Development of a recommender System for adaptive e-learning

Recent years have seen the web as an essential mean for Learning and Education, thanks to the almost infinite amount of information shared and to the explosion in development and adoption of e-Learning platforms that allow people to study any topic without the barriers of time, geography and physical participation. In addition to traditional learning content, online platforms allow user-centered approaches, creating an interactive and consequently very effective learning environment. The objective of the thesis is to develop an adaptive learning system for an Italian e-Learning Platform, leader on the market, being able to recommend an optimized learning path for each user. The developed system will be based on machine learning algorithms, which will learn from users’ performance and Learning characteristics – e.g. time spent learning a single topic, speed of improvement and learning abilities, test scores and completion times – in order to drive the user toward the next best new topic to study or the review on the most appropriate past topics to fill his/her knowledge gap.The focus will be mainly on mathematics courses, which are, currently, the most requested on the platform.