PROJECT TITLE :

Analogy-based effort estimation: a new method to discover set of analogies from dataset characteristics

ABSTRACT:

Analogy-based mostly effort estimation (ABE) is one among the efficient strategies for software effort estimation because of its outstanding performance and capability of handling noisy datasets. Standard ABE models typically use the same range of analogies for all projects within the datasets in order to make sensible estimates. The authors' claim is that using same range of analogies may manufacture overall best performance for the whole dataset however not essentially best performance for every individual project. Therefore there's a need to higher perceive the dataset characteristics in order to get the optimum set of analogies for every project instead of using a static k nearest projects. The authors propose a replacement technique based on bisecting k-medoids Clustering algorithm to return up with the simplest set of analogies for each individual project before making the prediction. With bisecting k-medoids it's possible to raised perceive the dataset characteristic, and automatically find best set of analogies for every take a look at project. Performance figures of the proposed estimation method are promising and higher than those of alternative regular ABE models.