TU Wien:Statistik und Wahrscheinlichkeitstheorie UE (Bura)/Übungen 2020W/HW10.4
Zur Navigation springen
Zur Suche springen
Processors - Part 3[Bearbeiten | Quelltext bearbeiten]
Perform pairwise t-tests (α = 5%) and correct the p-values according to Bonferroni
- (a) How many pairs are to be tested?
- (b) Give all p-values. Which null-hypotheses are to be rejected prior Bonferroni correction?
- (c) Which null-hypotheses are to be rejected after Bonferroni correction?
Lösungsvorschlag von Friday[Bearbeiten | Quelltext bearbeiten]
--Friday Sa 30 Jan 2021 17:47:36 CET
# Statistics and Probability - HW #10
# Friday
# Duedate: 14.12.2020
# Problem 4 - Processors - Part 3
# Perform pairwise t-tests (α = 5%) and correct the p-values according to
# Bonferron.
load("temperatures.Rdata")
alpha <- 5
# Problem 4a)
# How many pairs are to be tested?
k <- length(temp)
m <- (k*(k-1))/2
result_4a <- m
result_4a
# Problem 4b)
# Give all p-values. Which null-hypotheses are to be rejected prior
# Bonferroni correction?
draw_ttest <- function(data, alpha) {
colnames <- c()
rownames <- c('p-value','color')
pvalues <- c()
colors <- c()
len <- length(data)
for (i in 1:(len-1)) {
for (j in (i+1):len){
colnames <- c(colnames, sprintf("%d/%d", i, j))
p <- t.test(data[[i]], data[[j]])$p.value
pvalues <- c(pvalues, p)
if (p < alpha) {
colors <- c(colors, 'red')
} else {
colors<-c(colors, 'green')
}
}
}
tmp = matrix(NA, length(rownames), length(colnames))
rownames(tmp) = rownames
colnames(tmp) = colnames
tmp[1,] <- pvalues
tmp[2,] <- colors
print(length(colors))
return (tmp);
}
result_4b <- draw_ttest(temp, alpha)
result_4b
# Problem 4c)
# Which null-hypotheses are to be rejected after Bonferroni correction?
alpha_star <- alpha/m
result_4c <- draw_ttest(temp, alpha_star)
result_4c