Recommender systems are widely used these days in e-commerce, for the purpose of personalized recommendation. Based on each user’s profile, previous purchase history, and online behavior, they suggest products which they are likely to prefer. For example, Amazon.com is using recommender systems for books. When a user logs-in to the system, it suggests books similar to previously bought ones by the user. In this paper we compare some previous work done on personalized recommendations system for web applications, and try to find out what lacked in these previous work.