Bias-corrected random forests in regression