How to Calculate Training Error in Python

Before we dive into calculating the training error using Python, it is imperative to know that the training error is a measure of how well your model performs on the training data.

Improved training performance often leads to increased performance on unseen “test” data. Here, we’ll use machine-learning libraries sklearn and numpy, ensuring to highlight key elements as we progress.

Step 1: Install Required Libraries

This tutorial will utilize several Python libraries. If you don’t have them installed, you can do so using the pip command:
For the code,

1
pip install sklearn numpy pandas

Step 2: Import the Libraries

We’ll start by importing the libraries necessary for this task.

1
2
3
4
import numpy as np
from sklearn.model_selection import train_test_split
from sklearn import datasets, linear_model
from sklearn.metrics import mean_squared_error

Step 3:

Load and Split Dataset
We’re going to use the diabetes dataset available in sklearn for this tutorial. We’ll split it into training and testing data.

1
2
3
4
diabetes = datasets.load_diabetes()
X = diabetes.data
y = diabetes.target
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)

Step 4:

Fit the Model
We will use the linear regression model from sklearn, and fit it on the training data.

1
2
model = linear_model.LinearRegression()
model.fit(X_train, y_train)

Step 5:

Calculate the Training Error
We will predict the output on the training data and then calculate the MEAN SQUARED ERROR(MSE) which is our training error.

1
2
y_train_predict = model.predict(X_train)
training_error = mean_squared_error(y_train, y_train_predict)

Full Code:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
import numpy as np
from sklearn.model_selection import train_test_split
from sklearn import datasets, linear_model
from sklearn.metrics import mean_squared_error
 
diabetes = datasets.load_diabetes()
X = diabetes.data
y = diabetes.target
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)
 
model = linear_model.LinearRegression()
model.fit(X_train, y_train)
 
y_train_predict = model.predict(X_train)
training_error = mean_squared_error(y_train, y_train_predict)
print('Training Error(MSE): ', training_error)

Output:

Training Error(MSE): 2844.73157308

Conclusion

With Python’s sklearn, calculating the training error becomes a straightforward task. Understanding these metrics is crucial as they provide insight into how well the model is learning from the training data. Keep in mind that although a low training error is a good indicator, testing the model on unseen data is still vital to ensure that the model is not overfitting.