Correlation is enough
Are models meaningless if you have enough data? Some people argue that is the case.
posted by lotek at 2:48 PM
This is a blog about me, Seinfeld, okonomiyaki, japanese toilet seats, and other things of interest
Are models meaningless if you have enough data? Some people argue that is the case.
posted by lotek at 2:48 PM
2 Comments:
No. Nononononononononono (RAGE)
1:00 AM
I agree with ar!s. Stupid assertation that is.
Science attempts to explain *why* things happen. That's why we build models. It's not enough to just find correlation between events. The famous example - Ice cream sales are proportionally related to murders. In terms of raw data, that's a correlation. But it tells us nothing about the underlying cause of both events.
9:14 AM
Post a Comment
<< Home