How to use cv::ml::Boost for a regression problem?

The OpenCv documentation says that:

  • class cv::ml::Boost - Boosted tree classifier derived from DTrees
  • class cv::ml::RTrees - The class implements the random forest predictor

With the RTrees I received correct prediction for the regression data and I was unable to get any result with Boost for the same data.
I have tried to use all types (DISCRETE, REAL, LOGIT, GENTLE) and predict with the flag RAW_OUTPUT.

The class description clearly states that the cv::ml::Boost class is a classifier, however, there are many references in the Boost documentation about its use for regression.

For example: “Gentle AdaBoost. It puts less weight on outlier data points and for that reason is often good with regression data.”

Do you have any successful cases of using Boosted for regression? If so, what settings should be used?