International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 08 | Aug 2024
p-ISSN: 2395-0072
www.irjet.net
Course Recommendation Using Machine Learning Richa Sapre1, Arti Panchal2, Harsh Raut3 , Vaibhav Zarapkar4, Dr. Sunil Wankhade5 1,2,3,4Student, MCT’s Rajiv Gandhi Institute Of Technology, Mumbai, Maharashtra, India.
5Professor & Head Of Department ,Dept. of Information Technology ,MCT’s Rajiv Gandhi Institute Of Technology,
Mumbai, Maharashtra, India. ---------------------------------------------------------------------***--------------------------------------------------------------------developing, and implementing an intelligent course Abstract – This paper describes a course recommendation
recommendation system using Python and Streamlit—a modern framework for creating interactive online applications. Our solution is built around the integration of powerful machine learning methods such as course similarity analysis, user profile, clustering (augmented with Principal Component Analysis, PCA), K-Nearest Neighbours (KNN), and Non-Negative Matrix Factorization (NMF). Using these algorithms, our system seeks to provide personalized course recommendations that are closely aligned with each learner’s specific preferences, interests, and learning profile. The major goal of our course recommendation system is to empower students by making tailored choices that improve their educational experience. To accomplish this, our algorithm first examines all course metadata, including subject categories, descriptions, difficulty levels, and prerequisites. This information is utilized to create a complete course- user interaction matrix that includes specific user behaviours such as course completion rates, user ratings, and frequency of involvement. To create accurate and relevant recommendations, our system employs machine learning techniques such as clustering and dimensionality reduction with PCA. User profiling is an important aspect of our system that aims to capture individual learner preferences and behaviours. Our method uses K-Nearest Neighbours (KNN) and Non- Negative Matrix Factorization (NMF) to identify comparable users based on their course interactions, and it extracts latent components from user-course interactions.
system created with Python and Streamlit for interactive visualization. To provide personalized course recommendations, the system uses machine learning algorithms such as course similarity analysis, user profiling, clustering (using PCA), K-Nearest Neighbours (KNN), and Non-Negative Matrix Factorization (NMF). The course recommendation process begins by analyzing course metadata and user interactions to create a course-user matrix. This matrix is used to cluster related courses based on criteria such as subject, difficulty level, and prerequisites. Principal Component Analysis (PCA) is used to minimize dimensionality while keeping significant information, resulting in more efficient course clustering.
User profiling is then used to determine individual preferences and learning behaviours. The technology uses K-Nearest Neighbours (KNN) to identify comparable users based on how they engage with courses. Non-Negative Matrix Factorization (NMF) isused to extract latent components from user-course interactions, allowing for personalized suggestions based on the user’s preferences and learning history. The course recommendation systemis integrated into a user-friendly web interface called Streamlit, which allows users to enter preferences, browse recommended courses, and provide comments. User research and comparative analysis show that the system is effective at offering relevant and diverse course suggestions, which improves the learning experience across different topics and ability levels.
This personalized approach enables our system to adjust recommendations to each user’s unique learning history and interests. Furthermore, the course recommendation system is linked into a user- friendly web interface called Streamlit, which allows learners to enter their preferences, explore recommended courses, and provide comments, allowing the recommendation process to be refined over time. Our course suggestion system is more than just convenient; it aims to improve the educational experience by creating a more personalized and interesting learning environment. By bridging the gap between learners and the vast universe of online educational resources, our approach intends to improve course discovery, encourage study of new disciplines, and, eventually, improve the overall efficacy and enjoyment of learning. Through thorough evaluation and comparison analysis against baseline algorithms, we hope to demonstrate our system’s usefulness and utility in offering ac- curate, diversified, and personalized course recommendations targeted to individual learners across
Key Words: Course Recommendation System, Machine Learning, Streamlit, User Profiling, Clustering, KNN, NMF, PCA.
I.
INTRODUCTION
The emergence of online platforms offering a wide range of educational courses across many areas and disciplines has revolutionized the learning landscape in the era of digital education. This democratization of information provides learners with unprecedented options, but it also adds the task of navigating a large catalogue of courses to discover ones that best meet their own requirements and interests. This problem emphasizes the necessity of personalized course recommendation systems, which use powerful machine learning algorithms to provide targeted recommendations and speed upthe course discovery process. Our project aims to address this difficulty by designing,
© 2024, IRJET
|
Impact Factor value: 8.226
|
ISO 9001:2008 Certified Journal
|
Page 141