How to compute false positive rate of an imbalanced dataset for Stratified K fold cross validation?

Question

The below lines are the sample code where I am able to compute accuracy, precision, recall, and f1 score. How can I also compute a false positive rate (FPR) for Stratified K fold cross-validation?

from sklearn.metrics import make_scorer, accuracy_score, precision_score, recall_score, 
f1_score
from sklearn.model_selection import StratifiedKFold
from sklearn.model_selection import cross_validate

scoring = {'accuracy' : make_scorer(accuracy_score), 
       'precision' : make_scorer(precision_score),
       'recall' : make_scorer(recall_score), 
       'f1_score' : make_scorer(f1_score)}
skfold = StratifiedKFold(n_splits=10)
dt_clf = DecisionTreeClassifier()

results = cross_validate(estimator=dt_clf,
                      X=data_train_X,
                      y=target_train_Y,
                      cv=skfold,
                      scoring=scoring)
print("Results", results)

Answer 1

You could define a custom scorer as follows:

from sklearn.metrics import make_scorer, accuracy_score, precision_score, recall_score, f1_score
from sklearn.model_selection import StratifiedKFold, cross_validate
from sklearn.tree import DecisionTreeClassifier
from sklearn.datasets import make_classification

def false_positive_rate(y_true, y_pred):

    # false positive
    fp = ((y_pred == 1) & (y_true == 0)).sum()

    # true negative
    tn = ((y_pred == 0) & (y_true == 0)).sum()

    # false positive rate
    return fp / (fp + tn)

scoring = {
    'accuracy': make_scorer(accuracy_score),
    'precision': make_scorer(precision_score),
    'recall': make_scorer(recall_score),
    'f1_score': make_scorer(f1_score),
    'false_positive_rate': make_scorer(false_positive_rate),
}

skf = StratifiedKFold(n_splits=3)

clf = DecisionTreeClassifier(random_state=42)

X, y = make_classification(random_state=42)

results = cross_validate(estimator=clf, X=X, y=y, cv=skf, scoring=scoring)

print(results['test_false_positive_rate'])
# [0.11764706 0.11764706 0.0625]

Answer 2

hi i wrote this code with Logestic regression u can change it with whatever u want

#Importing required libraries
from sklearn.model_selection import KFold
from sklearn.datasets import load_breast_cancer
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import accuracy_score
from sklearn.metrics import confusion_matrix
data = load_breast_cancer(as_frame = True)
df = result.frame
X = result.iloc[:,:-1]
y = result.iloc[:,-1]


#shffling
X = X.sample(frac = 1)
y= y.sample(frac = 1)
#Implementing cross validation
kf = KFold(n_splits=10)#, random_state=None
model = LogisticRegression(max_iter=1000000)#(solver= 'liblinear')

acc_score = list()
res_tpr = list()
res_fpr = list()
for train_index , test_index in kf.split(X):
    #X_train , X_test = X.iloc[train_index,:],X.iloc[test_index,:]
    #y_train , y_test = y[train_index] , y[test_index]
    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3)
    model.fit(X_train,y_train)
    pred_values = model.predict(X_test)
    tn, fp, fn, tp = confusion_matrix(y_test, pred_values, labels=[0, 1]).ravel()
    print(f'True Positives: {tp}')
    print(f'False Positives: {fp}')
    print(f'True Negatives: {tn}')
    print(f'False Negatives: {fn}')
    tpr=(np.divide(tp,(tp+fn)))
    fpr=(np.divide(fp,(fp+tn)))
    if tp==0:
      tpr=0
    if fp==0:
      fpr=0

    print('tpr=%.4f fpr=%.3f' % ( mean(tpr), mean(fpr)))
    res_tpr.append(mean(tpr))
    res_fpr.append(mean(fpr))
    print('---------------------')

    acc = accuracy_score(pred_values , y_test)
    acc_score.append(acc)
 
avg_acc_score = np.sum(acc_score)/10
total_tpr=np.sum(res_tpr)/10
total_fpr=np.sum(res_fpr)/10
print('\n\n',' total_tpr=%.4f total_fpr=%.3f' % (total_tpr,total_fpr))
#print('\n\n','accuracy of each fold - {}'.format(acc_score))
print('\n\n','Avg accuracy : {}'.format(avg_acc_score))