gqqnbig / cs145project

0 stars 2 forks source link

Closed gqqnbig closed 4 years ago

gqqnbig commented 4 years ago

validation set准确率0.955。

gqqnbig commented 4 years ago

没问题了

gqqnbig commented 4 years ago

基准准确率0.6652100840336135

用StratifiedKFold

如果混合training和validation，准确率是0.6898395721925134。

只对training应用交叉验证，准确率反而低0.6605042016806723。

只对training应用交叉验证，0.6712605042016807

gqqnbig commented 4 years ago

全集基准(f7a4c6b13334ff95461e59a1a5e9e505a7c26dd6) 0.6601360859924246

KFold 0.6591

gqqnbig commented 4 years ago

Git master DecisionTree with accuracy 0.66013608 kaggle score 0.64245

Git kfold-simple K=5 DecisionTree submitted with accuracy 0.6591 kaggle score 0.64188

K=10 (81fe47fb02bf) 0.659524219125853 kaggle score 0.64170

gqqnbig commented 4 years ago

Not an improvement