Identifying the Research Specializations from the Publications using Text Mining and Linear Discriminant Analysis

摘要

In the past century data mining is one of the research area exercised by many scientists for delineating information from data and delivered good results. We attempted to find a scientific method to calculate few parameters like coefficients of Linear Discriminates and tabled to identify and range out reach of a scientist to the specializations. Reasonable conclusions are derived, and a good relationship is found from the data using LDA classification method. Coefficients of Linear Discriminates were computed from the key word percentages that are matched in 10 separate domains. The keywords were collected from the Microsoft Academic Search web site. From the independent datasets of author specializations in various domains, we found that the results are up to 80% to 90% accuracy.