<html>
<head>
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<div class="moz-cite-prefix">Hello,<br>
<br>
Le 24/07/2016 14:30, Dragonmaw a écrit :<br>
</div>
<blockquote
cite="mid:OLxJFglf3V50KhI264HT1YMPSF6JcG6L0aV98G_8jjxnA0xUttuGVnWf63-tZJ_UoqD6uTFTWz-nWa15WKxkcg==@protonmail.com"
type="cite">
<div>Hello there,<br>
</div>
<div><br>
</div>
<div>i need to align (unequally long) time series, computing their
minimal euclidean distance alignment. I believe this is being
done optimally with dynamic programming. Algorithms exist from
biocomputing or speech recognition, but the packages i've found
only provide functions applicable to either biosequences or
sound files. Since i'm interested in my numerical time series
only, those packages don't solve my problem, which is more
general. I lack the Scilab expertise to adapt the Algorithms for
my case, therefore i'd need some help finding a package that
does exactly that. Perhaps one of you knows such a package.<br>
</div>
<div>Thank you very much,<br>
</div>
<div>Victor K<br>
</div>
</blockquote>
.<br>
It depends on<br>
<ul>
<li>Quantization of values:</li>
<ul>
<li>is there a finite set of possible discrete values <br>
</li>
<ul>
<li>not orderable (like in genomics: A C G U) ?</li>
<li>orderable (ex: 1 3 7 -10)</li>
</ul>
<li>or is the series about a continuous variable?<br>
</li>
</ul>
<li>Are the sampling rates the same for both series?<br>
If it is not the case, do you know the stretching factor between
both, or you don't?</li>
</ul>
<p>...<br>
If your problem is more general, i am afraid that more general and
basic informations are needed about it.<br>
</p>
<p>Regards<br>
Samuel Gougeon<br>
<br>
</p>
</body>
</html>