Once you run the regression model with the module 'statsmodels' in Python, you can get the prediction with these confidence intervals: mean_ci_lower mean_ci_upper obs_ci_lower obs_ci_upper
I know what mean_ci means but I have no idea what obs stands for.
obs_ci is the prediction interval, i.e. a confidence interval for a new observation.
It combines the uncertainty about the mean estimate with the randomness of a new observation for given mean. It assumes that residuals are distributed close to normal distribution, uses either normal or t distribution for the prediction interval and is therefore only approximate.
