search

BrandonEugeneWalker / BrandonKevinDanielIntelligentSystemsProject

0 stars 0 forks source link

Run model with best single results vs all hyper-parameters. #9

Open BrandonEugeneWalker opened 4 years ago

BrandonEugeneWalker commented 4 years ago

Best Results: positive_ratings owners median_playtime

VS: positive_ratings negative_ratings owners average_playtime median_playtime

BrandonEugeneWalker commented 4 years ago

All parameters

---Linear Regression--- Mean MSE over 5 folds: 1.1354533408474803e-18 Total Time: 0:03:55.528134 ---Tree Regression--- Mean MSE over 5 folds: 34686106860.15331 Total Time: 0:03:56.698660 ---Random Forest--- Mean MSE over 5 folds: 67601188242.77806 Total Time: 0:06:08.836925

BrandonEugeneWalker commented 4 years ago

positive_ratings, owners, median_playtime

---Linear Regression--- Mean MSE over 5 folds: 3.2114453694881307e-19 Total Time: 0:03:51.922673 ---Tree Regression--- Mean MSE over 5 folds: 125908934964.69576 Total Time: 0:03:52.944860 ---Random Forest--- Mean MSE over 5 folds: 67326510977.71724 Total Time: 0:06:05.786569

BrandonEugeneWalker commented 4 years ago

Results: It seems despite the excluded columns having less correlation than the others, leaving them out results in a worse MSE for all models.

KevinFlynn12 commented 4 years ago

Results for Single parameter "positiveratings" ---Linear Regression--- Regression - Mean MSE over 5 folds: 62.51536028456036 Total Time: 0:00:00.046846 ---Tree Regression--- Decision Tree - Mean MSE over 5 folds: 69.14714376826723 Total Time: 0:00:00.105716 ---Random Forest--- Random Forest - Mean MSE over 5 folds: 63.80594498626552 Total Time: 0:00:08.480339 ---Bagging--- Bagging - Mean MSE over 5 folds: 66.67344973257234 Total Time: 0:00:08.512252 ---Boosting--- Boosting - Mean MSE over 5 folds: 131.40465915333078 Total Time: 0:00:00.260636 ---Voting--- Voting - Mean MSE over 5 folds: 60.57676752432933 Total Time: 0:00:01.104050