1. ## Orthogonal Min/Max Operator?

I have a cluster of 2 overlapping data populations, both of which can be described by a simple linear equation. I am looking for help to find a "simple" operator that maximizes the difference between the 2 populations. Realize that the overlapping data is still going to be in the overlapping cluster but I want to maximize the extremes.

2. ## Re: Orthogonal Min/Max Operator?

I have a few questions, if you don't mind.

1. What do you mean by "overlapping" data populations? Do you mean that, say, I have one time series that runs from t= 0 to 10, and another that runs from t = 5 to 15?

2. Can you exhibit the "simple linear equation" or do you not know it in advance?

3. What do you mean by "difference"? If I'm looking at a plot of your two data populations, what sort of thing am I doing by taking a difference?

3. ## Re: Orthogonal Min/Max Operator?

I know the linear regression for the green data population and the red data population, a simple y=ax+b. By overlapping, I mean the area where the polygons/data have similar/the same values. By difference I mean, I mean that in the transformed space, the differences in the new x and y has increased/maximized.

Make sense??