to be 'approx' for faster speed. to use old behavior(exact greedy algorithm on single machine), set tree_method我尝试过clf = xgb(tree_method='exact'),它返回了错误!
class Inner( tree_method}我正在传递这样一个json的输入
输入json val prparamas= CampaignSparkParameters( {“类别”:“x”,“数据”:{“增强器”:“gbtree”,"tree_methodmax_depth":"5","max_bin":&