ANALYSIS OF ON-LINE COURSES VIA STUDENT REVIEWS USING DATA MINING

Abstract

Over the last 15 years, educational technology has grown in significance. Currently, the educational technology umbrella encompasses a wide range of interesting online settings and subjects. Massive Open Online Courses (MOOCs) and Learning Analytics are two of the domain’s most significant new subjects. MOOCs excel in attracting a huge number of participants because they are free and available to everybody As a result, they may reach hundreds of thousands of people. Given how quickly this phenomenon has expanded, experts from many other fields have taken an interest in MOOCs. MOOCs have been shown to scale education across a wide range of subjects. Improvements in educational achievements, cost savings, and increased accessibility are just a few of the advantages they provide. The vast datasets generated by MOOC platforms need the use of specialized tools and approaches to fully examine them. This clearly shows the significance of learning analytics. MOOCs provide a wide range of learning analytics difficulties and strategies to consider. In this thesis, we have used a corpus of approx 14 lacs reviews for courses on coursera website. Preprocessing was carried out to cleanse the reviews by using certain approaches. On the cleaned data various Deep Learning and Machine learning Algorithms were utilised for determining various classification parameters and compared the results. We also propose Course Recommender Systems to recommend various courses to students based on reviews and ratings using various methodologies under courses main and sub categories.