extracurricular reading books.

[0009] Starting from the three comprehensive characteristics of book diversity, book complexity, and book cohesion, we analyzed and determined the multiple characteristics of five levels, namely, pictures, Chinese characters, vocabulary, sentences, and chapters, as the candidate feature sets for the graded reading evaluation method of primary school extracurricular books.

[0010] Use data analysis software to select the best comprehensive features to construct a graded reading evaluation method for primary school extracurricular books;

[0011] The graded reading evaluation method for primary school extracurricular books is used to predict the applicable school stage of books and is applied to the collaborative filtering recommendation algorithm to calculate user similarity. The Pearson correlation coefficient calculation method is used to calculate the similarity between users, and then the user's predicted score for unrated books is calculated.

[0012] On the other hand, the present invention provides a system for graded reading, evaluation and recommendation of extracurricular books for primary schools, which system includes: user management, book management, registration and login, personal center and book recommendation function modules, and the book recommendation function module is used to execute the method for graded reading, evaluation and recommendation of extracurricular books for primary schools.

[0013] The present invention can solve the problem of mismatch between the text content of extracurricular books for primary schools and the applicable school stages, and can accurately solve the problem of matching between the recommendation of extracurricular books for primary schools and the applicable school stages. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1A schematic diagram of a process flow of a method for evaluating and recommending graded reading of extracurricular books for primary schools provided by an embodiment of the present invention;

[0015] Figure 2 Design a flow chart for the evaluation method of extracurricular books for primary schools;

[0016] Figure 3 This is a structural diagram of the primary school extracurricular reading book recommendation system based on text content. DETAILED DESCRIPTION

[0017] Figure 1 A flow chart of a method for evaluating and recommending extracurricular books for primary school students according to an embodiment of the present invention. Figure 1 As shown, the method includes steps S101-S104:

[0018] Step S101, selecting books suitable for three stages of primary school from extracurricular books, and constructing a corpus of primary school extracurricular reading books;

[0019] Specifically, the embodiment of the present invention uses the recommended bibliography for primary and secondary schools issued by the Basic Curriculum Development Center of the Ministry of Education and the recommended bibliography for primary and secondary schools and the national primary and secondary school libraries as the selection criteria for extracurricular books, selects books suitable for three stages of primary school from the extracurricular books, and constructs a corpus of extracurricular reading books for primary schools.

[0020] Existing research on the readability formula of Chinese mainly selects features from four different levels: Chinese character level features, vocabulary level features, sentence level features, and paragraph level features. It does not comprehensively consider whether there are correlations and commonalities among the features at each level, and does not propose features with stronger comprehensive generalization capabilities to construct a readability formula.

[0021] In addition, in the grading and evaluation of extracurricular reading books in primary schools, the characteristics of pictures cannot be ignored. Pictures are involved in extracurricular books in all stages of primary schools, especially in the lower stages of primary schools, where picture books account for a large proportion of extracurricular reading books.

[0022] Step S102, starting from the three comprehensive features of book diversity, book complexity, and book coherence, multiple features at five levels of pictures, Chinese characters, vocabulary, sentences, and chapters are analyzed and determined as candidate feature sets for the graded reading evaluation method of primary school extracurricular books.

[0023] The embodiment of the present invention starts from the three comprehensive characteristics of book diversity, book complexity and book coherence, and after research and analysis, selects 59 features as candidate feature sets at five levels, namely pictures, Chinese characters, vocabulary, sentences and chapters, and uses a multivariate linear regression model for fitting.

[0024] 1. On the diversity of books

[0025] The diversity characteristics of books are mainly composed of 18 characteristics such as picture diversity, Chinese character diversity, vocabulary diversity, sentence diversity and chapter diversity, which are fitted by a multiple linear regression model. Table 1.1 is the information table of the diversity characteristics of books.

[0026] Table 1.1 Book diversity characteristics information table

[0027]

[0028]

[0029]

[0030] 2. About the complexity of books

[0031] The complexity of books is mainly composed of 25 features such as the complexity of pictures, the complexity of Chinese characters, the complexity of vocabulary, the complexity of sentences and the complexity of chapters, which are fitted by a multivariate linear regression model. Table 2.2 is the table of book complexity feature information.

