A method and system for recommending educational books based on big data

Abstract

The present invention discloses a method and system for recommending educational books based on big data, and relates to the technical field of book recommendation. An educational book recommendation system based on big data includes: an educational book grading evaluation module, a user personal ability evaluation module, an educational book recommendation module, and an educational book update recommendation module. The present invention optimizes educational book recommendations by using educational book recommendation grading labels and target user recommendation grading labels as internal recommendation labels, so that users can more easily find books suitable for their own knowledge level, so as to better absorb and understand new knowledge; by analyzing the user's historical reading information and personal educational background, it can provide more personalized educational book recommendations to meet the user's specific learning needs; by recommending books that match the user's ability, it can help users master the required knowledge points in the shortest time and improve learning efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of book recommendation, and in particular to a method and system for recommending educational books based on big data. Background Art

[0002] As a carrier of knowledge, educational books can impart information, inspire thinking, cultivate skills, and promote personal growth and lifelong learning. In traditional educational book recommendation methods, educational book recommendations are usually made based on users' search records and basic user preference information. However, users' historical search and preference information may be very limited, resulting in sparse data, which is insufficient to form accurate user portraits and provide personalized recommendations. At the same time, matching recommendation information based on search records has a broad search scope and may not match the user's current educational needs. Users need more time to select the educational books they currently need. Summary of the invention

[0003] The present invention aims to provide an educational book recommendation method and system based on big data, to provide users with educational books that meet current educational needs and to update user educational needs in a timely manner.

[0004] A method for recommending educational books based on big data, comprising:

[0005] S1.Educational book evaluation;

[0006] For any educational book, obtain the text information of the educational book; input the text information of the educational book into the educational book recommendation content analysis model for analysis to obtain the educational book recommendation classification label; the educational book recommendation classification label contains a number of educational book recommendation keywords and corresponding knowledge point levels; the educational book recommendation content analysis model includes an educational book text extraction layer, an educational book text classification layer, an educational book text analysis layer and an educational book label output layer, which is used to conduct a comprehensive analysis of any educational book, extract all educational book recommendation keywords and corresponding knowledge point levels contained in the educational book to form an educational book recommendation classification label;

[0007] S2. User personal evaluation;

[0008] Obtain the target user's historical reading information and the target user's personal education information; the target user's historical reading information includes all the educational books that have been read and the corresponding reading time; input all the educational books that have been read into the educational book recommendation content analysis model for analysis to obtain the recommended grading labels of the educational books that have been read; the recommended grading labels of the educational books that have been read include all the target user's educational book recommendation keywords and the corresponding knowledge point levels; combine the recommended grading labels of the educational books that have been read and the corresponding reading time to obtain the target user's read educational book data set; the target user's personal education information includes the target user's personal education category and self-evaluation level; input the target user's read educational book data set and the target user's personal education information into the educational book recommendation ability assessment model for analysis to obtain the target user's recommended grading labels; the educational book recommendation ability assessment model includes an ability initial matching layer, an ability precise assessment layer and an ability assessment result output layer, which is used to conduct an internal assessment of the target user's personal ability based on the target user's historical reading information and the target user's personal education information, and obtain the target user's recommended grading labels for subsequent educational book recommendations;

[0009] S3. User education book recommendations;

[0010] Obtaining target user search terms; inputting the target user search terms into the educational book recommendation model for analysis to obtain the target user search recommended educational book set; pushing educational books to the target user based on the target user search recommended educational book set; the educational book recommendation model includes a search term analysis layer, an internal term push layer, an educational book screening layer and an educational book push layer, which is used to use the improved firefly algorithm to screen and obtain the target user search recommended educational book set;

[0011] S4. Update of user education book recommendations;

[0012] The target user's continuous reading record is obtained, and the target user's continuous reading record includes the target user's continuous reading keywords, the corresponding knowledge point level and the continuous reading time; based on the target user's continuous reading record and the educational book recommendation fatigue analysis model, the target user fatigue factor and the target user fatigue recommended educational book set are analyzed; educational books are pushed to the target user according to the target user fatigue recommended educational book set to relieve the target user's continuous reading pressure; the target user search record term data set and the target user fatigue factor data set are obtained at preset time intervals to obtain an updated target user recommendation grading label; the educational book recommendation model is updated using the updated target user recommendation grading label to obtain a more accurate educational book recommendation result.

[0013] As a preferred technical solution of the present invention, the educational book recommendation content analysis model in step S1 includes an educational book text extraction layer, an educational book text classification layer, an educational book text analysis layer and an educational book label output layer;

[0014] The educational book text extraction layer is used to perform text preprocessing on the text information of educational books to obtain several text sentences;

[0015] The educational book text classification layer is used to classify all text sentences into transition sentence sets and knowledge sentence sets;

[0016] The educational book text analysis layer is used to perform feature recognition on all text sentences in the knowledge sentence set and obtain the text feature set T n , T n =(X i , U ij , V ijk );X i Represents the text feature set T n The education category to which the corresponding text sentence belongs, U ij Represents the text feature set T n The educational category knowledge point to which the corresponding text sentence belongs, V ijk Represents the text feature set T n The education category knowledge point level to which the corresponding text sentence belongs; n = 1, 2, ..., N, N is the total number of text sentences in the knowledge sentence set; i = 1, 2, ..., I, I is the total number of set education categories; j = 1, 2, ..., J, J is the education category X i The total number of education category knowledge points set under the following conditions; k = 1, 2, ..., K, K is the education category knowledge point U ij The total number of knowledge point levels of the education category set below;

[0017] The educational book label output layer is used to classify all text feature sets T n Perform feature fusion to obtain the recommended education book classification label; the recommended education book classification label contains several recommended education book keywords and corresponding knowledge point levels, and each recommended education book keyword consists of all the same X i and the corresponding U ij The corresponding knowledge point level is obtained by feature fusion. ijk Obtained by weighted calculation.

