Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Bias correction for TransformedTargetRegressor

See original GitHub issue

Description

If one is interested in predicting the (conditional on X) expected value of a target y, aka mean, then TransformedTargetRegressor should use a bias corrected inverse transform.

It would be nice to have an option for bias correction in TransformedTargetRegressor. At least, I would mention this in the user guide.

References

https://robjhyndman.com/hyndsight/backtransforming/

Issue Analytics

State:
Created 4 years ago
Comments:9 (8 by maintainers)

Top GitHub Comments

2reactions

robjhyndmancommented, Jul 5, 2020

It can also be done numerically. This is how bias correction is handled in fabletools (for R) so that any transformation can be used: https://github.com/tidyverts/fabletools/blob/master/R/transform.R#L102

1reaction

lorentzenchrcommented, Aug 19, 2020

@robjhyndman Thank you for pointing to the numerical solution. If I understand correctly, this is an additive correction, i.e. the biased corrected prediction is predict_biased + bias_correction.

Top Results From Across the Web

sklearn.compose.TransformedTargetRegressor

Meta-estimator to regress on a transformed target. Useful for applying a non-linear transformation to the target y in regression problems. This transformation ...

Correcting log-transformation bias in a linear model

I am trying to apply Miller's (1984) proposed bias correction to the predicted values (see p. 125). According to the study,.

Regression — Scikit-learn course - GitHub Pages

On this plot, correct predictions would lie on the diagonal line. ... allows us to detect if the model makes errors in a...

10 Sklearn Gems Buried In the Docs Waiting To Be Found

That's where you elegantly fix this dilemma with custom transformers. ... TransformedTargetRegressor is a class that accepts both a regressor pipeline on ...

regression | Andrew Wheeler

Again this does not cause bias in the equation (and so will not ... be the “right” choice (likely due to the degree...