The code of "Brier Score", "ePCP", "the separation plot". Please import from this file to your project. The "actual_outcome" is a list of real labels (0 or 1), and the "fitted_value" is a list of probabilities that the data's label is 1. This code requires "numpy" and "pandas".

assess.py

from matplotlib import pyplot as plt
from matplotlib import patches
import pandas as pd
import numpy as np

def BrierScores(actual_outcome, fitted_value):
    X = np.array(actual_outcome)
    p = np.array(fitted_value)
    B = sum((p-X)**2)/len(X)
    return B

def ePCP(actual_outcome, fitted_value):
    ps = 0
    for X,p in zip(actual_outcome,fitted_value):
        if X == 0:
            ps += 1 - p
        else:
            ps += p
    ePCP = ps / len(actual_outcome)
    return ePCP

def SeparationPlot(fitted_value, actual_outcome):
    data = np.vstack([fitted_value, actual_outcome])
    data = data[:, np.argsort(data[0])]
    l = len(actual_outcome)
    ENoE = np.sum(data[0])
    plt.figure(figsize=(10, 2))
    ax = plt.axes()
    fx = [p for p in data[0]]
    x = range(len(fx))
    lw = 1000 / len(x)
    plt.xlim([-(lw/2),l-lw/2])
    plt.ylim([0,1])
    for i, y in enumerate(data[1]):
        if y == 0:
            ax.add_patch(patches.Rectangle(xy=(float(i)-lw/2, 0), width=lw, height=1.0, fc="w", ec='w', fill=True))
        else:
            ax.add_patch(patches.Rectangle(xy=(float(i)-lw/2, 0), width=lw, height=1.0, fc="g", ec='g', fill=True))
    plt.plot(x,fx,color="#000000")
    plt.plot([float(len(fitted_value)-ENoE)-lw/2, float(len(fitted_value)-ENoE)-lw/2], [0,1], color="r")
    plt.title("Expected Number of Events = {:.2f}".format(ENoE))
    plt.xlabel("rank")
    plt.ylabel("probability")
    plt.tight_layout()
    plt.show()
    return data, ENoE