0

I am currently analyzing time series data relating to financial observations (revenue and number of deposits in particular). I would like to smoothen the data because a lot of outliers are present.

Which is the most suitable smoothing filter in this case? I have read about Spencer's 15 point moving average, Henderson's moving average, exponential smoothing and a number of other techniques but I am not sure which would be best to stabilize the data.

user120768
  • 375
  • What makes you think that there is some most suitable filter to remove outliers in financial data? – user21820 Feb 12 '16 at 13:56
  • I am just trying to see which filter I should use. Or in other words, is there a way to check which filter fits my data better? ( which filter would retain as much information as possible while removing any large outliers that are effecting the predictions in a negative way?) – user120768 Feb 12 '16 at 14:02
  • Define "fits" and "better" and "information" and "negative way". By the way it is "affecting" not "effecting". – user21820 Feb 12 '16 at 14:04
  • And if you still don't see what's wrong with your question, see https://en.wikipedia.org/wiki/Anscombe%27s_quartet. I think the best-fit line fits all four data sets very well! – user21820 Feb 12 '16 at 14:06
  • 1
    I understand what you are trying to say, thank you for your help – user120768 Feb 12 '16 at 17:46

0 Answers0