In okamumu/Rsrat: A Software Reliability Assessment Tool with NHPP-Based Models

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)

Rsrat

Rsrat provides the package to evalute the software reliability from the fault data collected in the testing phase. Rsrat can use two types of data; fault-detection time data and its grouped data. The fault-detection time data is a sequence of time intervals of fault detection times (CPU time, etc). Also its grouped data is a sequence of the number of detected faults for each time interval (per a working day, per a week, etc). The reliability evaluation is based on the software reliability growth model with NHPP (non-homogeneous Poisson process).

Installation

You can install Rsrat from GitHub with:

install.packages("devtools")
devtools::install_github("SwReliab/Rsrat")

Alternatively, you can use Remote to install Rsrat from GitHub

install.packages("remotes")
remotes::install_github("SwReliab/Rsrat")

Example

This is an example of the estimation of software reliability growth models from a fault data (tohma).

### load library
library(Rsrat)

### load example data
data(dacs)

### tohma is a grouped data
tohma

### Esimate all models and select the best one in terms of AIC
(result <- fit.srm.nhpp(fault=tohma))

### Draw the graph 
mvfplot(fault=tohma, srms=result)
rateplot(fault=tohma, srms=result)

The second example illustrates the estimation for two specified models.

### All models in the package
srm.models

### Estimate two models and no select
(result <- fit.srm.nhpp(fault=tohma, srm.names=c("exp", "gamma"), selection=NULL))

### Draw the graph
mvfplot(fault=tohma, srms=result)

### Draw the graph (dmvf)
dmvfplot(fault=tohma, srms=result)
rateplot(fault=tohma, srms=result)

The third example shows the case where the fault data are fault detection data.

### fault-detection time data
#### Time intervals for all faults
#### The last value is a negative value, that indicates the time interval in which there is no fault detection after the last fault detection.
sys1

### Esimate
(result <- fit.srm.nhpp(time=sys1[sys1>=0], te=-sys1[sys1<0]))

### Draw the graph
mvfplot(time=sys1[sys1>=0], te=-sys1[sys1<0], srms=result)

The fourth example illustrates the case where the mvfs for all the models are drawn.

### Esimate and return all the estimated results
(result <- fit.srm.nhpp(fault=sys1g, selection=NULL))

### Draw the graph
mvfplot(fault=sys1g, srms=result)