Answer-first summary for fast verification
Answer: To ensure the model's ability to generalize beyond the training data
Evaluating a predictive model solely on training data fails to assess its performance on new, unseen data. Selecting a model based on training data accuracy often results in poorer performance on test data due to overfitting, where the model becomes too tailored to the training dataset's specifics. This underscores the importance of a separate test set for validating the model's generalizability. Reference: [Machine Learning Mastery](https://machinelearningmastery.com/a-simple-intuition-for-overfitting/)
Author: LeetQuiz Editorial Team
Ultimate access to all questions.
No comments yet.
What is the primary reason for dividing a machine learning dataset into training and test sets?
A
To enable the testing of different feature sets
B
To ensure the model's ability to generalize beyond the training data
C
For the purpose of incorporating unit tests within your code
D
To allocate separate datasets for wide and deep model training