Implementation of linear classification algorithms on banknote authentication dataset. The models implemented include Fisher's linear discriminant, Probabilistic generative model and logistic regression.
Goal : Implement different linear models for binary classification.
Dataset: uci repository’s ’banknote authentication Data Set’ https://archive.ics.uci.edu/ml/datasets/banknote+authentication
I. Fisher’s linear discriminant
confusion matrix :
threshold: 0.966963 tp : 184 fp : 3 tn : 224 fn : 1
correct predictions: 408
incorrect predictions: 4
precision: 0.983957
recall: 0.994595
accuracy: 99.0291%
Confusion Matrix:
n = 412 Predicted: NO Predicted: YES
Actual: NO TN =224 FP =3
Actual: YES FN = 1 TP = 184
II. Probabilistic generative model
tp : 185 fp : 11 tn : 216 fn : 0
correct predictions: 401
incorrect predictions: 11
precision: 0.943878
recall: 1
accuracy: 97.3301%
Confusion Matrix:
n = 412 Predicted: NO Predicted: YES
Actual: NO TN =216 FP =11
Actual: YES FN = 0 TP = 185
III. Logistic Regression Model
true_positive: 185 true_negative: 224 false_positive: 3 false_negative: 0
correctly predicted: 409
incorrectly predicted: 3
precision : 0.984043
recall : 1
accuracy: 99.2718%
Confusion Matrix:
n = 412 Predicted: NO Predicted: YES
Actual: NO TN =224 FP =3
Actual: YES FN = 0 TP = 185