search for: deseason

Hello, I asked this question on the r-finance list server and didn't get a reply. Thought I would try here to. I am trying to deseasonalize some financial time series data and I wanted some feedback on the best methods for doing this. I found two Centered Moving Average and Holt-Winters. Which is better and/or more appropriate for financial time series data in your opinion? I understand that Holt-Winters is a type of exponential...

Dear Users, please help with the following DF test: ===== library(tseries) library(timeSeries) Y=c(3519,3803,4332,4251,4661,4811,4448,4451,4343,4067,4001,3934,3652,3768 ,4082,4101,4628,4898,4476,4728,4458,4004,4095,4056,3641,3966,4417,4367 ,4821,5190,4638,4904,4528,4383,4339,4327,3856,4072,4563,4561,4984,5316 ,4843,5383,4889,4681,4466,4463,4217,4322,4779,4988,5383,5591,5322,5404

....8.10), adehabitatLT (0.3.5), aod (1.3), asypow (2012.04-1), BayesDA (2012.04-1), bayesTFR (1.5-1), Bergm (2.3), BoolNet (1.50), bqtl (1.0-27), cccrm (1.1), censReg (0.5-10), cherry (0.2-6), coefplot (1.1.7), coloc (1.08), CompRandFld (1.0.2), ConjointChecks (0.0.7), copula (0.99-1), crawl (1.3-4), deseasonalize (1.03), digeR (1.3), DistributionUtils (0.5-1), effects (2.1-1), elec.strat (0.1.1), ElemStatLearn (2012.04-0), EVER (1.2), evtree (0.1-2), fanovaGraph (1.1), gbm (1.6-3.2), GeneralizedHyperbolic (0.8-0), GGEBiplotGUI (1.0-1), ggmap (2.0), growcurves (0.1.1), gWidgetsRGtk2 (0.0-81), hash (2.1....

...abble.com> Content-Type: text/plain; charset=us-ascii Nuncio, No, there is no requirement to subtract the mean. It is required that the residuals are N.I.I.D. (ie constant mean and constant variance). If you have an upward trending series, for example, then the series would need to be "deseasonalized" so that it is constant. There are many many steps to doing this right. Email me at sales at autobox.com to hear more. Tom -- View this message in context: http://r.789695.n4.nabble.com/arima-tp2993543p3054206.html Sent from the R help mailing list archive at Nabble.com. ------...