Abstract
Quality assurance of software is an important problem for a software development enterprise. A quality prediction using metrics is performed by many enterprises. There is Rayleigh model who used that the defect extraction rate per the unit time is based on Rayleigh distribution as the method to do a quality prediction based on the defect extraction history of the project concerned. There is a problem that anyone have to establish the defaults when doing parameter presumption. The defect extraction rate isn't based on Rayleigh distribution by a project, and there is a problem that it can't be predicted. Therefore the defect density obtained from the past Project Performance and the method to do a quality prediction using the Weibull distribution were considered.
