作者: Hiroyuki Okamura , Tadashi Dohi
DOI: 10.1109/HASE.2014.22
关键词:
摘要: This paper proposes a novel framework of software reliability growth models with metrics. Our approach is to integrate classical Poisson-regression-based fault prediction non-homogeneous Poisson process based models. The remarkable feature this handle time series data detections and metrics for number modules at the same time. In paper, we present modeling that combines models, also develop an efficient algorithm estimate model parameters on EM (expectation-maximization) algorithm. numerical experiments, by comparing proposed both discuss effectiveness using from viewpoints estimation prediction.