[0018] As a preferred technical solution of the present invention, the specific steps of training the educational book text analysis layer include:

[0019] Collecting a number of text analysis training samples, each of which contains an analyzed text sentence and an annotated text feature set, the text feature set containing the education category to which the text belongs, the knowledge points of the education category to which the text belongs, and the level of the knowledge points of the education category to which the text belongs; combining the several text analysis training samples to obtain a text analysis training set;

[0020] The text analysis training set is input into the educational book recommendation content analysis model to train the educational book text analysis layer with the annotated text feature set as the target, so as to obtain the initial educational book text analysis layer; the initial educational book text analysis layer is model evaluated to obtain the initial educational book text analysis layer model evaluation result; if the initial educational book text analysis layer model evaluation result is passed, the initial educational book text analysis layer is used as the educational book text analysis layer in the educational book recommendation content analysis model; otherwise, the model training is continued using the text analysis training set.

[0021] As a preferred technical solution of the present invention, the educational book recommendation capability assessment model in step S2 includes a capability preliminary matching layer, a capability precise assessment layer and a capability assessment result output layer;

[0022] The ability initial matching layer is used to extract data based on the target user's personal education information, the education category and self-evaluation level, and obtain the education category X m and the corresponding education category initial level X m ', m = 1, 2, ..., M, where M is the total number of education categories contained in the target user's personal education information;

[0023] The capability accurate assessment layer is used to evaluate the target user's read educational books dataset and educational category X m and the corresponding education category initial level X m 'Use the ability weighted evaluation function to calculate and obtain the target user's recommended classification label;

[0024] The specific steps for calculating using the capability weighted evaluation function are:

[0025] Recommend keywords for any target user's educational books in the target user's read educational books dataset, matching educational category X m and the corresponding education category initial level X m ', get the initial level of keywords;

[0026] The initial level of the keyword, the level of the knowledge point corresponding to the keyword of the educational book recommendation for the target user, and the corresponding reading time are input into the ability weighted evaluation function for calculation to obtain the current level of the knowledge point corresponding to the keyword of the educational book recommendation for the target user; the ability weighted evaluation function is obtained by fitting using historical prior data;

[0027] The ability assessment result output layer is used to output the target user recommendation classification labels; the target user recommendation classification labels contain all the target user education book recommendation keywords and the corresponding current knowledge point levels.

[0028] As a preferred technical solution of the present invention, the educational book recommendation model includes a search term analysis layer, an internal term push layer, an educational book screening layer and an educational book push layer;

[0029] The search term analysis layer is used to perform text analysis based on the target user's search terms to obtain the target user's search keyword labels;

[0030] The internal entry push layer is used to match the target user's search keyword tags in the target user's recommended graded tags. If the match is successful, the target user's internal push keyword and the corresponding current knowledge point level are obtained; if the match is unsuccessful, the target user's search keyword tags are directly pushed to obtain the target user's search recommended educational book set, and the target user's search recommended educational book set is output;

[0031] The educational book screening layer is used to match the target user's internally pushed keywords and the corresponding current knowledge point levels to obtain a candidate set of educational book recommendations; the candidate set of educational book recommendations includes candidate educational books that match the target user's internally pushed keywords;

[0032] The educational book push layer is used to use the improved firefly algorithm to screen the candidate set of educational book recommendations, obtain the target user search recommended educational book set, and output the target user search recommended educational book set.

[0033] As a preferred technical solution of the present invention, the specific steps of using the improved firefly algorithm to screen the candidate set of educational book recommendations include:

[0034] All candidate educational books in the candidate set of educational book recommendation are used as candidate positions of the population; R initial iterative firefly individuals G are generated by using the chaos algorithm; r , r=1, 2, ..., R; each initial iteration firefly individual G r contains a position sequence for selecting candidate positions of the population; R initial iteration firefly individuals G r Combine to get the initial iterative firefly population; set the maximum number of iterations;

[0035] For the initial iteration firefly individual G in the initial iteration firefly population r Perform simulation calculation to obtain the fitness S r ;

[0036] The specific steps for simulation calculation are:

[0037] For the initial iteration firefly individual G r , calculate the keyword similarity and knowledge point level matching of all selected candidate educational books to obtain the simulation evaluation index; sum up all simulation evaluation indexes to obtain the fitness S r ;

[0038] In the initial iteration firefly population iteration, Gaussian perturbation is performed on the initial iteration firefly individuals corresponding to the current larger fitness, and the initial iteration firefly population is iteratively updated;

[0039] When the maximum number of iterations is reached, the initial iteration firefly individual G corresponding to the maximum fitness is output. r , which is the optimal initial iterative firefly individual, the candidate educational books of the corresponding position sequence in the optimal initial iterative firefly individual are combined to obtain the target user search recommended educational book set.

[0040] As a preferred technical solution of the present invention, the educational book recommendation fatigue analysis model in step S4 includes a continuous reading fatigue judgment layer, a fatigue factor output layer and a fatigue recommendation educational book recommendation layer;

[0041] The continuous reading fatigue judgment layer is used to judge the continuous reading time. If the continuous reading time exceeds the preset continuous reading time, the target user is currently in a fatigue state, and the target user's continuous reading record is input into the fatigue factor output layer and the fatigue recommendation education book recommendation layer for analysis; otherwise, no operation is performed;

[0042] The fatigue factor output layer is used to perform fatigue factor analysis based on the target user's continuous reading keywords, the corresponding knowledge point level, and the continuous reading time to obtain the target user's fatigue factor;

[0043] Specific steps for fatigue factor analysis:

[0044] For the target user's continuous reading of keywords, the fatigue factor output function is used to calculate the knowledge point level and continuous reading time corresponding to the target user's continuous reading of keywords, and the target user's unit fatigue factor is obtained;

[0045] The fatigue factor of all target users is averaged to obtain the target user fatigue factor; the fatigue factor output function is obtained by fitting the historical data;

[0046] The fatigue recommendation educational book recommendation layer is used to push educational books that match the knowledge point level according to the target user's continuous reading keywords and the corresponding knowledge point level, and obtain the target user's fatigue recommended educational book set.

