User Profile Based Recommender System for E Learning Course

Abstract

Massive Open Online Courses (MOOCs) have created a huge boom in the education sector. With the help of MOOC, the learners can easily communicate with the instructors present worldwide. MOOC platforms engage learners in a significant way and help them to grow in whichever field they are interested. Due to the pandemic of COVID-19, many universities opted for MOOC platforms for their survival but most learners and instructors faced many challenges. The novice learners were not able to opt for the relevant courses from these platforms. Every platform has its own recommender system and these systems only recommend courses from their platform. As a result, learners are only able to opt for courses from a single platform and if they want to opt for a course from another platform then they need to register themselves to new platforms. The main objective of this research is to create a single platform for learners from where they can search for courses from multiple platforms like Coursera, Udemy, EdX, Udacity, etc., and then that platform recommends courses according to their preferences. To study the learner s behavior, a user profile is created through which data is fetched and some data is also fetched from parsing their resumes present on social media platforms like LinkedIn to generate a dataset. Once the final dataset is created, the hybrid model is applied to the dataset and for validating the model, multi models are used viz Random, User-based collaborative filtering, Item-based collaborative filtering and Matrix factorization. To check the results, different data-size is used and fundamental measures i.e RMSE, Precision, Recall and F1- score are examined. From all the models, the best result obtained from user-based collaboration filtering on 6000 size of dataset. The values of RMSE in case of user- based collaboration filtering is 0.101, value of precision is 0.82 and the value of recall is 0.822. The dataset consists of Coursera, Udemy and Edx courses. Further to improve this model more number of e-learning p