RE: [Wekalist] Help about output file with EM Clustering Model..
by Wu, Haibin (CAP, GE China)
how about using option '-p 0'. The result is to give you predicted cluster
for each instance.
From: Kiri Wagstaff [mailto:Kiri.Wagstaff@jhuapl.edu]
Sent: Friday, March 07, 2003 6:00 AM
To: Liliana Calderon
Subject: Re: [Wekalist] Help about output file with EM Clustering
On Fri, 7 Mar 2003, Liliana Calderon wrote:
> I did the training of the EM Clustering model (I did not use a test file).
> Now, I need to review the results provided by that model since I need to
> extract some information to carry out her processes. Using option -d,
> I obtain a file which apparently includes that information; however, when
> open the file, an unreadable (according to my knowledge), coded
> information is shown (please, see below). I would like to know if that
> is common and which is the way to obtain this nformation in a clearly
> readable format.
The -d option is not what you want. It is used to specify a file where
the classifier object (model) will be written. It is not meant to be
human-readable. Instead, you can use it to read the classifier back in
later. I don't know if there's a way to output the model (clusters) in
human-readable format when running Weka from the command line. However,
if you are running it from inside your own Java program, you can just
print it out, e.g.
Classifier EM = new EM();
Kiri L. Wagstaff, Ph.D. (kiri.wagstaff(a)jhuapl.edu) \|/
Science Applications Group, Space Department -O-
The Johns Hopkins University Applied Physics Lab /|\
Wekalist mailing list