Tell me more ×
Cross Validated is a question and answer site for statisticians, data analysts, data miners and data visualization experts. It's 100% free, no registration required.
up vote 3 down vote favorite
2

I have a vector which contains several values e.g.

a <- runif(8760)

I wish to perform continuous wavelet transform to the data but the data is required or preferred to be normally distributed. Can anyone provide any information regarding how I could transform the data to be normally distributed. With my own data set I have tried to standardise the data and perform a log transformation but these dont seem to help.

I have tried:

b <- log(a)
c <- scale(a)
share|improve this question
2  
Is the real data uniformly distributed or does it come from some unknown distribution? – Macro Jun 25 '12 at 21:16
If you're trying to coerce a into normal, you could apply a CDF. If you need data from a normal distribution, why don't you use rnorm? – Roman Luštrik Jun 26 '12 at 6:44

3 Answers

up vote 2 down vote

The Power Transformation is pretty good (try low values of $\lambda$)

share|improve this answer
+1 I've used the Box-Cox transformation with success. In R, stat.ethz.ch/R-manual/R-patched/library/MASS/html/boxcox.html – baha-kev Jun 26 '12 at 6:02
up vote 1 down vote

normally, you have to apply normality test to the residual after transformation

share|improve this answer

We're looking for long answers that provide some explanation and context. Don't just give a one-line answer: please explain why you're recommending it as a solution. Answers that don't explain anything will be deleted. See Good Subjective, Bad Subjective for more information.
up vote 1 down vote

Could you elaborate on why you think that the cwt requires your data to be normally distributed? Wavelet transforms get applied to all sorts of stuff which isn't even remotely normal (e.g., images, seismographs, electroencephalograms) and I've never heard anyone complain about that. I took a quick look in the references we have lying around, and none of them seem to mention anything about constraints on the data's distribution either.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.