误差、偏差与方差 (Error, Bias and Variance)
High Bias
$J_{train}(\theta)$ will be very high while $J_{train}(\theta) \approx J_{cv}(\theta)$
High Variance
$J_{train}(\theta)$ will be very low while $J_{train}(\theta) \gg J_{cv}(\theta)$
调参(Adjust $\lambda$)
调整参数$\lambda$，使得$J_{train}(\theta)$和$J_{cv}(\theta)$都处于比较低的水平。
Error along with Training Size
对比上面两张图，可以得出下面两个结论：
- If suffering from high bias, getting more training data will not (by itself) help much.
- If suffering from high variance, getting more training data is likely to help.
Summarization
Suppose you have implemented regularized linear regression to predict housing prices.However, when you test your hypothesis in a new set of houses, you find that it makes unacceptably large errors in its prediction. What should you try next?
