metrics

import numpy as np
from math import sqrt


def accuracy_score(y_true, y_predict):
    """计算y_true和y_predict之间的准确率"""
    assert len(y_true) == len(y_predict), \
        "the size of y_true must be equal to the size of y_predict"

    return np.sum(y_true == y_predict) / len(y_true)


def mean_squared_error(y_true, y_predict):
    """计算y_true和y_predict之间的MSE"""
    assert len(y_true) == len(y_predict), \
        "the size of y_true must be equal to the size of y_predict"

    return np.sum((y_true - y_predict)**2) / len(y_true)


def root_mean_squared_error(y_true, y_predict):
    """计算y_true和y_predict之间的RMSE"""

    return sqrt(mean_squared_error(y_true, y_predict))


def mean_absolute_error(y_true, y_predict):
    """计算y_true和y_predict之间的RMSE"""
    assert len(y_true) == len(y_predict), \
        "the size of y_true must be equal to the size of y_predict"

    return np.sum(np.absolute(y_true - y_predict)) / len(y_true)


def r2_score(y_true, y_predict):
    """计算y_true和y_predict之间的R Square"""

    return 1 - mean_squared_error(y_true, y_predict)/np.var(y_true)