#Estatística: Aplicação ao Sensoriamento Remoto - SER204, INPE, 2023 #http://www.dpi.inpe.br/~camilo/estatistica #Determinando o valor-P e o KDcrítico pata o Teste de Kolmogorov-Smirnov para 2 amostras regA<-c(81,78,61,89,69,58,64,84,89,83,88,56,87,95,75) regB<-c(56,55,76,54,83,97,85,66,78,80,61,69,71,55,91) min<-min(regA,regB) max<-max(regA,regB) dif<-rep(0,max-min+1) for (i in min:max) dif[i-min+1]<-abs(length(which(regA <= i))-length(which(regB <= i))) KDobs<-max(dif) KDobs regAB<-c(regA,regB) n<-10000 pKD<-rep(0,16) ValorP<-0 for (k in 1:n) { regAB<-sample(regAB) regAt<-regAB[1:length(regA)] regBt<-regAB[length(regA)+1:length(regAB)] dif<-rep(0,max-min+1) for (i in min:max) dif[i-min+1]<-abs(length(which(regAt <= i))-length(which(regBt <= i))) KD<-max(dif) if (KD <= KDobs) ValorP<- ValorP+1 pKD[KD+1]<- pKD[KD+1]+1 } pKDcum <-rev(cumsum(rev(pKD))/n) KDcrit<-min(which(pKDcum < 0.05))-1 KDcrit ValorP<-ValorP/n ValorP