[0047] An educational book recommendation system based on big data, comprising:

[0048] The educational book classification and evaluation module is deployed with an educational book recommendation content analysis model; for any educational book, the text information of the educational book is obtained; the text information of the educational book is input into the educational book recommendation content analysis model for analysis to obtain an educational book recommendation classification label; the educational book recommendation classification label contains a number of educational book recommendation keywords and corresponding knowledge point levels; the educational book recommendation content analysis model includes an educational book text extraction layer, an educational book text classification layer, an educational book text analysis layer and an educational book label output layer, which is used to conduct a comprehensive analysis of any educational book, extract all educational book recommendation keywords and corresponding knowledge point levels contained in the educational book to form an educational book recommendation classification label;

[0049] The user personal ability assessment module is deployed with an educational book recommendation ability assessment model to obtain the target user's historical reading information and the target user's personal education information; the target user's historical reading information includes all the educational books that have been read and the corresponding reading time; all the educational books that have been read are input into the educational book recommendation content analysis model for analysis to obtain the recommended graded labels of the educational books that have been read; the recommended graded labels of the educational books that have been read include all the target user's educational book recommendation keywords and the corresponding knowledge point levels; the recommended graded labels of the educational books that have been read and the corresponding reading time are combined to obtain the target user's read educational book data set; the target user's personal education information includes the target user's personal education category and self-evaluation level; the target user's read educational book data set and the target user's personal education information are input into the educational book recommendation ability assessment model for analysis to obtain the target user's recommended graded labels; the educational book recommendation ability assessment model includes an ability initial matching layer, an ability precise assessment layer and an ability assessment result output layer, which are used to conduct internal assessment of the target user's personal ability based on the target user's historical reading information and the target user's personal education information, and obtain the target user's recommended graded labels for subsequent educational book recommendations;

[0050] The educational book recommendation module deploys an educational book recommendation model to obtain the target user search terms; input the target user search terms into the educational book recommendation model for analysis to obtain the target user search recommended educational book set; push educational books to the target user based on the target user search recommended educational book set; the educational book recommendation model includes a search term analysis layer, an internal term push layer, an educational book screening layer and an educational book push layer, which are used to use the improved firefly algorithm to screen and obtain the target user search recommended educational book set;

[0051] The educational book update recommendation module deploys an educational book recommendation fatigue analysis model to obtain the target user's continuous reading record, which contains the target user's continuous reading keywords, corresponding knowledge point levels and continuous reading time; based on the target user's continuous reading record and the educational book recommendation fatigue analysis model, the target user fatigue factor and the target user fatigue recommended educational book set are analyzed; educational books are pushed to the target user based on the target user fatigue recommended educational book set to alleviate the target user's continuous reading pressure; the target user search record term dataset and the target user fatigue factor dataset are obtained at preset time intervals to obtain updated target user recommendation classification labels; the educational book recommendation model is updated using the updated target user recommendation classification labels to obtain more accurate educational book recommendation results.

[0052] The present invention has the following advantages:

[0053] 1. The present invention optimizes educational book recommendations by utilizing educational book recommendation grading labels and target user recommendation grading labels as internal recommendation labels, so that users can more easily find books that suit their knowledge level, thereby better absorbing and understanding new knowledge; by analyzing the user's historical reading information and personal educational background, more personalized educational book recommendations can be provided to meet the user's specific learning needs; by recommending books that match the user's abilities, the user can master the required knowledge points in the shortest possible time and improve learning efficiency; by analyzing the user's reading fatigue factor, some relaxing or different types of books can be recommended to help the user relieve reading pressure and maintain reading interest.

[0054] 2. The present invention ensures the relevance and timeliness of recommendations by dynamically updating recommended content based on users' continuous reading records and search behaviors; by using the improved Firefly algorithm to screen candidate educational books, educational books that better meet user needs are recommended to increase user satisfaction with educational book reading, and can provide more accurate and diverse book selections, enhance users' reading experience, and help build a lifelong learning environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 A flowchart of an educational book recommendation method based on big data adopted in an embodiment of the present invention.

[0056] Figure 2 A schematic diagram of the structure of an educational book recommendation system based on big data adopted in an embodiment of the present invention. DETAILED DESCRIPTION

[0057] In order to enable persons skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0058] Example 1, a method for recommending educational books based on big data, see Figure 1 As shown, including:

[0059] S1. Educational Book Evaluation

[0060] For any educational book, obtain the text information of the educational book; input the text information of the educational book into the educational book recommendation content analysis model for analysis to obtain the educational book recommendation classification label; the educational book recommendation classification label contains a number of educational book recommendation keywords and corresponding knowledge point levels; the educational book recommendation content analysis model includes an educational book text extraction layer, an educational book text classification layer, an educational book text analysis layer and an educational book label output layer, which is used to conduct a comprehensive analysis of any educational book, extract all educational book recommendation keywords and corresponding knowledge point levels contained in the educational book to form an educational book recommendation classification label;

[0061] For example, for popular science educational books, which may contain several knowledge points, keywords can be extracted and level matched for each knowledge point in the popular science educational books, so that accurate matching can be achieved in subsequent searches, and the possibility of recommendation is greater; in the analysis of educational books, big data technology is used for model training to mark the levels of knowledge points;

[0062] The educational book recommendation content analysis model in step S1 includes an educational book text extraction layer, an educational book text classification layer, an educational book text analysis layer and an educational book label output layer;

[0063] The educational book text extraction layer is used to perform text preprocessing on the text information of educational books to obtain several text sentences;

