Analogy-based effort estimation (ABE) is one of the efficient methods forsoftware effort estimation because of its outstanding performance andcapability of handling noisy datasets. Conventional ABE models usually use thesame number of analogies for all projects in the datasets in order to make goodestimates. The authors' claim is that using same number of analogies mayproduce overall best performance for the whole dataset but not necessarily bestperformance for each individual project. Therefore there is a need to betterunderstand the dataset characteristics in order to discover the optimum set ofanalogies for each project rather than using a static k nearest projects.Method: We propose a new technique based on Bisecting k-medoids clusteringalgorithm to come up with the best set of analogies for each individual projectbefore making the prediction. Results & Conclusions: With Bisecting k-medoidsit is possible to better understand the dataset characteristic, andautomatically find best set of analogies for each test project. Performancefigures of the proposed estimation method are promising and better than thoseof other regular ABE models
展开▼