Open PhilippPro opened 6 years ago
I've read the catboost paper and really like their approach.
AFAIK they have a (marginally) better performance than xgboost and others but have a worse runtime (which makes sense due to their more complex algorithm).
I might switch over to a different gbm implementation in the future if it really has advantages.
Did you consider in making (also) an autocatboost algorithm?
On their webpage they claim to be better regarding performance and runtime: https://catboost.yandex/
Would be interesting to compare this properly. Unfortunately it is not on CRAN yet.