[0064] The educational book text classification layer is used to classify all text sentences into transition sentence sets and knowledge sentence sets;

[0065] The educational book text analysis layer is used to perform feature recognition on all text sentences in the knowledge sentence set and obtain the text feature set T n , T n =(X i , U ij , V ijk );X i Represents the text feature set T n The education category to which the corresponding text sentence belongs, U ij Represents the text feature set T nThe educational category knowledge point to which the corresponding text sentence belongs, V ijk Represents the text feature set T n The education category knowledge point level to which the corresponding text sentence belongs; n = 1, 2, ..., N, N is the total number of text sentences in the knowledge sentence set; i = 1, 2, ..., I, I is the total number of set education categories; j = 1, 2, ..., J, J is the education category X i The total number of education category knowledge points set under the following conditions; k = 1, 2, ..., K, K is the education category knowledge point U ij The total number of knowledge point levels of the education category set below;

[0066] The educational book label output layer is used to classify all text feature sets T n Perform feature fusion to obtain the recommended education book classification label; the recommended education book classification label contains several recommended education book keywords and corresponding knowledge point levels, and each recommended education book keyword consists of all the same X i and the corresponding U ij The corresponding knowledge point level is obtained by feature fusion. ijk The weighted calculation is performed;

[0067] The specific steps for training the educational book text analysis layer include:

[0068] Collecting a number of text analysis training samples, each of which contains an analyzed text sentence and an annotated text feature set, the text feature set containing the education category to which the text belongs, the knowledge points of the education category to which the text belongs, and the level of the knowledge points of the education category to which the text belongs; combining the several text analysis training samples to obtain a text analysis training set;

[0069] Input the text analysis training set into the educational book recommendation content analysis model to train the educational book text analysis layer with the annotated text feature set as the target, and obtain the initial educational book text analysis layer; perform model evaluation on the initial educational book text analysis layer to obtain the initial educational book text analysis layer model evaluation result; if the initial educational book text analysis layer model evaluation result is passed, the initial educational book text analysis layer is used as the educational book text analysis layer in the educational book recommendation content analysis model; otherwise, continue model training using the text analysis training set;

[0070] The BP neural network model can be used as the basic model of the content analysis model for educational book recommendation. The preprocessing of the educational book text extraction layer can ensure the quality and analyzability of text sentences, laying a solid foundation for subsequent text analysis. The educational book text classification layer can effectively divide text sentences into transition sentences and knowledge sentences, which helps to concentrate resources to analyze sentences containing knowledge points. The educational book text classification layer is obtained after classification training. The educational book text analysis layer can deeply understand the educational content and structure of text sentences through feature recognition, and provide detailed text features for the recommendation of educational books. The educational book label output layer generates recommendation graded labels containing keywords and knowledge point levels through feature fusion, making the recommendation more specific and targeted. Through a precise text feature set, the relevance of educational book recommendations can be improved to ensure that the recommended content matches the actual needs and knowledge level of users.

[0071] S2. User personal evaluation

[0072] Obtain the target user's historical reading information and the target user's personal education information; the target user's historical reading information includes all the educational books that have been read and the corresponding reading time; input all the educational books that have been read into the educational book recommendation content analysis model for analysis to obtain the recommended grading labels of the educational books that have been read; the recommended grading labels of the educational books that have been read include all the target user's educational book recommendation keywords and the corresponding knowledge point levels; combine the recommended grading labels of the educational books that have been read and the corresponding reading time to obtain the target user's read educational book data set; the target user's personal education information includes the target user's personal education category and self-evaluation level; input the target user's read educational book data set and the target user's personal education information into the educational book recommendation ability assessment model for analysis to obtain the target user's recommended grading labels; the educational book recommendation ability assessment model includes an ability initial matching layer, an ability precise assessment layer and an ability assessment result output layer, which is used to conduct an internal assessment of the target user's personal ability based on the target user's historical reading information and the target user's personal education information, and obtain the target user's recommended grading labels for subsequent educational book recommendations;

[0073] For user personal evaluation, before the recommendation begins, with the user's consent, the user's other historical reading information can be obtained using big data technology, and a preliminary personal information survey of the user can be conducted; by analyzing the user's historical reading information and personal educational background, personalized recommendation entries can be customized for each user to improve learning efficiency; through the initial ability matching layer and the ability precise evaluation layer, educational books can be matched more effectively to ensure that users can obtain the most suitable learning materials for themselves;

[0074] The educational book recommendation capability assessment model in step S2 includes a capability preliminary matching layer, a capability precise assessment layer, and a capability assessment result output layer;

[0075] The ability initial matching layer is used to extract data based on the target user's personal education information, the education category and self-evaluation level, and obtain the education category X m and the corresponding education category initial level X m ', m = 1, 2, ..., M, where M is the total number of education categories contained in the target user's personal education information;

[0076] The capability accurate assessment layer is used to evaluate the target user's read educational books dataset and educational category X m and the corresponding education category initial level X m 'Use the ability weighted evaluation function to calculate and obtain the target user's recommended classification label;

[0077] The specific steps for calculating using the capability weighted evaluation function are:

[0078] Recommend keywords for any target user's educational books in the target user's read educational books dataset, matching educational category X m and the corresponding education category initial level X m ', get the initial level of keywords;

[0079] The initial level of the keyword, the level of the knowledge point corresponding to the target user's educational book recommendation keyword and the corresponding reading time are input into the ability weighted evaluation function for calculation, and the current level of the knowledge point corresponding to the target user's educational book recommendation keyword is obtained; the ability weighted evaluation function is obtained by fitting using historical prior data; the ability weighted evaluation function is set under the training of professional technicians to meet the calculation requirements;

[0080] The ability assessment result output layer is used to output the target user recommendation classification label; the target user recommendation classification label contains all the target user education book recommendation keywords and the corresponding current knowledge point level;

