Overview

1. Getting more training examples
2. Trying smaller sets of features
3. Trying additional features
4. Trying polynomial features
5. Increasing or decreasing λ

Split Into Training and Test Set

• A hypothesis may have a low error for the training examples but still be inaccurate (because of overfitting).
• Thus, to evaluate a hypothesis, given a dataset of training examples, we can split up the data into two sets:
  • a training set and
  • a test set.
• Typically, the training set consists of 70 % of your data and the test set is the remaining 30 %.

Evaluating the Hypothesis by looking at the Test Error value

1. Learn Θ, then minimise Jtrain (Θ) using the training set.
2. Compute the test set error Jtest(Θ)

How to compute the Test Error

The test error is a ratio value which gives an indication of the percentage of

For Linear Regression:

For Classification ~ Misclassification error (aka 0/1 misclassification error):