회귀 (연속형 변수) 측정하는법 MSE, RMSE, MAE, MPE, MAPE in python

<엔지니어는 구현을 못하면 모르는것이다>

회귀 (연속형 변수) 측정

MSE, RMSE, MAE, MPE, MAPE 사용하여 계산

모두 Mean Error 가 들어간다. => 즉 평균을 계산할수 있는 연속형 변수 측정에 사용한다.

import pandas as pd
import numpy as np
from sklearn import datasets
from sklearn.model_selection import train_test_split
datas = datasets.load_diabetes()
datas.target
df = pd.DataFrame(datas.data, columns = datas.feature_names)
y = pd.DataFrame(datas.target, columns = ['y'])
df = pd.concat([df, y], axis=1)
train, test = train_test_split(df, test_size=0.3, random_state=123)

from sklearn.linear_model import LinearRegression
lr = LinearRegression()
lr.fit(train.drop(columns='y'), train.y)
y_pred = lr.predict(test.drop(columns='y'))
y_true = test.y

# 평균 제곱 오차 mean squared error
from sklearn.metrics import mean_squared_error
mse = mean_squared_error(y_true, y_pred)

def MSE(y, pred):
  return np.mean(np.square(y-pred))

print(MSE(y_true, y_pred), mse)

# 평균 제곱근 오차 root mean squared error
from sklearn.metrics import mean_squared_error
rmse = np.sqrt(mean_squared_error(y_true, y_pred))

def RMSE(y, pred):
  return np.sqrt(np.mean(np.square(y-pred)))

print(RMSE(y_true, y_pred), rmse)

# 평균 절대 오차 mean absolute error
from sklearn.metrics import mean_absolute_error
mae = mean_absolute_error(y_true, y_pred)

def MAE(y, pred):
  return np.mean(np.abs(y-pred))

print(MAE(y_true, y_pred), mae)

# 평균 백분율 오차 mean absolute percentage error
def MPE(y, pred):
  return np.mean((y-pred)/y)*100

print(MPE(y_true, y_pred))

# 평균 절대 백분율 오차 mean percentage error
def MAPE(y, pred):
  return np.mean(np.abs((y-pred)/y)*100)

print(MAPE(y_true, y_pred))