Autocorrelogram

2020년 9월 20일

Fast estimation of an autocorrelogram, more commonly called autocorrelation function (ACF). The script sets the maximum lag as 10*log10(N)-1 and sets the autocorrelation at lag 0 to 1.

Length controls the number of past observations of Src to use as input, while Differentiate Src perform first order differencing to Src before calculating the autoccorelations.

The ACF can return a lot of information, for technical analysis, the ACF can be used to determine whether the market is trending or ranging. Without prior processing, we expect trending prices to have a slowly decaying ACF with significant autocorrelation at each lag, while ranging prices will have an ACF that decay faster toward 0.

I won't post a lot on this profile, I'll post most of my future work on the LuxAlgo profile instead, link in the signature.