使用Hotelling的T2进行置信区间的R PCA图

Question

For my test data, I calculated the first two principal components using the prcomp function. I then plotted the first 2 components using the plot function.

df <- data.frame(x1 = runif(10), x2 = runif(10), x3 = runif(10), x4 = runif(10))
rownames(df) <- letters[1:10]

res <- prcomp(df, center = TRUE, scale = TRUE, na.action = na.omit)

plot(res$x, xlim = c(-5, 5), ylim = c(-5, 5))

I now would like to place an ellipse around the data that represens a 95% confidence interval using Hotelling's T2 statistic.

The SensoMineR package in particular the panellipse function seems to be what I am looking for. However, my R knowledge is still quite limited and I was therefore wondering if you could point me in the right direction on to construct such a confidence interval.

Many thanks! Jacky

Answer 1

SensoMineR is good.

There is another package called Package 'Hotelling' which is easy to place an ellipse around the data.

Package 'Hotelling'