Analogy-based estimation methods have been researched for deriving accurate effort estimate required to complete a project; however, there were few activities to develop efficient variable selection methods. In this paper, we propose Correlation Based Selection (CBS) excluding variables that do not correlate with the total effort. The proposed method needs few calculation amounts because it calculates only the correlation coefficients between the total effort and each variable in the data. We experimentally evaluated accuracy of the proposed method by using the data collected from a softwar...