Path: news.mathworks.com!newsfeed-00.mathworks.com!newsfeed2.dallas1.level3.net!news.level3.com!postnews.google.com!i18g2000prn.googlegroups.com!not-for-mail
From: Nate <nateomason@gmail.com>
Newsgroups: comp.soft-sys.matlab
Subject: [MxN] for principle component analysis (PCA)
Date: Thu, 5 Jun 2008 15:14:27 -0700 (PDT)
Organization: http://groups.google.com
Lines: 17
Message-ID: <a30b7063-ed04-4ac3-9718-8c47fc9bad59@i18g2000prn.googlegroups.com>
NNTP-Posting-Host: 159.212.71.25
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
X-Trace: posting.google.com 1212704068 13452 127.0.0.1 (5 Jun 2008 22:14:28 GMT)
X-Complaints-To: groups-abuse@google.com
NNTP-Posting-Date: Thu, 5 Jun 2008 22:14:28 +0000 (UTC)
Complaints-To: groups-abuse@google.com
Injection-Info: i18g2000prn.googlegroups.com; posting-host=159.212.71.25; 
User-Agent: G2/1.0
X-HTTP-UserAgent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; .NET 
X-HTTP-Via: 1.0 proxy.ihc.com:8080 (squid/2.6.STABLE16), 1.0 
	proxy.ihc.com:8080 (squid/2.6.STABLE16)
Xref: news.mathworks.com comp.soft-sys.matlab:472446



I am trying to create a matrix [MxN] for principle component analysis.
My problem is that my waveforms do not all have the same vector length
N, there is a difference of 15 to 100 data points. I am unsure what
type of algorithm to use, or how in general to adjust the waveforms to
the same vector length N.

I know that this will change the frequency of the signal slightly but
I am not so concerned the, with that slight change.  I am looking for
identifying features over one rotation of the device one rotation may
take a few milliseconds longer so the waveform vector has more data
points. I need those vectors to contain the same number of points.

A thought that I had was to add zeros to the end of each waveform
making them all length N but I am unsure how that would effect PCA, if
any one knows that would also be helpful information.

Nate