IssueTracker

Catboost is a quite new boosting machine learning algorithm that  has the reputation of being a superior data mining tool to Random Forest in terms of accuracy  in recent machine learning competitions.  It could be something to consider incorporating catboost, which is open source, directly into Google Earth Engine in the same fashion as a machine learning algorithm as SVM, CART, Random Forest algorithm (or the GEE version, smileRandomForest).  One would expect to get better test confusion matrix results than Random Forest in general wth Catboost.  The algorithm is thought to be relatively easy to tune as is the case with Random Forest.

Here is the source of this code:
https://github.com/catboost/catboost

Just a suggestion.

Respectfully,

James Cormier-Chisholm