[0081] S3. User Education Book Recommendations

[0082] Obtaining target user search terms; inputting the target user search terms into the educational book recommendation model for analysis to obtain the target user search recommended educational book set; pushing educational books to the target user based on the target user search recommended educational book set; the educational book recommendation model includes a search term analysis layer, an internal term push layer, an educational book screening layer and an educational book push layer, which is used to use the improved firefly algorithm to screen and obtain the target user search recommended educational book set;

[0083] The educational book recommendation model includes a search term analysis layer, an internal term push layer, an educational book screening layer, and an educational book push layer;

[0084] The search term analysis layer is used to perform text analysis based on the target user's search terms to obtain the target user's search keyword labels;

[0085] The internal entry push layer is used to match the target user's search keyword tags in the target user's recommended graded tags. If the match is successful, the target user's internal push keyword and the corresponding current knowledge point level are obtained; if the match is unsuccessful, the target user's search keyword tags are directly pushed to obtain the target user's search recommended educational book set, and the target user's search recommended educational book set is output;

[0086] The educational book screening layer is used to match the target user's internally pushed keywords and the corresponding current knowledge point levels to obtain a candidate set of educational book recommendations; the candidate set of educational book recommendations includes candidate educational books that match the target user's internally pushed keywords;

[0087] The educational book push layer is used to use the improved firefly algorithm to screen the candidate set of educational book recommendations, obtain the target user search recommended educational book set, and output the target user search recommended educational book set;

[0088] The specific steps of using the improved firefly algorithm to screen the candidate set of educational book recommendations include:

[0089] All candidate educational books in the candidate set of educational book recommendation are used as candidate positions of the population; R initial iterative firefly individuals G are generated by using the chaos algorithm; r , r=1, 2, ..., R; each initial iteration firefly individual G r contains a position sequence for selecting candidate positions of the population; R initial iteration firefly individuals G r combination to obtain the initial iterative firefly population; set the maximum number of iterations; the maximum number of iterations is set by professional technicians according to actual conditions;

[0090] For the initial iteration firefly individual G in the initial iteration firefly population r Perform simulation calculation to obtain the fitness S r ;

[0091] The specific steps for simulation calculation are:

[0092] For the initial iteration firefly individual G r , calculate the keyword similarity and knowledge point level matching of all selected candidate educational books to obtain the simulation evaluation index; sum up all simulation evaluation indexes to obtain the fitness Sr ;

[0093] In the actual screening process, for example, the search keyword ABBC is matched with the keywords AB, BC, BB, and a number of educational books that match the keywords AB, BC, BB are matched. The keyword similarity is calculated based on the candidate educational books. The greater the keyword similarity, the greater the simulation evaluation index of the candidate educational book;

[0094] At the same time, for the same keyword AB, if the user's current knowledge point level is 5, the knowledge point level of AB in the educational book may be 3, 5, 6, 7, 9; at this time, the educational book with a knowledge point level of 6 has the highest matching degree; in the actual learning process, users need to master knowledge points that are higher than their current knowledge level to learn;

[0095] In the initial iteration firefly population iteration, Gaussian perturbation is performed on the initial iteration firefly individuals corresponding to the current larger fitness, and the initial iteration firefly population is iteratively updated;

[0096] When the maximum number of iterations is reached, the initial iteration firefly individual G corresponding to the maximum fitness is output. r , which is the optimal initial iterative firefly individual, combines the candidate educational books of the corresponding position sequence in the optimal initial iterative firefly individual to obtain the target user search recommended educational book set;

[0097] The internal entry push layer and the educational book screening layer combine the user's recommendation classification tags and search keywords to ensure that the pushed educational books match the user's personal educational background and interests; the improved firefly algorithm is used for screening to improve the efficiency and accuracy of the educational book recommendation process; through iterative updates and Gaussian perturbations, it can dynamically adapt to changes in user needs and continuously optimize recommendation results; through precise keyword similarity and knowledge point level matching calculations, it can provide more personalized and high-quality educational book recommendations, improve user experience, recommend educational books that match the user's current knowledge point level, and help users learn at an appropriate difficulty level, thereby improving learning efficiency;

[0098] S4. Update of user education book recommendations

[0099] Obtain the target user's continuous reading record, which includes the target user's continuous reading keywords, the corresponding knowledge point level and the continuous reading time; analyze the target user's continuous reading record and the educational book recommendation fatigue analysis model to obtain the target user's fatigue factor and the target user's fatigue recommended educational book set; push educational books to the target user based on the target user's fatigue recommended educational book set to relieve the target user's continuous reading pressure; obtain the target user's search record term data set and the target user's fatigue factor data set at preset time intervals to obtain updated target user recommendation classification labels; use the updated target user recommendation classification labels to update the educational book recommendation model to obtain more accurate educational book recommendation results; the preset time interval is set by professional and technical personnel according to actual conditions;

[0100] The educational book recommendation fatigue analysis model in step S4 includes a continuous reading fatigue judgment layer, a fatigue factor output layer and a fatigue recommendation educational book recommendation layer;

[0101] The continuous reading fatigue judgment layer is used to judge the continuous reading time. If the continuous reading time exceeds the preset continuous reading time, the target user is currently in a fatigue state, and the target user's continuous reading record is input into the fatigue factor output layer and the fatigue recommendation education book recommendation layer for analysis; otherwise, no operation is performed; the preset continuous reading time is set by professional technicians according to actual conditions;

[0102] The fatigue factor output layer is used to perform fatigue factor analysis based on the target user's continuous reading keywords, the corresponding knowledge point level, and the continuous reading time to obtain the target user's fatigue factor;

[0103] Specific steps for fatigue factor analysis:

[0104] For the target user's continuous reading of keywords, the fatigue factor output function is used to calculate the knowledge point level and continuous reading time corresponding to the target user's continuous reading of keywords, and the target user's unit fatigue factor is obtained;

