I'm aware that the typical EWMA approach is applied over larger time periods (say for Volatility, where lambda = 94% and all weights add up to 100% for stock returns data from last 5 years). Reference: Investopedia: EWMA

Is there a way to apply EWMA approach for weekly data? For example, I have Mon-Fri observations as 60, 35, 50, 80, 90 and I need to assign more weights to the recent observation (instead of using a simple average). Please suggest an alternative if that suits my situation better.

  • $\begingroup$ 94% is an ad-hoc parameter setting and the analysis behind it is no longer valid as markets became much more efficient. In general EWMA will not filter anything in price data series anymore. I would avoid using it. $\endgroup$ – Cowboy Trader Oct 10 '14 at 16:23
  • $\begingroup$ Thanks. Again, I'm not working on price data series. For e.g., I'm working on aircraft noise and it is quite volatile. My goal is to come up with a weighting scheme that assigns more W to the recent observation (in say weekly data, 5 observations). I was wondering if EWMA is even relevant here. $\endgroup$ – Maddy Oct 10 '14 at 16:39
  • $\begingroup$ EWMA is a filtering method. However it works correctly if the underlying process follows "local level" model. Otherwise it will do more harm than good by introducing artificial frequencies. $\endgroup$ – Cowboy Trader Oct 10 '14 at 16:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.