Smart art education resource recommendation system for full-scene teaching application

Abstract

The invention discloses an intelligent art education resource recommendation system for full-scene teaching application, and belongs to the technical field of artificial intelligence and art education resource recommendation. The system comprises a recommendation control module, a personalized learning recommendation module, a learning path recommendation module, a broadened learning recommendation module, an optimized course recommendation module and a multi-modal data module. According to the system, multiple modules are arranged, multi-aspect factors of users or articles are comprehensively considered, and education resource recommendation of the users is realized in combination with the synergistic effect of the user-based collaborative filtering sub-module and the article-based collaborative filtering sub-module, so that personalized requirements of the users can be better met; students can be more effectively guided to find the relation between different subjects, cross-domain learning is carried out, and the method has important significance for improving the comprehensive artistic quality of users.

Description

Technical Field

[0001] The present invention belongs to the technical field of artificial intelligence and art education resource recommendation, and in particular relates to an intelligent art education resource recommendation system for full-scenario teaching applications. Background Art

[0002] At present, with digital learning and personalized education receiving increasing attention, online art education is becoming an increasingly important way for people to learn and acquire knowledge. However, a mature, standardized and systematic mainstream online art education platform has not yet been formed. In particular, how users can better obtain recommendations for educational resources that are both professional and personalized in the process of studying art professional courses is of great significance to the effectiveness and efficiency of users' art education learning.

[0003] Therefore, while online education platforms provide users and learners with rich art course resources, how to help students find the most suitable courses from a large number of courses has become a key challenge for education platforms and one of the important issues that current online education platforms need to solve. Summary of the Invention

[0004] In view of this, the present invention provides a smart art education resource recommendation system for all-scenario teaching applications.

[0005] The present invention adopts the following technical solutions:

[0006] A smart art education resource recommendation system for full-scenario teaching applications, the system is used for full-scenario teaching, the system includes: a recommendation control module, a personalized learning recommendation module, a learning path recommendation module, a broadened learning recommendation module, an optimized course recommendation module, and a multimodal data module; the personalized learning recommendation module, the learning path recommendation module, the broadened learning recommendation module, the optimized course recommendation module, and the multimodal data module are all connected to the recommendation control module;

[0007] The recommendation control module makes course recommendations based on the multimodal data module through the personalized learning recommendation module, the learning path recommendation module, the broadened learning recommendation module, the optimized course recommendation module and the multimodal data module;

[0008] The personalized learning recommendation module provides personalized course recommendations to users based on their historical learning data;

[0009] The learning path recommendation module provides course recommendations from entry-level to advanced levels based on the user's current learning progress and skill level;

[0010] The broadened learning recommendation module recommends relevant courses based on the user's current learning content, which is used to broaden the current learning content and enrich the user's course options;

[0011] The optimized course recommendation module recommends courses based on user interests and needs;

[0012] The multimodal data module is used to integrate course content, professional settings, tutor information, student background, and interaction data to build a multimodal information database.

[0013] Furthermore, the recommendation control module includes a user-based collaborative filtering submodule; the user-based collaborative filtering submodule includes: finding previous users similar to the current user, and screening and recommending the previous users' selected courses to the current user.

[0014] Furthermore, the recommendation control module further includes an item-based collaborative filtering submodule; the item-based collaborative filtering submodule includes: recommending new courses similar to the user's previous interest courses to the user based on the user's previous interest courses.

[0015] Furthermore, the user-based collaborative filtering submodule further includes: constructing a user-item matrix, calculating user similarity, and generating a recommendation list.

[0016] Furthermore, the item-based collaborative filtering submodule further includes: constructing a user-item matrix, calculating item similarity, and generating a recommendation list.

[0017] Furthermore, the user similarity assessment content includes the user's education level, major, occupation, gender, age or interests.

[0018] Furthermore, the content of the examination of the similarity of the items includes the type, name, application field or place of use of the items.

[0019] Furthermore, the methods for calculating user similarity and item similarity include: cosine similarity, Pearson correlation coefficient and Jaccard similarity coefficient.

[0020] Furthermore, the user-based collaborative filtering submodule and the item-based collaborative filtering submodule are connected via a collaborative control submodule.

[0021] Furthermore, the user similarity assessment also includes academic background or career planning.