[0105] The fatigue factor of all target users is calculated as the average to obtain the target user fatigue factor; the fatigue factor output function is obtained by fitting historical data; the fatigue factor output function is set by professional technicians based on actual calculation and evaluation to meet actual usage requirements;

[0106] The fatigue recommendation educational book recommendation layer is used to push educational books that match the knowledge point level according to the target user's continuous reading keywords and the corresponding knowledge point level, and obtain the target user fatigue recommendation educational book set;

[0107] Through the continuous reading fatigue judgment layer, it is possible to identify the user's fatigue state and adjust the recommendation strategy accordingly to manage the user's reading fatigue in a personalized way; through the fatigue factor output layer, it is possible to calculate the user's fatigue level and recommend appropriate educational books based on the fatigue level to optimize the user's reading experience; the fatigue recommendation educational book recommendation layer can dynamically adjust the recommended content based on the user's fatigue state and reading history to ensure that users can obtain suitable educational books in different states; when calculating the fatigue factor, the user's actual knowledge point level increases with the generation of the fatigue factor, and the target user's fatigue recommended educational book set is used to expand the user's educational book reading record, which has strong adaptability.

[0108] Example 2, an educational book recommendation system based on big data, see Figure 2 As shown, including:

[0109] The educational book classification and evaluation module is deployed with an educational book recommendation content analysis model; for any educational book, the text information of the educational book is obtained; the text information of the educational book is input into the educational book recommendation content analysis model for analysis to obtain an educational book recommendation classification label; the educational book recommendation classification label contains a number of educational book recommendation keywords and corresponding knowledge point levels; the educational book recommendation content analysis model includes an educational book text extraction layer, an educational book text classification layer, an educational book text analysis layer and an educational book label output layer, which is used to conduct a comprehensive analysis of any educational book, extract all educational book recommendation keywords and corresponding knowledge point levels contained in the educational book to form an educational book recommendation classification label;

[0110] The user personal ability assessment module is deployed with an educational book recommendation ability assessment model to obtain the target user's historical reading information and the target user's personal education information; the target user's historical reading information includes all the educational books that have been read and the corresponding reading time; all the educational books that have been read are input into the educational book recommendation content analysis model for analysis to obtain the recommended graded labels of the educational books that have been read; the recommended graded labels of the educational books that have been read include all the target user's educational book recommendation keywords and the corresponding knowledge point levels; the recommended graded labels of the educational books that have been read and the corresponding reading time are combined to obtain the target user's read educational book data set; the target user's personal education information includes the target user's personal education category and self-evaluation level; the target user's read educational book data set and the target user's personal education information are input into the educational book recommendation ability assessment model for analysis to obtain the target user's recommended graded labels; the educational book recommendation ability assessment model includes an ability initial matching layer, an ability precise assessment layer and an ability assessment result output layer, which are used to conduct internal assessment of the target user's personal ability based on the target user's historical reading information and the target user's personal education information, and obtain the target user's recommended graded labels for subsequent educational book recommendations;

[0111] The educational book recommendation module deploys an educational book recommendation model to obtain the target user search terms; input the target user search terms into the educational book recommendation model for analysis to obtain the target user search recommended educational book set; push educational books to the target user based on the target user search recommended educational book set; the educational book recommendation model includes a search term analysis layer, an internal term push layer, an educational book screening layer and an educational book push layer, which are used to use the improved firefly algorithm to screen and obtain the target user search recommended educational book set;

[0112] The educational book update recommendation module deploys an educational book recommendation fatigue analysis model to obtain the target user's continuous reading record, which contains the target user's continuous reading keywords, corresponding knowledge point levels and continuous reading time; based on the target user's continuous reading record and the educational book recommendation fatigue analysis model, the target user fatigue factor and the target user fatigue recommended educational book set are analyzed; educational books are pushed to the target user based on the target user fatigue recommended educational book set to alleviate the target user's continuous reading pressure; the target user search record term dataset and the target user fatigue factor dataset are obtained at preset time intervals to obtain updated target user recommendation classification labels; the educational book recommendation model is updated using the updated target user recommendation classification labels to obtain more accurate educational book recommendation results.

[0113] It should be understood that those skilled in the art can make improvements or changes based on the above description, and all such improvements and changes should fall within the scope of protection of the appended claims of the present invention. Parts not described in detail in this specification belong to the prior art known to those skilled in the art.

