A Personalized Recommendation System for Academic Events

Abstract

Academic events are growing in numbers worldwide annually for researchers to discuss their work. The research on recommendation systems in academic domains has high significance for researchers. The classical approach to the recommender system uses content-based and collaborative filtering that tends to produce poor results. The focus of the study is to determine the factors involving the selection of academic events and create a user-based personalised recommender system for academic events. A survey will be conducted to identify the factors affecting the choice of events. The system will filter the results of the events using a matching matrix by conducting a factor analysis and receiving input to find the most relevant academic events from the database. The study's approach evaluates the result based on the pre-processed data and the similarity measures between a similar user (Top-n) and an active user for events with a higher probability of participation. The weighted average of the neighbour’s ratings will be generated for the predictions of the events. The study’s outcome will prove that the personalised recommendation system is better than the classical approach in finding the most relevant events. The recommendation system can be optimised in domains.