A novel approach for unsupervised learning of software components

Abstract

Clustering and classification are two important tasks in data mining and machine learning. These tasks have various applications in other related areas of research such as software engineering, text mining, image processing, and bio-informatics. Clustering is an NP-Hard problem, i.e. there is no proved polynomial time algorithm that can cluster a given set of input instances. However, approaches for evaluating cluster quality exist in the literature. This paper gives a new approach for software component learning by introducing an incremental learning approach for component clustering. Experiments are conducted by applying proposed approach on synthetic dataset and results proved the importance of proposed approach in terms of execution time and memory consumed.