search

ageron / handson-ml2

A series of Jupyter notebooks that walk you through the fundamentals of Machine Learning and Deep Learning in Python using Scikit-Learn, Keras and TensorFlow 2.
Apache License 2.0
28k stars 12.8k forks source link

[Chapter 2] How to calculate confidence intervals for MAE? #155

Open vasili111 opened 4 years ago

vasili111 commented 4 years ago

First of all thank you for creating wonderful book! I find it very well written and easy to follow.

My question is: At page 80 there is a method for estimation confidence intervals for RMSE but not for MAE. I find MAE helpful too, because it is more easy to explain it to layman person than RSME. How can I estimate confidence intervals for MAE?

Thank you.

Rosseel commented 4 years ago

My best guess would be confidence=0.95 errors=final_predictions-y_test stats.t.interval(confidence,len(errors)-1,loc=errors.mean(),scale=stats.sem(errors))

So a simplified version of the RMSE case.

vasili111 commented 3 years ago

@Rosseel

Thank you. Why errors are not squared here as with RMSE?