Semi-OT: outliers and artifacts

[josef...@gmail.com]

http://xkcd.com/1781/
"I didn't even realize you could HAVE a data set made up entirely of outliers."

Josef

- Wasn't there robust regression or something?

- Yes, but some of those thingies are stuck in PRs.

- Do you have anything with a breakdown point of 100%?

- ??? Yes, but then the answer is always 42.

def fit(data, breakdown_point=1):

    if breakdown_point == 1:

        return 42

    else:

        return RLM(data).fit()