Clustering Question

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Clustering Question

Kel Lam
Hi group,

My case has N physicians with each seeing M patients.
One physician could have seen a group of patients, or,
a patient could have been seen by multiple number of
physicians.  In order words, there are overlaps.  Now,
I have the following NxM matrix

              Patient#1 Patient#2 Patient#3 .......
Patient#m
Physician#1       1        0          1     .......  
 0
Physician#2       1        1          1     .......  
 1
Physician#3       0        1          0     .......  
 1
.                 .        .          .     .......  
 .      
.                 .        .          .     .......  
 .      
Physician#n       1        1          0     .......  
 0

"1" indicates previous encouter and "0" otherwise.  My
aim is to identify physician group practice based on
the common patients they see.  Any suggestion on which
R package would best serve this purpose?  Thank you so
much!

Regards,
Kelvin

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Reply | Threaded
Open this post in threaded view
|

Re: Clustering Question

Markus Preisetanz
For the dissimilarity metric I would suggest manhattan, as provided by dist (base package), daisy, agnes (both cluster package), for in your case a common "0" is meaningful - means that both pysicians didn't see the patient.

 

When using complete linkage you can see exactly how many patients (seen or not seen) the pysicians in one cluster have at least in common. If the height goes up too fast so that you would have to extract to many clusters you can use average linkage.

 

For the clustering you can use hclust from the base package, agnes from the cluster package, or, when hclust or agnes run out of memory, clara (see thread  [R] cluster analysis for 80000 observations)

 

sincerely, Markus

___________________

Markus Preisetanz

Consultant

 

Client Vela GmbH

Albert-Roßhaupter-Str. 32

81369 München

fon:          +49 (0) 89 742 17-113

fax:          +49 (0) 89 742 17-150

mailto:[hidden email] <mailto:[hidden email]>



Diese E-Mail enthält vertrauliche und/oder rechtlich geschützte Informationen. Wenn Sie nicht der richtige Adressat sind oder diese E-Mail irrtümlich erhalten haben, informieren Sie bitte sofort den Absender und vernichten Sie diese Mail. Das unerlaubte Kopieren sowie die unbefugte Weitergabe dieser E-Mail ist nicht gestattet.

This e-mail may contain confidential and/or privileged infor...{{dropped}}


______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html