For a paper I’m currently working on, I’m dealing with several types of Gaussian distributions that signal to onset and decay of an event. Events can have a slow onset and slow decay, resulting in a wide distribution, or they can have a fast onset and slow decay or vice versa. I’m exploring different ways to compare these distributions.

- comparing the entropy of the distribution, which gives a general statistic
- comparing the (linear) slopes of either side of distribution
- calculating the skewness and kurtosis of the distributions.

I was wondering if others had pointers and possible feedback on the proposed methods.