COMPOR Experiment

analysis.R

library(simpleboot)

taskGroup <- function(row) {
  task = row[[2]]
  
  if (task == 'T 1.01' | task == 'T 1.02' | task == 'T 1.03') {
    return("G1")
  }

  if (task == 'T 1.04' | task == 'T 1.05' | task == 'T 1.06') {
    return("G2")
  }
  
  if (task == 'T 1.07' | task == 'T 1.08' | task == 'T 1.09') {
    return("G3")
  }
  
  if (task == 'T 1.10' | task == 'T 1.11' | task == 'T 1.12') {
    return("G4")
  }
  
  if (task == 'T 1.13' | task == 'T 1.14' | task == 'T 1.15') {
    return("G5")
  }
  
  if (task == 'T 1.16' | task == 'T 1.17' | task == 'T 1.18') {
    return("G6")
  }
  
  if (task == 'T 1.19' | task == 'T 1.20' | task == 'T 1.21') {
    return("G7")
  }
  
  if (task == 'T 1.22' | task == 'T 1.23' | task == 'T 1.24') {
    return("G8")
  }

  if (task == 'T 1.25' | task == 'T 1.26' | task == 'T 2.01') {
    return("G9")
  }
  
  if (task == 'T 2.02' | task == 'T 2.03' | task == 'T 3.01') {
    return("G10")
  }

  return ("G11")
}

calculateIC <- function(x, out) {
  b1 <- one.boot(x, mean, R=2000, tr=.2)
  b2 = boot.ci(b1, type="perc")
  
  print(paste(b2$percent[4], b2$t0, b2$percent[5], sep = " "))
  
  out = rbind(out, c(b2$percent[4], b2$t0,  b2$percent[5]))
  return(out)
}

evaluateResponseVariable <- function(responseVariable, description, plotMin, plotMax) {
  groups = c("G2", "G3", "G4", "G5", "G6", "G7", "G8", "G9", "G10")
  xgroups = c(2:10)
  
  dfO <- data.frame(l = numeric(), m = numeric(), u = numeric())
  dfC <- data.frame(l = numeric(), m = numeric(), u = numeric())
  
  for (group in groups) {
    loopdata = data[data$TaskGroup == group, ]
    COMPOR = loopdata[loopdata$Tech == 'COMPOR' ,]
    OO     = loopdata[loopdata$Tech == 'OO' ,]
    
    dfO = calculateIC(OO[,responseVariable], dfO)
    dfC = calculateIC(COMPOR[,responseVariable], dfC)
  }
  
  plot(xgroups, c(plotMin, plotMin, plotMin, plotMin, plotMin, plotMin, plotMin, plotMin, plotMax), col='white', xlab = "Task groups", ylab = description)
  axis(1, at = xgroups)
  lines(xgroups, dfC[,2], lwd=1.5, col = 'blue')
  lines(xgroups, dfC[,1], lty = "dotted", col = 'blue')
  lines(xgroups, dfC[,3], lty = "dotted", col = 'blue') 
  polygon(c(rev(xgroups), xgroups), c(rev(dfC[,1]), dfC[,3]), col = rgb(0, 0, 1, 0.2), border = NA)
  
  lines(xgroups, dfO[,2], lwd=1.5, col = 'red')
  lines(xgroups, dfO[,1], lty = "dotted", col = 'red')
  lines(xgroups, dfO[,3], lty = "dotted", col = 'red')  
  polygon(c(rev(xgroups), xgroups), c(rev(dfO[,1]), dfO[,3]), col = rgb(1, 0, 0, 0.2), border = NA)
  
  legend('topright', c("COMPOR","OO"), lty=c(1,1), lwd=c(2,2), col=c("blue","red"))
}


data <- read.csv(file = "UFCG/comporSAC/COMPOR-results-2.csv", header = T, sep = ",")

#Normality tests

COMPOR = data[data$Tech == 'COMPOR' ,]
OO     = data[data$Tech == 'OO' ,]

shapiro.test(COMPOR$Time)
shapiro.test(COMPOR$LOC)
shapiro.test(COMPOR$Coverage)
shapiro.test(COMPOR$Compl)

shapiro.test(OO$Time)
shapiro.test(OO$LOC)
shapiro.test(OO$Coverage)
shapiro.test(OO$Compl)

# Statistics tests
data$TaskGroup <- apply(data,1,taskGroup)

evaluateResponseVariable("Time", "Time in minutes", 0, 150)
evaluateResponseVariable("LOC", "Lines of code", 0,150)
evaluateResponseVariable("Coverage", "Test coverage (%)", 90, 101)
evaluateResponseVariable("Compl", "Complexity x 100", 115, 255)


#wilcox.test(COMPOR$Compl, OO$Compl, alternative = "two.sided")
#kruskal.test(Compl ~ Tech, data = data)

COMPOR-results-2.csv

COMPOR-results.csv