Re: Find new cluster centers in k-means for mixed data (containing numbers and nominals)

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

Re: Find new cluster centers in k-means for mixed data (containing numbers and nominals)

Eibe Frank-2
Administrator
Sorry, your message (copied below) must have been accidentally deleted.

SimpleKMeans in WEKA uses the mode of the nominal attribute values that
occur in a cluster as the representative value for that cluster.

Cheers,
Eibe

===========

From: leanhcntt <[hidden email]>
Subject: Find new cluster centers in k-means for mixed data (containing
numbers and nominals)
Date: 23 October 2019 at 3:59:33 AM NZDT
To: [hidden email]


https://weka.8497.n7.nabble.com/Nominal-and-Numerical-Attributes-in-Simple-K-Means-td30528.html

I understood the conversion of data from nominal to number.
Another problem when defining new cluster centers is that the center
attribute values will be averaged from the corresponding properties.
For example: gender: Male, faler => 0, 1, and the distance then gender will
produce a decimal value (eg 1.5). then the distance between elements cannot
be compared with the central value because gender is different from 0 and 1.
For mixed data, how to determine the cluster center?
Thank you so much!



--
Sent from: https://weka.8497.n7.nabble.com/
_______________________________________________
Wekalist mailing list -- [hidden email]
Send posts to: To unsubscribe send an email to [hidden email]
To subscribe, unsubscribe, etc., visit
https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html