如何使用三个自变量拟合线性回归模型并使用sklearn计算均方误差？

Question

I'm trying to fit a linear regression model using three independent variables and calculate the mean squared error using sklearn, but I seem not be able to get it right.

My data is the Boston Housing, and three independent variables are as follow: 1. CRIM (per capita crime rate by town) 2. RM (average number of rooms per dwelling) 3. PTRATIO (pupil-teacher ratio by town)

Fit model:

import numpy as np
import pandas as pd
import matplotlib as mpl
import matplotlib.pyplot as plt
import sklearn
lm = LinearRegression()
lm.fit(X[['CRIM']['RM'], ['PTRATIO']], boston_df.PRICE)

Calculate the mean square error

from sklearn.metrics import mean_squared_error
y_true = ['CRIM', 'RM', 'PTRATIO']
y_pred = ['PRICE']
mean_squared_error(y_true, y_pred)

Any advice or hints are much appreciated!

Answer 1

Try X[['CRIM', 'RM', 'PRTATIO']] instead of X[['CRIM']['RM'], ['PTRATIO']] for fitting the model

For prediction you need to compare these two vectors:

y_true = boston_df.PRICE
y_pred = lm.predict(X[['CRIM', 'RM', 'PRTATIO']])
mean_squared_error(y_true, y_pred)

Basically your y_pred should be the predicted values from your model which is lm in this case.

Answer 2

sklearn has great documentation. Here is a super thorough example complete with example data set: http://scikit-learn.org/stable/auto_examples/linear_model/plot_ols.html

The biggest problem that you are having is your data set. Like your code here:

y_true = ['CRIM', 'RM', 'PTRATIO']
y_pred = ['PRICE']

That isn't even real data, it is just 2 lists of string labels, so of course this won't work:

mean_squared_error(y_true, y_pred)

From the example I posted, you could try this "hello world" type code (using existing data sets) just to make sure you are getting the code working, then all you need to do is replace the dataset with your own data. As you can see most of the code is dedicated to preparing the data so it will load correctly into the linear regression function:

import matplotlib.pyplot as plt
import numpy as np
from sklearn import datasets, linear_model

# Load the diabetes dataset
diabetes = datasets.load_diabetes()

# Use only one feature
diabetes_X = diabetes.data[:, np.newaxis, 2]

# Split the data into training/testing sets
diabetes_X_train = diabetes_X[:-20]
diabetes_X_test = diabetes_X[-20:]

# Split the targets into training/testing sets
diabetes_y_train = diabetes.target[:-20]
diabetes_y_test = diabetes.target[-20:]

# Create linear regression object
regr = linear_model.LinearRegression()

# Train the model using the training sets
regr.fit(diabetes_X_train, diabetes_y_train)

print("Mean squared error: %.2f" % np.mean((regr.predict(diabetes_X_test) - diabetes_y_test) ** 2))