I just found out that I made a mistake in my previous post (it is fixed now). There, I argued that in order to calculate the distance between a point and a set (which has unequal variances along different dimensions), we can use the following formula:

Advertisements