Claims

1. A method for recommending educational books based on big data, characterized in that: The following steps are involved: Step S1. Educational book evaluation; For any educational book, obtain the text information of the educational book; input the text information of the educational book into the educational book recommendation content analysis model for analysis to obtain the educational book recommendation classification label; the educational book recommendation classification label contains a number of educational book recommendation keywords and corresponding knowledge point levels; the educational book recommendation content analysis model includes an educational book text extraction layer, an educational book text classification layer, an educational book text analysis layer and an educational book label output layer, which is used to conduct a comprehensive analysis of any educational book, extract all educational book recommendation keywords and corresponding knowledge point levels contained in the educational book to form an educational book recommendation classification label; Step S2. User personal evaluation; Obtain the target user's historical reading information and the target user's personal education information; the target user's historical reading information includes all the educational books that have been read and the corresponding reading time; input all the educational books that have been read into the educational book recommendation content analysis model for analysis, and obtain the recommended classification labels of the educational books that have been read; The recommended classification labels of read educational books contain all the recommended keywords of educational books for target users and the corresponding knowledge point levels; Combine the recommended graded labels of the educational books that have been read and the corresponding reading time to obtain the educational book dataset that the target user has read; the personal educational information of the target user includes the educational categories and self-evaluation levels that the target user has studied; input the educational book dataset that the target user has read and the personal educational information of the target user into the educational book recommendation capability evaluation model for analysis to obtain the target user recommendation graded labels; the educational book recommendation capability evaluation model includes a capability initial matching layer, a capability precise evaluation layer and a capability evaluation result output layer, which is used to conduct an internal evaluation of the target user's personal capability based on the target user's historical reading information and the target user's personal educational information, and obtain the target user recommendation graded labels for subsequent educational book recommendations; Step S3. User educational book recommendation; Obtaining target user search terms; inputting the target user search terms into the educational book recommendation model for analysis to obtain the target user search recommended educational book set; pushing educational books to the target user based on the target user search recommended educational book set; the educational book recommendation model includes a search term analysis layer, an internal term push layer, an educational book screening layer and an educational book push layer, which is used to use the improved firefly algorithm to screen and obtain the target user search recommended educational book set; Step S4. Update user education book recommendations; Obtaining the target user's continuous reading record, which includes the target user's continuous reading keywords, corresponding knowledge point levels and continuous reading time; Based on the target user's continuous reading record and the educational book recommendation fatigue analysis model, the target user fatigue factor and the target user fatigue recommended educational book set are analyzed; educational books are pushed to the target user according to the target user fatigue recommended educational book set to relieve the target user's continuous reading pressure; the target user search record term data set and the target user fatigue factor data set are obtained at preset time intervals to obtain updated target user recommendation classification labels; the educational book recommendation model is updated using the updated target user recommendation classification labels to obtain more accurate educational book recommendation results; The educational book recommendation content analysis model in step S1 includes an educational book text extraction layer, an educational book text classification layer, an educational book text analysis layer and an educational book label output layer; The educational book text extraction layer is used to perform text preprocessing on the text information of educational books to obtain several text sentences; The educational book text classification layer is used to classify all text sentences into transition sentence sets and knowledge sentence sets; The educational book text analysis layer is used to perform feature recognition on all text sentences in the knowledge sentence set and obtain the text feature set T n , T n =(X i , U ij , V ijk );X i Represents the text feature set T n The education category to which the corresponding text sentence belongs, U ij Represents the text feature set T n The educational category knowledge point to which the corresponding text sentence belongs, V ijk Represents the text feature set T n The education category knowledge point level to which the corresponding text sentence belongs; n=1, 2, ..., N, where N is the total number of text sentences in the knowledge sentence set; i=1, 2, …, I, where I is the total number of education categories set; j=1, 2, …, J, where J is the education category X i The total number of education category knowledge points set under the following; k = 1, 2, ..., K, K is the education category knowledge point U ij The total number of knowledge point levels of the education category set below; The educational book label output layer is used to classify all text feature sets T n Perform feature fusion to obtain the recommended education book classification label; the recommended education book classification label contains several recommended education book keywords and corresponding knowledge point levels, and each recommended education book keyword consists of all the same X i and the corresponding U ij The corresponding knowledge point level is obtained by feature fusion. ijk Obtained by weighted calculation.

2. The method for recommending educational books based on big data according to claim 1, characterized in that: The specific steps for training the educational book text analysis layer include: Collecting a number of text analysis training samples, each of which contains an analyzed text sentence and an annotated text feature set, the text feature set containing the education category to which the text belongs, the knowledge points of the education category to which the text belongs, and the level of the knowledge points of the education category to which the text belongs; combining the several text analysis training samples to obtain a text analysis training set; The text analysis training set is input into the educational book recommendation content analysis model to train the educational book text analysis layer with the annotated text feature set as the target, and obtain the initial educational book text analysis layer; the initial educational book text analysis layer is model evaluated to obtain the initial educational book text analysis layer model evaluation result; if the initial educational book text analysis layer model evaluation result is passed, the initial educational book text analysis layer is used as the educational book text analysis layer in the educational book recommendation content analysis model; otherwise, the model training is continued using the text analysis training set.

3. The method for recommending educational books based on big data according to claim 2, characterized in that: The educational book recommendation capability assessment model in step S2 includes a capability preliminary matching layer, a capability precise assessment layer, and a capability assessment result output layer; The ability initial matching layer is used to extract data based on the target user's personal education information, the education category and self-evaluation level, and obtain the education category X m and the corresponding education category initial level X m ', m = 1, 2, ..., M, where M is the total number of education categories contained in the target user's personal education information; The capability accurate assessment layer is used to evaluate the target user's read educational books dataset and educational category X m and the corresponding education category initial level X m 'Use the ability weighted evaluation function to calculate and obtain the target user's recommended classification label; The specific steps for calculating using the capability weighted evaluation function are: Recommend keywords for any target user's educational books in the target user's read educational books dataset, matching the educational category X m and the corresponding education category initial level X m ', get the initial level of keywords; The initial level of the keyword, the level of the knowledge point corresponding to the keyword of the educational book recommendation for the target user, and the corresponding reading time are input into the ability weighted evaluation function for calculation to obtain the current level of the knowledge point corresponding to the keyword of the educational book recommendation for the target user; The capability weighted evaluation function is obtained by fitting using historical prior data; The capability assessment result output layer is used to output the target user's recommended classification labels; The target user recommendation classification labels include all target user educational book recommendation keywords and the corresponding current knowledge point levels.

4. The method for recommending educational books based on big data according to claim 3, characterized in that: The educational book recommendation model includes a search term analysis layer, an internal term push layer, an educational book screening layer, and an educational book push layer; The search term analysis layer is used to perform text analysis based on the target user's search terms to obtain the target user's search keyword labels; The internal entry push layer is used to match the target user's search keyword tags with the target user's recommended graded tags. If the match is successful, the target user's internal push keyword and the corresponding current knowledge point level are obtained; If the match is unsuccessful, the target user searches for keywords and tags and directly pushes the recommended educational books to obtain the target user's searched educational books, and outputs the target user's searched educational books; The educational book screening layer is used to match the target user's internally pushed keywords and the corresponding current knowledge point level to obtain a candidate set of educational book recommendations; the candidate set of educational book recommendations includes candidate educational books that match the target user's internally pushed keywords; The educational book push layer is used to use the improved firefly algorithm to screen the candidate set of educational book recommendations, obtain the target user search recommended educational book set, and output the target user search recommended educational book set.

