Measuring Changeability of Object Oriented Classes and Packages Using Hybrid Probabilisticlsa and SVM-ELM Model for Fault Prediction Viji*, Kumar Raj**, Dr. Duraisamy S.*** *Assistant Professor, Department of Computer Science and Engineering, SVS College of Engineering, Coimbatore, Tamil Nadu, India **Professor and Head, Department of Computer Applications, Sri Krishna College of Engineering and Technology, Coimbatore, India ***Department of Computer Applications, Sri Krishna College of Engineering and Technology, Coimbatore, India Online published on 14 October, 2016. Abstract The quality of a product is very important in the development of any software product as the quality resembles the satisfaction of the customer without compromising the profit of the organization. The object oriented coding is used in the software product development. The detection of the faults in these codes is very important in enhancing the quality which must be detected in prior. Hence in this paper, the fault prediction is performed in software class and package changeability measurements which are mined based on the three categories of the measurements such as Changeability directory measurement, coupling measurement and semantic measurement. While the changeability directory and coupling measurements are calculated, the Hybrid Probabilistic Latent Semantic Analysis (H-PLSA) method with optimized weighted scheme using QPSO is used to compute the semantic measurements from the past history. The prediction of the faults is done by classifying them using SVM-ELM classifier. Thus the software class and package changeability measurements can be obtained and the faults are detected more accurately. The performance evaluation results also show that the proposed methodology provides better performance than the existing methodologies. Top Keywords Class and package changeability, Probabilistic Latent Semantic Analysis, Quantum Particle Swarm optimization, Support Vector Machine, Extreme Learning Machine. Top |