Term
overfitting
The mistake of fitting a model too closely to past data, so it takes random noise for a pattern. It looks good on the past and gets the future wrong.
The Signal and the Noise: Why So Many Predictions Fail, but Some Don't, Nate Silver
Explained in
- The Signal and the Noise: Why So Many Predictions Fail, but Some Don't, Nate Silver
- Algorithms to Live By: The Computer Science of Human Decisions, Brian Christian, Tom Griffiths
The error of a model fitted too closely to past data, which makes it worse at predicting the future.