roc_plot.ipynb

Kernel: Unknown Kernel

# ROC curves for two hypothetical classification systems
# A is better than B. Plots true positive rate, (tpr) vs false positive
# rate, (fpr).


import numpy as np
import matplotlib.pyplot as plt

try:
    import probml_utils as pml
except ModuleNotFoundError:
    %pip install -qq git+https://github.com/probml/probml-utils.git
    import probml_utils as pml


fA = np.vectorize(lambda x: x ** (1.0 / 3))
fB = np.vectorize(lambda x: x ** (2.0 / 3))
x = np.arange(0, 1, 0.01)

plt.plot(x, fA(x), "r-")
plt.plot(x, fB(x), "b-")
plt.fill_between(x, fB(x), 0, color="gray", alpha=0.2)
plt.plot(x, 1 - x, "k-")

inter_a = 0.3177  # found using scipy.optimize.fsolve(x**(1.0/3)+x-1, 0)
inter_b = 0.4302  # found using scipy.optimize.fsolve(x**(2.0/3)+x-1, 0)

plt.plot(inter_a, fA(inter_a), "ro")
plt.plot(inter_b, fB(inter_b), "bo")

plt.text(inter_a, fA(inter_a) + 0.1, "A", color="red", size="x-large")
plt.text(inter_b, fB(inter_b) + 0.1, "B", color="blue", size="x-large")

plt.xlabel("FPR", fontsize=14)
plt.ylabel("TPR", fontsize=14)
pml.savefig("ROChand.pdf")
plt.show()