http://en.wikipedia.org/wiki/Prediction_interval
http://en.wikipedia.org/wiki/Confidence_interval

2# warren_619
thanks

不明白，可否翻译一下？~~~~(>_<)~~~~

If you ask the difference in regression analysis, then think it the following simple way.
Y=X*beta+delta
Confidence interval considers only the delta.
When you want to predict, you estimate beta first. Since beta is a linear combination of Ys, beta can be looked as random variable. The prediction Y=X*beta_hat+delta. Now you should consider the variance brought by delta and beta.

Thank you!