Bias

• The ability of our model to predict the values/labels correctly
• Assumptions made by a model to make a function easier to learn.

• It is the error rate of the training data

• high_bias

  • model is simple, not able to learn from data
  • Error rate has a high value
  • Underfitting model will have #high_bias

• low_bias

  • model is complex enough, ability to learn the pattern and predictions are close to the target.
  • Error rate has a low value