ROC-AUC score

roc_auc.py

# demonstrates plotting ROC curve and getting AUC score
import numpy as np

def roc_curve(y_true, y_pred):
    fpr = []
    tpr = []
    thresholds = np.arange(0.0, 1.01, .01)

    P = sum(y_true)
    N = len(y_true) - P

    for thresh in thresholds:
        FP=0
        TP=0
        for y_t, y_p in zip(y_true, y_pred):
            if y_p > thresh:
                if y_t == 1:
                    TP = TP + 1
                else:
                    FP = FP + 1
        fpr.append(FP/float(N))
        tpr.append(TP/float(P))
    # plt.plot(fpr, tpr)
    return tpr, fpr


def integrate(x_s, y_s):
    # integartion by trapezoidal rule
    x_diffs = [abs(x_s[i]-x_s[i-1]) for i in range(1,len(x_s))]
    y_sum = [y_s[i]+y_s[i-1] for i in range(1,len(y_s))]
    return .5*sum([a*b for a, b in zip(y_sum, x_diffs)])


# example
from sklearn.datasets import make_classification
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression

X, y = make_classification(n_samples=1000, n_classes=2, weights=[1,1], random_state=1)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.5, random_state=2)
model = LogisticRegression()
model.fit(X_train, y_train)
y_pred = model.predict_proba(X_test)[:,1]

tpr, fpr = roc_curve(y_test, y_pred)
auc = integrate(x_s=fpr, y_s=tpr)
plt.plot(fpr, tpr)
print("AUC score = {}".format(auc))

# verify
import sklearn.metrics as metrics, auc

fpr, tpr, threshold = metrics.roc_curve(y_test, y_pred)
roc_auc = metrics.auc(fpr, tpr)

fig = plt.figure(figsize=(10,6))
ax = fig.add_subplot(111)
ax.set_title('Receiver Operating Characteristic')
ax.plot(fpr, tpr, 'b', label = 'AUC = %0.2f' % roc_auc)
ax.legend(loc = 'lower right')
ax.plot([0, 1], [0, 1],'r--')
plt.xlim([0, 1])
plt.ylim([0, 1])
ax.set_ylabel('True Positive Rate')
ax.set_xlabel('False Positive Rate')