Comparison of two quotation charts with non-linear distortions on the X-axis - page 12

 
alsu: patterns based on DTW?
no, nothing worked for me with DTW, went back to fractal studies (fractal sequences, not fractal indicator)
 
IgorM:
no, nothing worked for me with DTW, went back to fractal studies (fractal sequences, not fractal indicator)


any particular grain to generate a fractal?

and another question - are you interested in fractal compression algorithms by any chance?

 
alsu: Any particular grain to generate a fractal you use?

Well, the fractals literally turned out sort of by themselves :), the case was something like this: I read somewhere about the TS by candlestick ranges - I turned it around, it works then not, just at that time I was interested in non-standard TFs, I tried to display candlestick ranges for all possible periods for all available TFs, I got https://c.The only thing I noticed is that the price moves 1/3 or 2/3 of it at higher timeframes, you won't be able to guess which one it is coming from. Then somewhere in the discussion I think Sorento? told me to google "bloody stairs" Googled http://fractalworld.xaoc.ru/Cantor_set indeed my charts are very similar, well that's actually why fractals.

I started to gather statistics with historical data and it turned out that price approximately 30% of history repeats its movement style from the observer's side, there is no definite order (i.e. pattern 1, then 2, then 3... the constant mixing of 123, 213,331) and the rest 60% (10% of holes and curve history)) for some unknown reason - they never have any repeats in history

and another question - have you ever been interested in fractal compression algorithms?
i did, i have ready wavelet transforms somewhere, but i haven't got back to them yet - i don't have time to embrace them all