MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

We have financial some data (500-1000 samples), which is not normally distributed (well known fact from the literature). I have some ideas to do parametric transformations of this data (using some other data) to produce "adjusted" series. My goal is to find a transformation that makes the series normally distributed (with mean 0 and std deviation 1). What is the most appropriate statistic and corresponding test to optimize my parameters and determine if the outcome can be considered normally distributed?

I apologize if this question is too basic for some - I come from financial mathematics and my statistics knowledge is rather limited.

share|cite|improve this question
I think you have much better chances of getting a good answer by posting to <>; instead of here. – Angelo Jan 1 '12 at 12:50
Ah thanks @Angelo, I didn't realise this other site existed! – Grzenio Jan 1 '12 at 13:54
up vote 1 down vote accepted

The Anderson-Darling test is considered one of the best tests for normality, I think.

share|cite|improve this answer

Sort all your data points in increasing value. If you have $n$ point (1 to $n$), transform point $i$ into $\mathcal{N}^{-1}\left( \frac{i-1/2}{n} \right)$. It does exactly what you want but it's dumb because it's not going to have a lot of predictive power.

One of the main source of non-normality in financial time series is heteroskedasticity. Model a stochastic variance first, then work on the residuals.

share|cite|improve this answer

Your Answer


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.