A novel hybrid based recommendation system based on clustering and association mining

In recent years, E-commerce had made a tremendous impact on the world. However before the emergence of E-commerce, individuals can't skim the information about the products within short time of the period, so therefore recommendation system was introduced. The principle point of the recommendation system is to prescribe the most appropriate items to the user. Many of the recommendation systems mainly use content based method, collaborative filtering method, demographic based method and hybrid method. In this paper, the major challenges such as “data sparsity” and “cold start problem” are addressed. To overcome these challenges, we propose a new methodology by combining the clustering algorithm with Eclat Algorithm for better rules generation. Firstly we cluster the rating matrix based on the user similarity. Then we convert the clustered data into Boolean data and applying Eclat Algorithm on Boolean data efficient rules generation takes place. At last based on rules generation recommendation takes place. Our experiments shows that approach not only decrease the sparsity level but also increase the accuracy of a system.

I. Introduction

In recent years with springing of the internet, large amount of information is available on the web, however it is difficult to handle all the information and that results into information overload, To overcome the problem of information overloading recommendation system was introduced. The main aim of the recommendation system is to recommend the most suitable items to the user. Now a days, users rely on recommendations from other people by spoken words, reference letters, news reports from news media, general surveys, travel guides and such others. Therefore recommendation system plays an important role in finding the best items. A recommendation system filtered the data through data analysis techniques which is useful to recommend the suitable items to the user. Recommendation systems work from a specific type of information filtering system technique that attempts to recommend information items (movies, TV program/show/episode, music, books, news, images, web pages, scientific literature etc.) or social elements (e.g. people, events or groups) that are likely to be of interest to the user [1]. The recommendation system also compares the user profiles and seeks to predict the ratings. With the help of Recommendation systems, filtering and sorting data can be easily done. Moreover the Recommendation system use opinions about the community of users and to determine content of interest using certain rules extractions. Recommendation systems are classified into 3 approaches which are collaborative, content-based or knowledge-based method to have a better recommendation.