computeNormIndex
## function (my.data)
## {
## my.data <- data.frame(my.data)
## d <- ncol(my.data)
## indices <- NULL
## for (i2 in 1:(d - 1)) {
## for (i1 in (i2 + 1):d) {
## indices <- rbind(indices, c(i1, i2))
## }
## }
## Tstat <- nrow(my.data) * computeT(my.data, indices)
## vd <- getVd.sam(my.data)
## alphas <- eigen(vd)$values
## alphas <- alphas[alphas > 1e-05]
## normindex <- (Tstat - sum(alphas))/sqrt((2 * sum(alphas^2)))
## return(normindex)
## }
Underlying correlation matrix \(p=5\): And \(p=10\)
We simulated 30 samples in each conditions.
## # A tibble: 6 x 4
## # Groups: p [2]
## p variable mean sd
## <dbl> <fct> <dbl> <dbl>
## 1 5 normal -0.0367 0.865
## 2 5 vitaclayton 3.36 2.61
## 3 5 vitajoe 5.54 3.58
## 4 10 normal 0.0128 0.726
## 5 10 vitaclayton 4.42 3.72
## 6 10 vitajoe 6.12 4.75
## # A tibble: 6 x 4
## # Groups: n [2]
## n variable mean sd
## <dbl> <fct> <dbl> <dbl>
## 1 500 normal 0.0377 0.741
## 2 500 vitaclayton 2.36 1.81
## 3 500 vitajoe 3.67 2.49
## 4 1000 normal -0.0616 0.850
## 5 1000 vitaclayton 5.41 3.64
## 6 1000 vitajoe 7.99 4.47
## # A tibble: 9 x 4
## # Groups: k [3]
## k variable mean sd
## <dbl> <fct> <dbl> <dbl>
## 1 3 normal -0.00558 0.972
## 2 3 vitaclayton 6.94 3.79
## 3 3 vitajoe 10.0 4.45
## 4 7 normal -0.0453 0.791
## 5 7 vitaclayton 2.82 1.44
## 6 7 vitajoe 4.52 2.02
## 7 10 normal 0.0150 0.589
## 8 10 vitaclayton 1.91 0.984
## 9 10 vitajoe 2.98 1.46
## # A tibble: 3 x 3
## variable mean sd
## <fct> <dbl> <dbl>
## 1 normal -0.0120 0.798
## 2 vitaclayton 3.89 3.25
## 3 vitajoe 5.83 4.21