[0022] Beneficial effects of the present invention:

[0023] The system of the present invention comprehensively considers multiple factors of users or items through the setting of multiple modules, and realizes the recommendation of educational resources for users by combining the synergy of user-based collaborative filtering sub-module and item-based collaborative filtering sub-module. It can better adapt to the personalized requirements of users, provide users with more accurate course recommendations, and more effectively guide students to discover the connections between different subjects and conduct cross-disciplinary learning, which is of great significance for improving the comprehensive artistic literacy of users. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0025] Figure 1 This is a schematic diagram of the control flow of the user collaborative filtering submodule of the present invention;

[0026] Figure 2 This is a schematic diagram of the control flow of the item collaborative filtering submodule of the present invention. DETAILED DESCRIPTION

[0027] In order to better understand the technical solution of the present invention, the embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0028] It should be understood that the embodiments described are only a portion of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative work are within the scope of protection of the present invention.

[0029] Example 1

[0030] A smart art education resource recommendation system for full-scenario teaching applications, the system is used for full-scenario teaching, the system includes: a recommendation control module, a personalized learning recommendation module, a learning path recommendation module, a broadened learning recommendation module, an optimized course recommendation module, and a multimodal data module; the personalized learning recommendation module, the learning path recommendation module, the broadened learning recommendation module, the optimized course recommendation module, and the multimodal data module are all connected to the recommendation control module;

[0031] The recommendation control module makes course recommendations based on the multimodal data module through the personalized learning recommendation module, the learning path recommendation module, the broadened learning recommendation module, the optimized course recommendation module and the multimodal data module;

[0032] The personalized learning recommendation module provides personalized course recommendations to users based on their historical learning data;

[0033] The learning path recommendation module provides course recommendations from entry-level to advanced levels based on the user's current learning progress and skill level;

[0034] The broadened learning recommendation module recommends relevant courses based on the user's current learning content, which is used to broaden the current learning content and enrich the user's course options;

[0035] The optimized course recommendation module recommends courses based on user interests and needs;

[0036] The multimodal data module is used to integrate course content, professional settings, tutor information, student background, and interaction data to build a multimodal information database.

[0037] Furthermore, the recommendation control module includes a user-based collaborative filtering submodule; the user-based collaborative filtering submodule includes: finding previous users similar to the current user, and screening and recommending the previous users' selected courses to the current user.

[0038] Furthermore, the recommendation control module further includes an item-based collaborative filtering submodule; the item-based collaborative filtering submodule includes: recommending new courses similar to the user's previous interest courses to the user based on the user's previous interest courses.

[0039] Furthermore, the user-based collaborative filtering submodule further includes: constructing a user-item matrix, calculating user similarity, and generating a recommendation list.

[0040] Furthermore, the item-based collaborative filtering submodule further includes: constructing a user-item matrix, calculating item similarity, and generating a recommendation list.

[0041] Furthermore, the user similarity assessment content includes the user's education level, major, occupation, gender, age or interests.

[0042] Furthermore, the content of the examination of the similarity of the items includes the type, name, application field or place of use of the items.

[0043] Furthermore, the methods for calculating user similarity and item similarity include: cosine similarity, Pearson correlation coefficient and Jaccard similarity coefficient.

[0044] Furthermore, the user-based collaborative filtering submodule and the item-based collaborative filtering submodule are connected via a collaborative control submodule.

[0045] Furthermore, the user similarity assessment also includes academic background or career planning.

[0046] Example 2

[0047] The personalized learning recommendation module provides users with personalized course recommendations based on their historical learning data. Each learner's historical learning data in art education is unique, and the course recommendation system can provide personalized course recommendations based on their historical learning data. Such personalized recommendations not only enhance learners' engagement and satisfaction, but also help them more efficiently find courses that meet their needs, avoiding information overload.

[0048] A learning path recommendation module provides course recommendations from entry-level to advanced levels based on the user's current learning progress and skill level. Art education encompasses a variety of forms, including music, dance, drama, visual arts, design, and film and television. The learning paths and advanced courses in each field have a certain logical relationship. The course recommendation system can provide course recommendations from entry-level to advanced levels based on the student's learning progress and skill level. Through reasonable course planning, learners can avoid leapfrogging and gain a more systematic and gradual learning experience. For example, if a learner is new to piano, they can obtain a course path from basic to advanced through the recommendation system, avoiding the frustration of jumping directly into advanced courses.

