Statistics Help @ Talk Stats Forum - Applied Statistics
http://www.talkstats.com/
Statistics in finance, economics, engineering. Actuarial science. Econometrics. Operations research.enThu, 31 Jul 2014 13:13:26 GMTvBulletin60http://www.talkstats.com/images/misc/rss.pngStatistics Help @ Talk Stats Forum - Applied Statistics
http://www.talkstats.com/
Time Series Analysis
http://www.talkstats.com/showthread.php/57024-Time-Series-Analysis?goto=newpost
Thu, 31 Jul 2014 09:39:42 GMTI have data that is spatial and temporal. Each data point at each time contains multiple signals of interest (some are spatially autocorrelated, some...I have data that is spatial and temporal. Each data point at each time contains multiple signals of interest (some are spatially autocorrelated, some are temporally autocorrelated, some are both spatially and temporally autocorrelated and some are not autocorrelated).

At the moment I analyse the data on a pixel by pixel basis by looking at each individual time series at a time (which is ignoring the spatial autocorrelation unfortunately). I do a linear regression of my measurement with respect to 3 other variables (e.g. elevation, time, etc...) to extract the trend from the time series and I get the error information from this also (again ignoring the autocorrelations).

I then perform spatial and temporal filtering to two remaining signals - one is spatially correlated but not temporally correlated, and the other is both spatially and temporally correlated. Then I am left with an estimate of the noise. I am trying to isolate one of the linear trend signals and the temporally and spatially correlated signal from the filtering. So I have a number of signals that have been extracted from the original variable, with some error information available from the regressions.

Is this a sensible method to go about analysing this data? I am not sure it is because the autocorrelations are not being accounted for. Also I ideally want to provide error bars for the values I am trying to isolate but I am unsure of a method to do so - these error bars would also time dependent I imagine.

I have done a lot of Googling to try to find some guidance on the best way to go about this but haven't had much luck so I would really appreciate a nudge in the right direction from anyone on here.

Thanks in advance

Harry
]]>Applied Statisticsharry_http://www.talkstats.com/showthread.php/57024-Time-Series-AnalysisAcf Pacf interpretation for ARMA modeling
http://www.talkstats.com/showthread.php/56982-Acf-Pacf-interpretation-for-ARMA-modeling?goto=newpost
Mon, 28 Jul 2014 20:23:12 GMTHi,
I have trouble interpreting acf and pacf of the stationary series depicted.
Could I receive some suggested interpretations, with focus on...Hi,

I have trouble interpreting acf and pacf of the stationary series depicted.

Could I receive some suggested interpretations, with focus on determining ARMA(p, q) order? Thanks, please let me know if i should give more information.

Edit: I added a spectrum plot. Because if I'm correct the acf is suggesting a seasonal component in my time series?

Kind regards, sander

]]>Applied Statisticss0050506http://www.talkstats.com/showthread.php/56982-Acf-Pacf-interpretation-for-ARMA-modelingNon-parametric methods
http://www.talkstats.com/showthread.php/56981-Non-parametric-methods?goto=newpost
Mon, 28 Jul 2014 20:10:52 GMTA book I am reading now on non-parametrics argues that t test may be still accurate in large sample sizes even if the distribution assumptions are...A book I am reading now on non-parametrics argues that t test may be still accurate in large sample sizes even if the distribution assumptions are not met because of the central limit theorem while even in small samples simulations suggest that it is robust to type I error. They suggest that the choice between non-parametric approaches and t test comes down to power.

But what if you have ordinal not interval data. In this case isn't the t test always invalid and you have to apply a non-parametric approach?
]]>Applied Statisticsnoetsihttp://www.talkstats.com/showthread.php/56981-Non-parametric-methods