- Bias in machine learning refers to the error introduced by approximating a real-world problem, which may be extremely complex, by a much simpler model. If our model is too simple, it may have high bias, meaning it consistently predicts inaccurately. It’s like always missing the answer to a hard math problem because you’re only using simple addition and subtraction when you should be using multiplication and division.
- Variance in machine learning refers to the amount that our model’s predictions would change if we used a different training dataset. If our model is too complex, it might have high variance, meaning it doesn’t generalize well to new, unseen data. It’s like memorizing the answers to a specific set of math problems, then doing poorly on a test because the problems aren’t exactly the same.
In machine learning, the goal is to balance bias and variance to ensure the model is both generally accurate (low bias) and reliable when presented with new data (low variance). This is called the bias-variance tradeoff.
« Back to Glossary Index