I made a Classification model, I would like to understand the leaf values. So when I print the trees:
0:[repy__v_user_cashloan_active_repy_trade_cnt_30d<0.999998987] yes=1,no=2,missing=2,gain=18276.9141,cover=374122.75
1:[repy__p_user_cashloan_last_repyt_succ_date_t_cur_datediff<114] yes=3,no=4,missing=4,gain=8097.74072,cover=363191.75
3:[visit_behavior_r_m_on_6m_pv_days_rencentrepay<0.0454535484] yes=7,no=8,missing=7,gain=1420.46899,cover=86702
7:[loan__r_user_cashloan_unsettled_loan_terms_cnt_90d<0.882398963] yes=15,no=16,missing=15,gain=565.593018,cover=59721
15:leaf=-0.126535222,cover=20534.25
16:leaf=-0.143589199,cover=39186.75
8:[loan__r_user_cashloan_principal_balance_90d<0.838899016] yes=17,no=18,missing=17,gain=372.071869,cover=26981
17:leaf=-0.10015884,cover=8358.25
18:leaf=-0.121303588,cover=18622.75
…
and How can I understand the cover?
This is defined on the document ：coverage is defined as the number of samples affected by the split
but why the cover is not integers?
I’m sorry if I asked something that someone else has already done, I did not find nothing.
Best wishes,
gg