SAX implementation in Wolfram Mathematica

Question

I have to deal with time series to find some patterns there. The method selected is SAX (based on PAA).

There's a built-in function to convert a numeric times series into a series of letters with a specific length and alphabet in R: http://rug.mnhn.fr/seewave/HTML/MAN/SAX.html And AFAIK, there's no such function in Wolfram Mathematica.

So, my question is: what is the simplest way to implement SAX in WM? Which functions in which combinations do I use to apply this operation to a time series?

update

More info on transformations:

• PAA - https://jmotif.github.io/sax-vsm_site/morea/algorithm/PAA.html
• SAX - https://jmotif.github.io/sax-vsm_site/morea/algorithm/SAX.html

Answer 1

Well, I managed it by myself.

...
normalized = Standardize[data];
size = 100;
quantile = N[1/26];
quantiles = Table[Quantile[normalized, i], {i, 0, 1, quantile}];
paa = TimeSeriesAggregate[normalized, size];
sax = Table[
      Alphabet[][[FirstPosition[quantiles, 
           SelectFirst[quantiles, # > paa[[i]] &]][[1]] - 1]], {i, 1, 
       Length[paa]}]