Recommender Systems Using Item Sequences and Ranking Mechanisms

Now days, huge amount of information is available on the web and it is difficult to user to find relevant information. Recommender systems are very useful to give suggestions. Recommendation quality depends on number of criteria, diversity is one major criterion i.e. recommending less popular and more personalized items. The proposed system uses collaborative filtering and content based recommendation techniques. In collaborative filtering, items rating predictions are calculated. Based on these rating predictions and items popularity, ranking algorithm generates the recommendations with more diversity while maintaining the accuracy. Consumer/ manufacturer oriented ranking mechanism ranks the reviews of items which are recommended by collaborative filtering and content based techniques. Thus user will get diverse recommendations with useful reviews of those recommendations. Also depending upon past sequence of user purchases, recommender system will recommend items with the help of learning techniques.