Beyond Matrix Completion of the traditional Recommender System

Recommendation Systems apply Information Retrieval techniques to select the online information relevant to a given user. Collaborative Filtering (CF) is currently most widely used approach to build Recommendation System. CF techniques uses the user' behavior in form of user-item ratings as their information source for prediction. There are major challenges like sparsity of rating matrix and growing nature of data which is faced by CF algorithms. These challenges are been well taken care by Matrix Factorization (MF). In this paper we attempt to present an overview on the role of different MF model to address the challenges of CF algorithms, which can be served as a roadmap for research in this area.

User Modeling, Adaptation, and …, 2012

Users in recommender systems often express their opinions about different items by rating the items on a fixed rating scale. The rating information provided by the users is used by the recommender systems to generate personalized recommendations for them. Few recent research work on rating based recommender systems advocate the use of preference relations instead of absolute ratings in order to produce better recommendations. Use of preference relations for neighborhood based collaborative recommendation has been looked upon in recent literature. On the other hand, Matrix Factorization algorithms have been shown to perform well for recommender systems, specially when the data is sparse. In this work, we propose a matrix factorization based collaborative recommendation algorithm that considers preference relations. Experimental results show that the proposed method is able to achieve better recommendation accuracy over the compared baseline methods.

2017

With current projections regarding the growth of Internet sales, online retailing raises many questions about how to market on the Net. A Recommender System (RS) is a composition of software tools that provides valuable piece of advice for items or services chosen by a user. Recommender systems are currently useful in both the research and in the commercial areas. Recommender systems are a means of personalizing a site and a solution to the customer's information overload problem. Recommender Systems (RS) are software tools and techniques providing suggestions for items and/or services to be of use to a user. These systems are achieving widespread success in ecommerce applications now days, with the advent of internet. This paper presents a categorical review of the field of recommender systems and describes the state-of-the-art of the recommendation methods that are usually classified into four categories: Content based Collaborative, Demographic and Hybrid systems. To build ou...

The explosive growth and availability of data on the internet has caused information overload. Searching for a query is not easy in the sources of information available for the interest of an individual user. Collaborative filtering systems recommend items based upon opinions of people with similar tastes. Collaborative filtering overcomes some difficulties faced by traditional information filtering by eliminating the need for computers to understand the content of the items. Further, collaborative filtering can also recommend articles that are not similar in content to items rated in the past as long as like-minded users have rated the items. Collaborative filtering (CF) is one of the most frequently used techniques in personalized recommendation systems. But currently used CF techniques are based on item rating prediction. We proposed an improved personalized recommended CF algorithm. Hybrid recommender systems or content-boosted technologies are quickly produce high quality recommendations. We have explored content-boosted CF technique which analyzes the user-item matrix to identify relationships between different items, and then use these relationships to indirectly compute recommendations for users. In this paper we analyze different Memorybased CF and Model-based CF techniques. Finally, we experimentally evaluate our results and compare them. The testing results show that in most cases, the improved algorithm that we put forward can improve recommendation quality.

The ever-increasing number of E-commerce sites on the Internet has brought about information overload. This has made it difficult for consumers of certain products to find information about such products in an attempt to purchase products that best satisfies them. It has equally reduced the volume of product sales in the E-commerce domain. Hence, this paper proposes a personalized recommender system driven by fuzzy logic technique. The proposed system intelligently mine information about the features of laptop computers and provide professional services to potential buyers by recommending optimal products based on their personal needs. Fuzzy Near Compactness (FNC) concept is employed to measure the similarity between consumer needs and product features in order to recommend optimal products to potential buyers. Experimental result of the proposed system with 50 laptop computers consisting of Acer, Dell, HP, Sony, and Toshiba proves its effectiveness.

The explosive growth and availability of data on the internet has caused information overload. Searching for a query is not easy in the sources of information available for the interest of an individual user. Collaborative filtering systems recommend items based upon opinions of people with similar tastes. Collaborative filtering overcomes some difficulties faced by traditional information filtering by eliminating the need for computers to understand the content of the items. Further, collaborative filtering can also recommend articles that are not similar in content to items rated in the past as long as like-minded users have rated the items. Collaborative filtering (CF) is one of the most frequently used techniques in personalized recommendation systems. But currently used CF techniques are based on item rating prediction. We proposed an improved personalized recommended CF algorithm. Hybrid recommender systems or content-boosted technologies are quickly produce high quality recommendations. We have explored content-boosted CF technique which analyzes the user-item matrix to identify relationships between different items, and then use these relationships to indirectly compute recommendations for users. In this paper we analyze different Memorybased CF and Model-based CF techniques. Finally, we experimentally evaluate our results and compare them. The testing results show that in most cases, the improved algorithm that we put forward can improve recommendation quality.

In the era of e-commerce, the demand for custom based recommender systems has emerged. Collaborative Filtering is a technique used for building recommended systems. In this technique, a user-item sparse matrix is analysed to predict recommendations based on the ratings given for the same product by the neighbours. Neighbours are those users whose item purchases matches closely with active user. In this case study, a comparative performance of two algorithms, Memory based and Gradient descent matrix factorization is evaluated.

Procedia Computer Science, 2015

Collaborative filtering (CF) is widely used in recommendation systems. Traditional collaborative filtering (CF) algorithms face two major challenges: data sparsity and scalability. In this study, we propose a hybrid method based on item based CF trying to achieve a more personalized product recommendation for a user while addressing some of these challenges. Case Based Reasoning (CBR) combined with average filling is used to handle the sparsity of data set, while Self-Organizing Map (SOM) optimized with Genetic Algorithm (GA) performs user clustering in large datasets to reduce the scope for item-based CF. The proposed method shows encouraging results when evaluated and compared with the traditional item based CF algorithm.

2016

Recommender systems provide useful recommendations to a collection of users for items or products that might be of concern or interest to them. Its design depends on the data characteristics and the domain for which it will be applicable. The applications where recommender systems are used range from electronic products, books, movies, websites to even news articles and so on. The recommendations deal with predicting user's ratings that are yet to be consumed by users, based on the items already rated by them and recommend the top N items with the highest predicted ratings. Many algorithms try to improve the accuracy in predicting the top N items thereby improving the recommendation quality. This paper presents an overview of recommender systems and explores the various approaches to recommender systems.

2011

Recommender systems provide relevant items to users from a large number of choices. In this work, we are interested in personalized recommender systems where user model is based on an analysis of usage. Collaborative filtering and content-based filtering are the most widely used techniques in personalized recommender systems. Each technique has its drawbacks, so hybrid solutions, combining the two techniques, have emerged to overcome their disadvantages and benefit from their strengths. In this paper, we propose a hybrid solution combining collaborative filtering and content-based filtering. With this aim, we have defined a new user model, called userfeature model, to model user preferences based on items' features and user ratings. The user-feature model is built from the user item model by using a fuzzy clustering algorithm: the Fuzzy C Mean (FCM) algorithm. Then, we used the user-feature model in a user-based collaborative filtering algorithm to calculate the similarity between users. Applying our approach to the MoviesLens dataset, significant improvement can be noticed comparatively to the main CF algorithm, denoted as user-based collaborative filtering.