Analogy based Software Effort Estimation and Accuracy Improvements
R. Kamalraj, Dean · 2013
Rathi. J, Kamalraj. R, Karthik. S Abstract— Accurate software development estimation in the software development is essential in the software project management practices and it will be affected completely by the irrelevant and the misleading information's. Impact of an excessive budget and project schedule concentrate becomes a rule in the today's IT industry. So that the software effort estimation is being a big challen ge to software engineers for give effective management strategies. For software effort or cost estimation, analogy-based problem solving is extensively accepted by the software experts. Similarity measures play a vital role in the analogy based effort estimation. Distance between the project being estimated and the historical or the existing projects gives the final decision to generate effort estimation. The nearest neighbors of a project give the immutable conclusions for the present project. Here we taking the Desharnais data set as a sample data set and proceeding the Subset selection, Feature weighting, Discretization, Adaptation mechanism, Analogy selection, BRICH Clustering was applied in this system to give the performance accuracy and performance comparison with the Greedy Agglomerative Clustering(GAC), Neural Networks and the BRICH Clustering. Using the BRICH clustering gives better results to compare the other techniques.