News clustering using relations between named entities

News clustering is the process of bringing together news articles which are similar in content. By the growth of the Internet, it is an important problem to cluster distributed and constantly updated news. There are a lot of studies for this problem. Google News is one of the major works to cluster news articles. After collecting thousands of news from the sources in the world, it provides user those that relate to each other. ?Bag of wrods? is the most well-known method for solving news clustering. This method makes clustering operation by looking at the frequency of words in the document. Recently, there are some approaches for clustering documents by the information extracted from news sources such as Wikipedia and Wordnet. It is observed that these kind of methods gives beter results than ?bag of words?. With the advent of Semantic Web, Internet is no longer document network (web of document) and becomes data network (web of data). This technology allows us to reach information more quickly on the internet. Linking Open Data, is a project developed for the publication and interconnection of open data with the approach of the semantic web. Through this project the semantic web data network and linked data is growing day by day. Linked data was used in several information extraction and artifical intelligence application.In this thesis, we developed a new approach to the problem of news clustering by using semantic relations of the named entites in the document. According to the experiments, it is obtained that, our approach shows beter results than ?bag of word? approach.Keywords: Named Entity, News Clustering, Semantic Web, Linked Data