Machine Learning

Regression

December 24, 2020

Why do we do a train test split?

We do a test train split to ensure that the model prediction we are getting is due to actual learning and not overfitting. We train the model on training set and then make the predictions, using the equation learned from training set, on test set. If the accuracy of prediction is similar for both training and test sets then the model is working properly and has not over fitted.

by : Monis Khan

Quick Summary:

We do a test train split to ensure that the model prediction we are getting is due to actual learning and not overfitting. We train the model on training set and then make the predictions, using the equation learned from training set, on test set. If the accuracy of prediction is similar for both training […]