Recommender Systems Using Category Correlations Based on WordNet Similarity Recently, many internet users are not only information consumers but also information providers. There is lots of information on the Web and most people can search information what they want through the Web. One problem of the large number of data in the Web is that we often spend most of our time to find a correct result from search results. Thus, people start looking for a better system that can suggest relevant information instead of letting users go through all search results: We call such systems recommendation systems. Conventional recommendation systems are based on collaborative filtering (CF) approaches. The CF approaches have two problems: sparsity and cold-start. Some researchers have studied to alleviate the problems in CF approaches. One of them is the recommendation algorithm based on category correlations. In this study, researchers utilize genre information in movie domain as category. They have drawn genre correlations using genre counting method. This approach can alleviate the user-side cold-start problems, however, there exists one problem that extensions of the approach are less likely. If a domain has singular category, then we cannot apply previous approaches. It means that we cannot draw category correlations. Because of this reason, we propose a novel approach that can draw category correlations for not only multiple categories but also singular one. We utilize word similarities provided by WordNet.