We implemented the six classic software reliability models, i.e., JM, GO, MO, Schneidewind, Duane, LV models. And based on non-parametric and parametric bootstrap resampling methods, we improved these six modesl.
Requirements: Matlab
We take JM and its improved models to show how to use our implementations.
To run the orignal JM model:
[predict_data, parameters] = JM_predict(input_data);Note: the input_data are culmulated failire times (MTTFs), the predict_data are also model calculated culmulated failire times. If you want to convert from/to MTBFs, please use tbf2ttf.m/ttf2tbf.m.
To run the non-parametric bootstrap improved JM model:
[predict_data, parameters] = JM_predict_NPB(input_data);To run the parametric bootstrap improved orignal JM models:
[predict_data, parameters] = JM_predict_PB(input_data);The other models can be executed as above commands, we do not give each for brief.
Besides, we give 14 popular failure datasets (data1.mat - data14.mat) in verification of software reliability models.