Software Fault Prediction Model Based on Adaptive Dynamical and Median Particle Swarm Optimization

Software quality prediction can play a role of importance in software management, and thus in improve the quality of software systems. By mining software with data mining technique, predictive models can be induced that software managers the insights they need to tackle these quality problems in an efficient way. This paper deals with the adaptive dynamic and median particle swarm optimization (ADMPSO) based on the PSO classification technique. ADMPSO can act as a valid data mining technique to predict erroneous software modules. The predictive model in this paper extracts the relationship rules of software quality and metrics. Information entropy approach is applied to simplify the extraction rule set. The empirical result shows that this method set of rules can be streamlined and the forecast accuracy can be improved.