Autocorrelation

It refers to correlation between values of the same time series at different time intervals, that is, how the value of the t period is related to the value of t-1 period, how the value of t-1 period is related to the value of t-2 period, and so on. In regression modeling, a model is believed to be robust if there is no autocorrelation among residuals (difference between actual value and fitted value), meaning residuals are not correlated with each other. It is also called serial correlation.