Machine Learning Based Text Document Classification for E-Learning

Abstract

The number of e-learning websites as well as e-contents are increasing exponentially over the years and most of the time it become cumbersome for a learner to find e-content suitable for learning as the learner gets overwhelmed by the enormity of the content availability. The proposed work focus on evaluating the efficiencies of the different classification algorithm for the identification of the e-learning content based on difficulty levels. The data is collected from many e-learning web sites through web scraping. The web scraper downloads the web pages and parse to text file. The text files were made to run through many machine learning classification algorithms to find out the best classification model suitable for achieving the highest score with minimum training and testing time. This method helps to understand the performance of different text classification algorithms on e-learning contents and identifies the classifier with high accuracy for document classification.