[0032] Table 2.2 Complexity feature information table of books

[0033]

[0034]

[0035]

[0036] III. About the coherence of books

[0037] The coherence of books is mainly composed of 16 features such as the coherence of pictures, the coherence of Chinese characters, the coherence of vocabulary, the coherence of sentences, and the coherence of chapters, which are fitted through a multiple linear regression model. Table 2.3 is a table of information on the coherence characteristics of books.

[0038] Table 2.3 Information table on the cohesiveness characteristics of books

[0039]

[0040]

[0041] The embodiment of the present invention uses the recommended bibliography for primary and secondary schools and the recommended bibliography for primary and secondary school libraries (rooms) nationwide as the selection criteria for extracurricular books, and the constructed primary school extracurricular reading book corpus is used to fit the primary school extracurricular reading book grading evaluation model more accurately and reliably. In order to effectively improve the accuracy and reliability and validity of the primary school extracurricular book grading evaluation model, the embodiment of the present invention starts from the three comprehensive features of book diversity, book complexity and book cohesion, and after research and analysis, selects 59 features as the candidate feature set at five levels such as pictures, Chinese characters, vocabulary, sentences and chapters, and uses a multivariate linear regression model to fit the primary school extracurricular book grading reading evaluation model.

[0042] Step S103, using data analysis software to select the best comprehensive features to construct a graded reading evaluation method for extracurricular books in primary schools;

[0043] Specifically, the embodiment of the present invention uses a multiple linear regression model to fit the primary school extracurricular reading book grading evaluation model for the three selected comprehensive features.

[0044] Specifically, Figure 2 As shown in the figure, the word segmentation tools in natural language processing and Python data processing programs are used to extract the text features of primary school extracurricular books, and then the 59 extracted text feature sets are screened, mainly by analyzing the "Solutions Statistical Package for the Social Sciences" (SPSS) software for factor analysis. Based on the idea of ​​dimensionality reduction, the correlation coefficient matrix between variables is explored, and the variables are grouped according to the size of the variable correlation. Many variables are aggregated into a few common factors, in order to reduce the difficulty of data collection and analysis. When performing factor analysis, the correlation coefficient of the variables is required to be greater than 0, and the KMO test and Bartiett's sphericity test are required to be satisfied.

[0045] Then, factor analysis was performed to obtain the correlation coefficient r between the text features and the applicable stage of study. The features were sorted according to their absolute values ​​and multiple linear diagnosis was performed using the variance expansion factor method. If the variance expansion factor (VIF) was less than 10, it was determined that the candidate features did not have collinearity problems. At the same time, the unique explanation of the dependent variable applicable stage of study (ΔR) after the feature was added to the candidate feature set was calculated. 2 ), if ΔR 2 If the percentage of feature extraction is greater than 2%, the feature will be retained; otherwise, the feature will be deleted from the candidate feature set.

[0046] Then, the regression model was fitted. The selected text features were used as independent variables, and the applicable school stage was used as the dependent variable to fit the linear regression model. The main evaluation indicators of the fitted model were the multiple correlation coefficient R, which reflects the linear correlation between all independent variables and the dependent variable. The larger the value, the closer the linear correlation; the determination coefficient R 2 The closer the R-squared is to 1, the better the model fits the data. 2 represents the adjusted coefficient of determination, which is one of the important standard indicators of good model fit.

[0047] Then the overall regression effect F test is performed on the model, which is mainly to test the overall significance of the linear regression equation, to determine whether the linear relationship between the explanatory variable and all explanatory variables is significant, and to determine whether it is appropriate to use a linear model to fit the relationship between these variables. Only when the significance of the model is <0.05 can it pass the significance test. If it does not pass the F test, it is necessary to consider using nonlinear regression to refit.

[0048] Then, the model regression coefficient T test is needed. In the constructed multiple regression model, it is necessary to consider whether each introduced variable has an impact and whether the impact is significant, and finally decide whether to retain the variable in the model. It is generally believed that if the significance value of the variable is less than 0.05, it means that the variable is statistically significant. If the regression coefficient T test fails, it is necessary to re-screen the text features and re-execute the above steps.

[0049] After completing the above steps, a multivariate linear regression model can be output. In order to test whether the model effect is significant, it is necessary to use the test data set to verify the constructed elementary school extracurricular reading book evaluation method and test the accuracy of its evaluation performance.

