Random forests for exoplanets

One of today’s arXivals examines the use of random forest regression to predict exoplanetary radii given a training set with approx 10 observed properties of exoplanets and their host stars.  I don’t quite understand the motivation for a random forest approach in this instance because (a) the learned relationship seems not to be particularly complex and (b) the observables all come with a range of non-inconsequential heteroscedastic uncertainties.  A regularised, low-order polynomial regression with errors-in-variables framework could handle this problem easily and would offer advantages in terms of interpretability and shrinkage on the uncertain observables.

This entry was posted in Uncategorized. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s