5. The method for recommending educational books based on big data according to claim 4, characterized in that: The specific steps of using the improved firefly algorithm to screen the candidate set of educational book recommendations include: All candidate educational books in the candidate set of educational book recommendation are used as candidate positions of the population; R initial iterative firefly individuals G are generated by using the chaos algorithm; r , r=1, 2, ..., R; each initial iteration firefly individual G r contains a position sequence for selecting candidate positions of the population; R initial iteration firefly individuals G r Combine to get the initial iterative firefly population; set the maximum number of iterations; For the initial iteration firefly individual G in the initial iteration firefly population r Perform simulation calculation to obtain the fitness S r ; The specific steps for simulation calculation are: For the initial iteration firefly individual G r , calculate the keyword similarity and knowledge point level matching of all selected candidate educational books to obtain the simulation evaluation index; sum up all simulation evaluation indexes to obtain the fitness S r ; In the initial iteration firefly population iteration, Gaussian perturbation is performed on the initial iteration firefly individuals corresponding to the current larger fitness, and the initial iteration firefly population is iteratively updated; When the maximum number of iterations is reached, the initial iteration firefly individual G corresponding to the maximum fitness is output. r , which is the optimal initial iterative firefly individual, the candidate educational books of the corresponding position sequence in the optimal initial iterative firefly individual are combined to obtain the target user search recommended educational book set.

6. The method for recommending educational books based on big data according to claim 5, characterized in that: The educational book recommendation fatigue analysis model in step S4 includes a continuous reading fatigue judgment layer, a fatigue factor output layer and a fatigue recommendation educational book recommendation layer; The continuous reading fatigue judgment layer is used to judge the continuous reading time. If the continuous reading time exceeds the preset continuous reading time, the target user is currently in a fatigue state, and the target user's continuous reading record is input into the fatigue factor output layer and the fatigue recommendation education book recommendation layer for analysis; otherwise, no operation is performed; The fatigue factor output layer is used to perform fatigue factor analysis based on the target user's continuous reading keywords, the corresponding knowledge point level, and the continuous reading time to obtain the target user's fatigue factor; Specific steps for fatigue factor analysis: For the target user's continuous reading of keywords, the fatigue factor output function is used to calculate the knowledge point level and continuous reading time corresponding to the target user's continuous reading of keywords, and the target user's unit fatigue factor is obtained; The fatigue factor of all target users is averaged to obtain the target user fatigue factor; the fatigue factor output function is obtained by fitting the historical data; The fatigue recommendation educational book recommendation layer is used to push educational books that match the knowledge point level according to the target user's continuous reading keywords and the corresponding knowledge point level, and obtain the target user's fatigue recommended educational book set.

7. An educational book recommendation system based on big data, characterized in that: The system applies the educational book recommendation method based on big data as described in any one of claims 1 to 6, including: The educational book classification and evaluation module is deployed with an educational book recommendation content analysis model; for any educational book, the text information of the educational book is obtained; the text information of the educational book is input into the educational book recommendation content analysis model for analysis to obtain an educational book recommendation classification label; the educational book recommendation classification label contains a number of educational book recommendation keywords and corresponding knowledge point levels; the educational book recommendation content analysis model includes an educational book text extraction layer, an educational book text classification layer, an educational book text analysis layer and an educational book label output layer, which is used to conduct a comprehensive analysis of any educational book, extract all educational book recommendation keywords and corresponding knowledge point levels contained in the educational book to form an educational book recommendation classification label; The user personal ability assessment module is deployed with an educational book recommendation ability assessment model to obtain the target user's historical reading information and the target user's personal education information; the target user's historical reading information includes all the educational books that have been read and the corresponding reading time; all the educational books that have been read are input into the educational book recommendation content analysis model for analysis to obtain the recommended graded labels of the educational books that have been read; the recommended graded labels of the educational books that have been read include all the target user's educational book recommendation keywords and the corresponding knowledge point levels; the recommended graded labels of the educational books that have been read and the corresponding reading time are combined to obtain the target user's read educational book data set; the target user's personal education information includes the target user's personal education category and self-evaluation level; the target user's read educational book data set and the target user's personal education information are input into the educational book recommendation ability assessment model for analysis to obtain the target user's recommended graded labels; the educational book recommendation ability assessment model includes an ability initial matching layer, an ability precise assessment layer and an ability assessment result output layer, which are used to conduct internal assessment of the target user's personal ability based on the target user's historical reading information and the target user's personal education information, and obtain the target user's recommended graded labels for subsequent educational book recommendations; The educational book recommendation module deploys an educational book recommendation model to obtain the target user search terms; input the target user search terms into the educational book recommendation model for analysis to obtain the target user search recommended educational book set; push educational books to the target user based on the target user search recommended educational book set; the educational book recommendation model includes a search term analysis layer, an internal term push layer, an educational book screening layer and an educational book push layer, which are used to use the improved firefly algorithm to screen and obtain the target user search recommended educational book set; The educational book update recommendation module deploys an educational book recommendation fatigue analysis model to obtain the target user's continuous reading record, which contains the target user's continuous reading keywords, corresponding knowledge point levels and continuous reading time; based on the target user's continuous reading record and the educational book recommendation fatigue analysis model, the target user's fatigue factor and the target user's fatigue recommended educational book set are analyzed; educational books are pushed to the target user based on the target user's fatigue recommended educational book set to relieve the target user's continuous reading pressure; the target user's search record term dataset and the target user's fatigue factor dataset are obtained at preset time intervals to obtain the updated target user recommendation classification label; The educational book recommendation model is updated by updating the target user's recommended rating labels to obtain more accurate educational book recommendation results.

Images