Using Aggloramative Clustering to Asses and Improve Software Reliability

Abstract

For the reduction of cost in software testing we propose a novel technique for testing and classifying methods based on clustering methods for classifying test cases for powerful and non-viable groups. This technique is based on data treatment obtained by pre-release of program while testing. Here we introduce 2 new clustering algorithms such as centroid and hierarchical based clustering. The test study expresses that the experiment bunching results can be distinguished viably with high review proportion and noteworthy rate exactness. The present paper tells about the presentation of clustering which move towards by comparing and investigating the factors like criteria coverage, features of construction and pre-release faults quality.