Hi Guys,
Just wondering if anybody could help me to understand why we shoud deseason the time series data before applying Trend Analysis in Earth Trend Modeller. Anybody?
Thanks in advance.
Cheers
Hi Guys,
Just wondering if anybody could help me to understand why we shoud deseason the time series data before applying Trend Analysis in Earth Trend Modeller. Anybody?
Thanks in advance.
Cheers
Hi Eddie,
Deseason removes the dramatic trends that accompany changing seasons. After removing seasonal trends many less apparent trends are more easily observed. You can read more about this important tool in the IDRISI Tutorial document (chapter 7, exercise 2).
Best,
Andrew Hostetler
IDRISI Technical Support Team
Thank you very much Andrew. If I may share additional comments by Prof. Eastman here for the benefit of others.
"For a non-parametric technique such as the Theil-Sen trend operator or the monotonic tend, it makes no difference. However, for the correlation measure, it makes an enormous difference because we remove a huge source of variance (the seasons). The OLS slope measure should be about the same, but the significance would be affected for the same reason. In general, deseasoning is advised."