[0049] The Broadening Learning Recommendation Module recommends relevant courses based on the user's current learning content, broadening their current learning content and enriching their course options. If a student has made good progress in a certain type of art course, the system can recommend other courses in similar fields or styles to help them broaden their artistic horizons. This cross-disciplinary recommendation helps cultivate students' diverse artistic abilities and stimulates their interest in continuous exploration. For example, after learning basic piano lessons, the system can recommend that students try music theory, composition, or other instrumental studies.

[0050] An optimized course recommendation module recommends courses based on user interests, skill levels, progress, learning goals, and needs. Art education often encompasses a variety of creative forms and styles. This recommendation system can address students' diverse needs by providing a rich selection of courses tailored to their interests, skill levels, progress, learning goals, and needs. Whether it's traditional art, modern art, or interdisciplinary art forms, course recommendations help students find the content they're most interested in, while avoiding the limitations of a single learning resource.

[0051] On online art education platforms, students are faced with a vast array of course options. Without a recommendation system, learners often spend considerable time browsing and sifting through courses, potentially leading to poor learning outcomes due to excessive choices, duplicate content, or mismatched difficulty levels. A course recommendation system can precisely recommend the most relevant courses based on students' needs and interests, saving learners time in sifting and selecting, thereby improving learning efficiency. For example, a student may be interested in digital painting but unfamiliar with the relevant software. An optimized course recommendation system can recommend a range of courses, from basic software operations to advanced painting techniques, eliminating the need for students to waste time and effort on their own.

[0052] For online art education platforms, course recommendation systems can improve user engagement and engagement. Through precise recommendations, the platform not only maintains the engagement of existing users but also attracts new users through word-of-mouth. If the recommendation system can accurately match students' needs and interests, user satisfaction will be greatly improved, thereby boosting the platform's retention rate. For example, after a student has completed several courses on the platform, the system may recommend new art forms or related courses. This can strengthen the student's long-term learning motivation and encourage further course purchases and participation.

[0053] Modern art education increasingly emphasizes interdisciplinary collaboration, with the integration of art with technology and the humanities on the rise. Course recommendation systems can broaden their learning recommendation modules to recommend relevant courses, guiding students to discover connections between different disciplines and engage in cross-disciplinary learning, thereby improving their overall artistic literacy. For example, a recommendation system could recommend courses related to photography, digital design, and other subjects to students studying visual arts, inspiring their cross-disciplinary creativity.

[0054] Course recommendations in online art education not only provide students with a more personalized and flexible learning experience, but also optimize the platform's teaching resources and user engagement. By making recommendations based on students' interests, historical behavior, and learning goals, educational platforms can significantly improve learning efficiency and student satisfaction. Furthermore, with technological advancements, recommendation systems that incorporate multimodal learning data (such as video, audio, and text) will further enhance the accuracy and diversity of recommendations, promoting the sustainable development of online art education.

[0055] Example 3

[0056] (1) Implementation of recommendation tasks - collaborative filtering recommendation.

[0057] Collaborative filtering is a widely used technology in recommendation systems. It leverages historical user behavior data to predict items that users might be interested in. Collaborative filtering algorithms are broadly classified into two categories: user-based collaborative filtering and item-based collaborative filtering.

[0058] User-based collaborative filtering: recommend items that similar users like to the user.

[0059] The user-based collaborative filtering submodule is based on the user-based collaborative filtering; the user-based collaborative filtering submodule includes: finding previous users similar to the current user, and recommending the previous user's selected courses to the current user. This algorithm finds other users who are similar to the target user. The user similarity is examined based on the user's education, major, occupation, gender, age, and interests, and then recommends items that these similar users like to the target user. This method believes that if user A and user B are similar, then user A may like the items that user B likes. The steps to implement this algorithm usually include constructing a user-item matrix, calculating user similarity, and generating a recommendation list. For example Figure 1 shown.

[0060] User-based collaborative filtering recommends related courses or majors by finding similar user groups. For similar users, the recommendation system automatically pushes the majors or courses they have selected.

