A clustering based on optimization for object oriented quality prediction

In this paper software metric estimation has been established through object oriented parameters measures and quality estimation techniques. First the dataset has been prepared based on the object oriented programming. The four object oriented parameters have been considered for the experimentation are class, object, inheritance and dynamic memory allocation (DMA). The data is pre-processed based on these parameters and then cluster for the specified grouping. For the software strength analysis the data filtering has been done by the chi square testing method. Then software metrics F-Measure, Odd Ratio and Power have been used for the strength analysis. Finally by using binary particle swarm optimization (BPSO) quality estimation measures are optimized and retrieved the final accuracy based on the metrics input set. The result after BPSO shows better accuracy as comparison to the previous methods.