ABSTRACT

Object Oriented Software development technique has become very popular and is being used by most of the software development companies. Although maintenance may be turn out to be easier for Object Oriented System. But it is unlikely that the maintenance burden will be completely disappearing. It has always been a crucial issue for software project developers. Still, maintenance consumes a large portion of software development cost. Therefore it is worthwhile to develop Object Oriented System keeping maintainability as a key issue in design phase. This paper investigates the use of clustering technique of data mining in maintenance of software system using object oriented metrics. The presented work evaluates the K-means clustering method by applying it to the commercial software system. The experimental work of software maintenance for the sample data is being simulated on Matlab.