Item-item collaborative filtering

Abstract: Item-based collaborative filtering is one of the most popular techniques in the recommender system to retrieve useful items for the users by finding the correlation among the items. Traditional item-based collaborative filtering works well when there exists sufficient rating data but cannot calculate similarity for new items, known as a cold-start problem. Usually, for the lack of rating data, the identification of the similarity among the cold-start items is difficult. As a result, existing techniques fail to predict accurate recommendations for cold-start items which also affects the recommender system’s performance. In this paper, two item-based similarity measures have been designed to overcome this problem by incorporating items’ genre data. An item might be uniform to other items as they might belong to more than one common genre. Thus, one of the similarity measures is defined by determining the degree of direct asymmetric correlation between items by considering their association of common genres. However, the similarity is determined between a couple of items where one of the items could be cold-start and another could be any highly rated item. Thus, the proposed similarity measure is accounted for as asymmetric by taking consideration of the item’s rating data. Another similarity measure is defined as the relative interconnection between items based on transitive inference. In addition, an enhanced prediction algorithm has been proposed so that it can calculate a better prediction for the recommendation. The proposed approach has experimented with two popular datasets that is Movielens and MovieTweets. In addition, it is found that the proposed technique performs better in comparison with the traditional techniques in a collaborative filtering recommender system. The proposed approach improved prediction accuracy for Movielens and MovieTweets approximately in terms of 3.42% & 8.58% mean absolute error, 7.25% & 3.29% precision, 7.20% & 7.55% recall, 8.76% & 5.15% f-measure and 49.3% and 16.49% mean reciprocal rank, respectively.

Abstract: The one stop to all problems is the Internet. But finding relevant information is difficult. The interest of the user lies in different forms of information content such as images, text, audio, or videos. The recommendation system is a process of information filtering that helps users to find better products, financial plans, and other related information by personalizing the suggestions. There are different recommendations techniques such as collaborative filtering, demographic recommendation, knowledge-based recommendations, content-based recommendation, and utility-based recommendation system. These techniques fail to eliminate the drawbacks such as data sparsity, new user cold start problem, new item cold start problem, overspecialization, and shilling attacks. In today’s generation, saving income is very important. In this work, a recommendation system for financial planning is proposed. Here, the idea is to modify the recommendation process to improve the recommendations in the best possible way. The above-mentioned drawbacks are eliminated using hybrid approach. In the hybrid approach, the techniques of collaborative filtering, i.e., user–user and item–item similarity along with demographic filtering, are combined. The experimental result is evaluated using performance metrics precision and recall. An ROC curve is used for evaluating the system.

Abstract: Recommender Systems, these days, are no longer personal recommender systems, rather they are group recommender systems which list out recommendations for a group of users. Also, they are an integral part of today's web sites (mainly shopping, search engine etc.) who want to keep track of their users' preferences. Although we cannot build a recommender system for every individual, we can build a recommender system which considers the preferences of a group of users. Hence the concept of group recommendation is even more difficult. Recent researches transpire that there is no efficient group recommendation technique available in the market and also the techniques developed till date are good for individual application or web sites. In this paper we have proposed a new approach to group recommender system using collaborative filtering technique which is one of the two techniques of building recommender systems. In our proposed method we have combined the features of item-item collaborative filtering as well as user-user collaborative filtering to make efficient group recommendation by making homogeneous groups. We have also made a sincere attempt to list out the precision of our group recommender system by using the movie lens data set which is mostly used worldwide for recommender system testing.

Abstract: Recommender System is a technique which is used to recommend an item or product to a user based on the user's preference'. Collaborative filtering is an approach that is vastly used in recommender systems. Item-item-based collaborative filtering is a collaborative filtering recommender system technique where the user got the recommendation based on the similarity among the item ratings. Here, we present an approach where we calculate the similarity among the items based on the genre of items. Any item may belong to more than one genre or category. Based on items propensity to a specific genre or category we propose a new item-item-based similarity metric and a little improvement in prediction method that can efficiently compute the ratings and provide more accurate recommendation compare to the state-of-art works. Our model addresses the problem of the cold start since traditional similarity model takes the user ratings into account whereas our model can calculate the similarity based on item genre or category among them. We also show the extensive simulation results based on sparsity and other recommender system evaluation techniques. We also distinguish that our result outperforms than the traditional collaborative filtering recommender systems.