topic Re: How to identify seasonality or periodicity in data? in Statistical Procedures

How to identify seasonality or periodicity in data?

AshP — Mon, 25 Apr 2011 17:43:28 GMT

I have a very simple time series related question.

I have data spanning a four month time period. I have one data record for each day, so a total of approx. 120 data points.

All I want to know is whether there is a periodicity/seasonality to the data. Descriptively, it look like there is a weekly periodicity, but how do I identify this statistically?

Re: How to identify seasonality or periodicity in data?

SPR — Tue, 26 Apr 2011 16:25:16 GMT

Hello AshP,

You could use proc UCM:

http://support.sas.com/onlinedoc/913/docMainpage.jsp

Sincerely,
SPR

Re: How to identify seasonality or periodicity in data?

mfisher — Tue, 26 Apr 2011 16:38:41 GMT

The simplest method to model seasonal patterns is to use dummy variables to represent the seasons. With such a short time series, a weekly periodicity is about the longest effect you can detect (i.e., you would not be able to see a monthly effect with only 4 months). A regresion model with the seasonal dummy variables as your inputs will serve you, but if there is any residual autocorrelation remaining, your MSE may seriously underestimate the variance of the error terms, so your p-values could be way off.

I'd use proc autoreg with the dummy variables and ask for the Durbin-Watson statistic to check for autocorrelation. Proc autoreg can handle autoregressive errors, if present.

Mark

Re: How to identify seasonality or periodicity in data?

proctice — Tue, 26 Apr 2011 18:41:08 GMT

Proc Spectra or Proc Arima: Dickey Fuller test, Seasonal Decomposition Plot