How to set the minimum number of instances in one node

Hi,

for data protection reasons I want to make sure that splitting stops when a minimum number of observation in one node is reached. In other tree boosting implementations there is a parameter to specify this. In case of Xgboost the manual says that min_child_weight specifys the minimum number of observations in regression mode.

Is there a possibility to specify the minimum number of observations in classification mode?

Thank you for your help.

No repository file or test is named. Start with the XGBoost manual's min_child_weight entry and trace how the parameter is applied in classification, comparing the documented behavior with existing tests. Done means establishing whether a minimum observation count can be configured for classification and, if not, defining the required feature and test coverage.

Written by the indexing model from the issue text.