Since the late 20 th century, the number of Internet users has noticeably increased. Recently, the number of Internet queries and the quantity of information available on the web has increased drastically. A large amount of new information is uploaded to the Web on a daily basis. However, search results are not always reliable due to the vast amount of data available on-line. As a result, users often have to repeat their searches in order to find exactly what they are looking for. To remedy this, some researchers have suggested recommendation systems. Since a recommendation system proposes information relevant to a particular query, users no longer need to repeat a search to obtain desired data. In the Web 2.0 era, recommendation systems often rely on the collaborative filtering approach, which is based on user information such as age, location, or preference. However, the traditional approach is affected by the cold-start and spar-sity problems. The reason for these problems is the fact that the traditional system requires user information to operate properly. In this paper we address the sparsity problem associated with the current recommendation systems. We also suggest a new recommendation system approach and compare the performance of the proposed method with that of the traditional approach.