[0061] Item-based collaborative filtering: recommend items similar to items that the user has liked before.

[0062] The item-based collaborative filtering submodule is based on the item-based collaborative filtering; the item-based collaborative filtering submodule includes: recommending new courses similar to the user's previous interest courses based on the user's previous interest courses; different from user-based collaborative filtering, this algorithm finds other items similar to the target item by analyzing the user's preferences for different items, and then recommends these similar items to users who like the target item. The content of the item similarity examination includes the type, name, application field or place of use of the item. The core of this method is that if two items are liked by many of the same users, then there is a certain similarity between the two items, so one item can be recommended to users who like the other item. Figure 2 shown.

[0063] Item-based collaborative filtering is more suitable for recommendation scenarios where items are highly similar. For example, in summer camp and short-term study abroad recommendations, if a user has previously selected a specific summer camp, the system will recommend other similar programs based on historical user selection data. This type of recommendation method can help users discover new but similar experience opportunities when making their choices.

[0064] When implementing collaborative filtering algorithms, similarity calculation is a key step. Commonly used similarity calculation methods include cosine similarity, Pearson correlation coefficient, and Jaccard similarity coefficient.

[0065] Jaccard similarity coefficient:

[0066]

[0067] Cosine similarity:

[0068]

[0069] Pearson correlation coefficient:

[0070]

[0071] The advantages of collaborative filtering algorithms include their ability to provide personalized recommendations and handle large datasets. However, they also have some drawbacks. For example, recommendations for new users or items may not be ideal due to a lack of sufficient historical behavioral data to support recommendations. Furthermore, collaborative filtering algorithms rely heavily on user behavioral data and may require regular updates to reflect changes in user interests.

[0072] In practical applications, to improve the accuracy and diversity of recommendations, user-based collaborative filtering and item-based collaborative filtering algorithms can be combined. Furthermore, various factors, such as time and user personalization, can be incorporated to adjust the recommendation strategy. The user-based collaborative filtering submodule and the item-based collaborative filtering submodule are connected via a collaborative control submodule. Through the collaborative control module, the user-based collaborative filtering submodule and the item-based collaborative filtering submodule can better perform course recommendations.

[0073] (2) Application scenarios of user-based collaborative filtering and item-based collaborative filtering.

[0074] On the one hand, because collaborative filtering makes recommendations based on user similarity, it possesses a stronger social nature. Users can quickly learn what others with similar interests are currently liking. Even if a particular interest was previously outside their own scope, it is possible to quickly update their recommendation list based on their friends' updates. This makes it ideal for news recommendation scenarios. Because news interests are often scattered, the timeliness and popularity of news are often more important attributes than user preferences for different news items. Collaborative filtering is therefore ideal for discovering hot topics and tracking their trends.

[0075] On the other hand, item-based collaborative filtering is more suitable for applications where interest changes are relatively stable. For example, in Amazon's e-commerce scenario, users tend to look for a certain type of product within a certain period of time. In this case, using item similarity to recommend related items is in line with user motivation. In Netflix's video recommendation scenario, users' interests in watching movies and TV series are often relatively stable, so using item-based collaborative filtering to recommend videos of similar styles and types is a more reasonable choice.

[0076] Therefore, using item-based collaborative filtering can achieve better recommendation results in tasks such as recommending majors, courses, and summer camps.

[0077] (3) Advantages of collaborative filtering algorithms in cross-domain recommendations.

[0078] The multimodal data module utilizes this multimodal data by integrating multimodal information such as course content, program settings, instructor information, student background, and interaction data. The collaborative filtering algorithm can provide more accurate and personalized recommendations for different users. For example, by combining a user's behavioral data in a specific subject area with their participation in specific courses, the recommendation system can not only make course recommendations based on item similarity, but also provide more customized recommendations based on user personal information such as education, career goals, and learning objectives.

[0079] Combining expert course and practice camp recommendations, high-level course recommendations (such as master classes and observation courses) can be accurately matched to the user's academic background, career plans, and historical choices by comparing them with the most appropriate course. Furthermore, short-term visiting studies and practice camps, as extensions of the curriculum, can provide users with academic practice opportunities. These activities often rely on item similarity for recommendation.