[0050] The multiple correlation coefficient R of the model was calculated by SPSS software, which reflects the linear correlation between all independent variables and the dependent variable. The coefficient of determination R square and adjusted R square were calculated, which are important indicators for measuring good model fit. Then the overall regression effect F test was performed on the model to test whether the linear relationship between the explanatory variable and all explanatory variables is significant, and to determine whether it is appropriate to use a linear model to fit the relationship between these variables. Finally, the regression coefficient T test was performed on the model to test the significance of each variable on the dependent variable, and the evaluation method of graded reading of extracurricular books in primary schools was obtained.

[0051] Y=αL 1 +βL 2 +γL 3 ++C

[0052] Where Y is the applicable stage of study, L 1 For book diversity, L 2 is the book complexity, L3 is the book coherence, C is the bias term, α is the coefficient of book diversity, β is the coefficient of book complexity, and γ is the coefficient of book coherence.

[0053] The above-mentioned multivariate linear regression model uses SPSS software to determine the optimal combination of multiple independent variables to jointly predict and evaluate the most suitable academic stage. Through a series of operations such as text feature extraction, text feature screening, factor analysis, regression model fitting, model F test and regression coefficient T test, the final method of graded reading evaluation of extracurricular books in primary schools is more effective and in line with actual conditions.

[0054] Step S104, applying the primary school extracurricular book graded reading evaluation method to predict the applicable school stage of the book to the collaborative filtering recommendation algorithm to calculate the user similarity, using the Pearson correlation coefficient calculation method to calculate the similarity between users, and then calculating the user's predicted score for the unrated book.

[0055] The traditional collaborative filtering recommendation algorithm is suitable for situations where the number of items is significantly smaller than the number of users. If the number of items is too large, the cost of calculating the similarity matrix is ​​very high. Therefore, the algorithm needs to be improved in the process of recommending extracurricular books in elementary schools.

[0056] The improved recommendation algorithm mainly applies the primary school extracurricular book graded reading evaluation method to predict the applicable school stage of books to the collaborative filtering recommendation algorithm to calculate the user similarity process, and uses the Pearson correlation coefficient calculation method to calculate the similarity between users, and then calculates the user's predicted score for unrated books. The calculation method is as follows:

[0057]

[0058]

[0059] Where i represents the i-th book; I u represents the set of books evaluated by user u in a certain academic period; I v represents the set of books rated by user v in a certain academic period; r u,i represents the rating of the i-th book by user u in a certain academic period; r v,i represents the rating of the i-th book by user v in a certain academic period; Indicates the average rating of books of user u in a certain academic period; represents the average rating of books of user v in a certain period of study. s(u,u') represents the similarity between user u and user u'. p u,i The predicted score of the user for the unrated book.

[0060] Figure 3 This is a structural diagram of the primary school extracurricular reading book recommendation system based on text content. Figure 3The server side of the system shown is based on the Django framework of the Python language, which has the advantages of completeness and versatility. The data method uses ORM technology to optimize database functions and improve the efficiency of interacting with the database, so that users can use the system more smoothly.

[0061] The text-based primary school extracurricular reading book recommendation system has five functional modules: user management, book management, registration and login, personal center, and book recommendation. The book recommendation module is used to implement the graded reading evaluation and recommendation method of primary school extracurricular books. Each functional module was designed in detail from the aspects of demand analysis, database design, front-end and back-end interaction, and relevant tests were carried out on each functional module of the system after the system was built.

[0062] By adopting the Django framework to design and implement a text-based primary school extracurricular reading book recommendation system, the system can make graded reading recommendations based on the stage characteristics of primary school students. By analyzing and studying the content level of primary school extracurricular reading texts, the characteristics of extracurricular reading materials in the same stage are extracted to make targeted recommendations. The extracurricular reading materials recommended for primary school students can reflect their reading interests and match their reading abilities. The system has good robustness and stability, complete functions, good interface interactivity, and conforms to user habits.

[0063] The embodiment of the present invention proposes a method for evaluating the graded reading of extracurricular books for primary schools. Starting from the three comprehensive features of book diversity, book complexity and book cohesion, 59 features are selected as candidate feature sets at five levels, including pictures, Chinese characters, vocabulary, sentences and chapters, after research and analysis, and a graded evaluation model for primary school extracurricular reading books is fitted. The influencing factors of various features at various levels on the graded evaluation of primary school extracurricular reading books are fully integrated, and finally the accurate matching of the text content of primary school extracurricular books and the applicable learning stage can be achieved.

