Analyzing and Improving Object Oriented Software Design using CK Metric and Neural Network

Object Oriented Software Engineering (OOSE) is used to build a new quality software system by taking the help of existing classes. Chidamber and Kemerer (CK) metrics is calculated on OOSE model to analyze reusability. The research work presents a general model to analyze, evaluate and improve CK metric values for object oriented software using Matlab’s Neural Network Toolbox. Object Oriented Design Patterns are used to analyze various metrics and number of useful conclusions can be drawn by evaluation of these metrics that may effect on quality factors of OOSE. Various metrics have been calculated that affects the performance of Object Oriented Software Engineering and proposed a model that analyzes and improves CK metric by automating the calculation of CK metric values. Finally, assessment is performed of the theoretical and empirical evaluation procedures to determine the extent of producing useful findings. Here we are proposing a model that will use CK metric values of Object Oriented Design Patterns (obtained from Rational Rose) as input and Learning Vector Quantization (LVQ) & Back Propagation Neural Network (BPNN) for processing of input values in order to calculate and automate CK metric values. Finally, Mean Squared Error (MSE) and Peak Signal to Noise Ratio (PSNR) have been calculated to measure accuracy as well as improvement in CK metric values.