[0080] (4) Solve the cold start problem of collaborative filtering

[0081] In collaborative filtering recommendation systems, the cold start problem is often a challenge, especially for new users, new courses, or new projects. The following are some common cold start strategies that are particularly suitable for recommendation systems in the education field:

[0082] Leverage user registration information: We collect basic information provided by users during registration (such as age, gender, and educational background) and use this information to provide preliminary recommendations. For academic fields, we can recommend relevant courses based on the user's subject area and academic level.

[0083] User-defined interests: Allow users to proactively select the majors or course areas of interest, providing a personalized recommendation experience for new users. For short-term visiting studies and internship camps, recommendations can be made based on the user's voluntarily selected areas of interest.

[0084] Integrate user behavior data: Even during the cold start phase, the system can capture users' interests by analyzing their browsing history, search keywords, and other behavioral data. Multimodal recommendations, in particular, rely not only on click data but also on user interaction data across text, video, or other resources in different fields.

[0085] Therefore, this system is equipped with a user information registration module, a user-defined interest module and a user behavior data module. The user information registration module, the user-defined interest module and the user behavior data module are all connected to the recommendation control module. The user himself enters the user's registration information, user interest information and user behavior data information, and the cold start problem of collaborative filtering is solved through the linkage of multiple modules including the user information registration module, the user-defined interest module, the user behavior data module and the recommendation control module.

[0086] The embodiments of the present invention are described in detail above. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core ideas. At the same time, for those skilled in the art, according to the ideas of the present invention, there may be changes in the specific implementation methods and application scopes. In summary, the contents of this specification should not be understood as limiting the present invention.

Claims

1. A smart art education resource recommendation system for all-scenario teaching applications, characterized by: The system includes: a recommendation control module, a personalized learning recommendation module, a learning path recommendation module, a broadened learning recommendation module, an optimized course recommendation module and a multimodal data module; the personalized learning recommendation module, the learning path recommendation module, the broadened learning recommendation module, the optimized course recommendation module and the multimodal data module are all connected to the recommendation control module; The recommendation control module makes course recommendations based on the multimodal data module through the personalized learning recommendation module, the learning path recommendation module, the broadened learning recommendation module, the optimized course recommendation module and the multimodal data module; The personalized learning recommendation module provides personalized course recommendations to users based on their historical learning data; The learning path recommendation module provides course recommendations from entry-level to advanced levels based on the user's current learning progress and skill level; The broadened learning recommendation module recommends relevant courses based on the user's current learning content, which is used to broaden the current learning content and enrich the user's course options; The optimized course recommendation module recommends courses based on user interests and needs; The multimodal data module is used to integrate course content, professional settings, tutor information, student background, and interaction data to build a multimodal information database.

2. The recommendation system according to claim 1, characterized in that The recommendation control module includes a user-based collaborative filtering submodule; the user-based collaborative filtering submodule includes: finding previous users similar to the current user, screening and recommending the previous users' selected courses to the current user.

3. The recommendation system according to claim 2, characterized in that The recommendation control module further includes an item-based collaborative filtering submodule; the item-based collaborative filtering submodule includes: recommending new courses similar to the user's previous interest courses to the user based on the user's previous interest courses.

4. The recommendation system according to claim 3, characterized in that The user-based collaborative filtering submodule further includes: constructing a user-item matrix, calculating user similarity, and generating a recommendation list.

5. The recommendation system according to claim 4, characterized in that: The item-based collaborative filtering submodule further includes: constructing a user-item matrix, calculating item similarity, and generating a recommendation list.

6. The recommendation system according to claim 5, characterized in that: The content of the user similarity assessment includes the user's education background, major, occupation, gender, age or interests.

7. The recommendation system according to claim 6, characterized in that: The content of the examination of the similarity of the items includes the type, name, application field or place of use of the items.

8. The recommendation system according to claim 7, characterized in that: The methods for calculating user similarity and item similarity include: cosine similarity, Pearson correlation coefficient and Jaccard similarity coefficient.

9. The recommendation system according to claim 8, characterized in that: The user-based collaborative filtering submodule and the item-based collaborative filtering submodule are connected via a collaborative control submodule.

10. The recommendation system according to claim 9, characterized in that: The user similarity assessment also includes academic background or career planning.

Images