[0064] In addition, the embodiment of the present invention proposes an improved collaborative filtering recommendation algorithm by integrating the graded reading evaluation method, and applies the applicable school stage predicted by the graded reading evaluation method for primary school extracurricular books to the collaborative filtering recommendation algorithm to achieve accurate matching.

Claims

1. A method for evaluating and recommending graded reading of extracurricular books for primary schools, characterized in that: The following steps are involved: We selected books suitable for three primary school stages from extracurricular books and constructed a corpus of primary school extracurricular reading books. Starting from the three comprehensive characteristics of book diversity, book complexity, and book cohesion, we analyzed and determined the multiple characteristics of five levels, namely, pictures, Chinese characters, vocabulary, sentences, and chapters, as the candidate feature set for the graded reading evaluation method of extracurricular books in primary schools. Use data analysis software to select the best comprehensive features to construct a graded reading evaluation method for primary school extracurricular books; The method of graded reading evaluation of primary school extracurricular books is used to predict the applicable school stage of books in the collaborative filtering recommendation algorithm to calculate the user similarity. The Pearson correlation coefficient calculation method is used to calculate the similarity between users, and then the predicted score of users for unrated books is calculated. The steps of using data analysis software to screen out the optimal comprehensive features to construct a graded reading evaluation method for elementary school extracurricular books include: The three comprehensive characteristics selected were used to fit the graded evaluation model of extracurricular reading books in primary schools by using a multiple linear regression model; the negative correlation coefficient R of the model was calculated by SPSS software, which reflects the linear correlation between all independent variables and the dependent variable, and the determination coefficient R square and adjusted R square were calculated; then the overall regression effect F test was performed on the model to test whether the linear relationship between the explanatory variable and all explanatory variables was significant, and to determine whether it was appropriate to use a linear model to fit the relationship between these variables; finally, the regression coefficient T test was performed on the model to test the significance of each variable to the dependent variable, and the graded reading evaluation method of extracurricular books in primary schools was obtained; The formula for calculating the predicted score of unrated books is: Where i represents the i-th book; I u represents the set of books evaluated by user u in a certain academic period; I v represents the set of books rated by user v in a certain academic period; r u,i represents the rating of the i-th book by user u in a certain academic period; r v,i represents the rating of the i-th book by user v in a certain academic period; Indicates the average rating of books of user u in a certain academic period; represents the average rating of books of user v in a certain period of study; s(u,u') represents the similarity between user u and user u'; p u,i The predicted score of the user for the unrated book.

2. The method according to claim 1, characterized in that The steps of selecting books suitable for three primary school stages from extracurricular books and constructing a corpus of primary school extracurricular reading books include: The selection criteria for extracurricular books are based on the recommended reading lists for primary and secondary schools by the Basic Curriculum Development Center of the Ministry of Education and the recommended reading lists by primary and secondary school libraries across the country.

3. The method according to claim 1, characterized in that The multiple features include: book diversity features, book complexity features and book coherence features.

4. The method according to claim 3, characterized in that The book diversity features are fitted by a multiple linear regression model using 18 features including picture diversity, Chinese character diversity, vocabulary diversity, sentence diversity and chapter diversity.

5. The method according to claim 3, characterized in that: The book complexity is fitted by a multiple linear regression model using 25 features including picture complexity, Chinese character complexity, vocabulary complexity, sentence complexity and chapter complexity.

6. The method according to claim 3, characterized in that: The book cohesion features are fitted by a multiple linear regression model using 16 features including picture cohesion, Chinese character cohesion, vocabulary cohesion, sentence cohesion and chapter cohesion.

7. The method according to claim 1, characterized in that The method for evaluating the graded reading of extracurricular books for primary schools is implemented by the following formula: Y=αL1+βL2+γL3+C Among them, Y is the applicable stage of study, L1 is the diversity of books, L2 is the complexity of books, L3 is the coherence of books, C is the bias term, α is the coefficient of book diversity, β is the coefficient of book complexity, and γ is the coefficient of book coherence.

8. A graded reading evaluation and recommendation system for extracurricular books in primary schools, characterized in that: include: User management, book management, registration and login, personal center and book recommendation function modules, wherein the book recommendation function module is used to execute the method as claimed in any one of claims 1 to claim 7.