I have two sets of time series data (series1 and series2). Each data set has 20 samples for 20 time intervals (one sample per each time interval). I want to see if these two data sets are significantly different. What test should I use?
More precisely, I am trying to predict the population of people for 20 time intervals. I have used two different methods to make this prediction that is why I have two sets of data. I want to see which prediction is better and more accurate. I also have the real numbers of population to evaluate my prediction. I have calculated the RMSE
for the two series of prediction. The mean RMSE
for series2 is less than series1 but I am not sure if it is a sufficient criteria to say that series2 is better prediction than series1.
Wrong forum :) but you can use the Granger Causality Test or some other t or f-test.
https://stats.stackexchange.com/questions/172226/proving-similarities-of-two-time-series