A COMPARISON STUDY BETWEEN CONTENTBASED AND POPULARITY-BASED FILTERING VIA IMPLEMENTING A BOOK RECOMMENDATION SYSTEM

Recommender systems are systems that filter information and suggest items based on user profiles or preferences. As recommender systems are increasingly incorporated in our daily lives and users become more dependent upon these systems during their decision-making process in their daily life, researchers and developers are continuously looking for better approaches to improve the performance of recommender systems and alleviate existing problems in the systems. The cold-start is a common problem in recommendation systems. This work introduces a book recommendation website, which employs content-based and popularity-based techniques based on Google Trends. The recommendation algorithms used in the system are compared and their performance in cold-start scenarios are evaluated. It is found that content-based recommendation performs with lower precision and recall with higher personalization and coverage whereas popularity recommendation performs with higher precision and recall with low personalization and coverage. The results also show that recommendation algorithms implementing Google Trends are able to perform with higher precision and recall with a lower personalization score