Publication

Abstract : Patents and Research papers are published in various fields. These are stored in various conferences and journals database. If a user (researcher or any general user) want to search for any patent or research paper in any particular field, then there is lack of search criteria available for this. In this paper, we have used nearest neighbor algorithm with cosine similarity to categorize patents and research papers. In this paper, experimental results show that if a user want to search for the patent or research paper in any particular field or category, then user would get better results. The advantage of the approach presented in this paper is that the search area becomes very small and so waiting time of user to get answer of query reduces to a large extent. To take decision about category of particular research paper or patent, there have been a lot of research work but categorizing was not that much accurate. In this paper, we have calculated threshold based on the similarity of terms between query and research paper or patent. This proposed calculation of threshold value is not based on numerical values. So, this novel approach of threshold calculation categorize more accurately than previous research work.