Team Choice: Classification
Joshua Staples
Model: GBT
Accuracy: 75%
Precision: 83% (0's) 55% (1's)
F1-Score: 83% (0's) 55% (1's)
Does really good identifying 0's, barely better than a guess for 1's.
Excluded the >100000 Income column.
ROC Curve:
Kody Smith
Model: GBT
Area Under ROC Curve: 79.79%
Accuracy: 77.2%
Recall: 0's: 93.00% 1's: 40.40%
Precision: 0's: 78.56% 1's: 71.08%
F1-Score: 0's: 85.17% 1's: 52.52%
Excluded the >100000 Income column.
ROC Curve:
Matthew Goodsell
Model: Random Forest Classifier
Area Under ROC Curve: 77.9%
Accuracy: 76.7%
Precision on 0s: 80.12% 1s: 61.75%
F1 Score 0s: 84.85% 1s: 49.65%
Recall: 0s: 90.17% 1s: 41.51%
Excluded the >100000 Income column.
Random Forest Classifier with 77.9 Area Under ROC Curve
Model's parameters:
- Number of trees set to 15
- Max Depth of the trees set to 15
Top ten most important Features
