Software Effort Estimation Using Grey Relational Analysis with K-Means Clustering

Software effort estimation is described as a method of predicting the amount of person/months ratio to build a new system. Effort estimation is calculated in terms of persons involved per month for the completion of a project. During the launch of any new project into the market or in industry, the cost and effort of a new project is estimated. In this context, a numerous models have been developed to measure the effort and cost. This becomes a challenging task for the industries to predict the effort. In the present paper, a novel method is proposed called the Grey Relational Analysis (GRA) for estimating the effort of a particular project by considering the most influenced parameters. To achieve the same, one-way ANOVA and Pearson correlation methods are combined. Experimental results obtained with the help of clustering and without clustering by using the proposed method on the data set are presented. An attempt has been made to show the minimum error rate by using GRA for predicting the effort estimation on COCOMO 81 data set and clustered data set. The proposed method demonstrated better results compared to the traditional techniques used for estimation. The efficiency of the proposed system is illustrated through experimental results.