preparation of data for kmeans

[jose]

hello, my question is this ..
I want to use the k Means to group data from texts. such as the following:
ugly cat
cute dog
cat intestine
barking dog
cow gives milk
cow is in the valley
chicken eats corn ..
etc.
I want to group data by animals .. I can do this?
if only I have the text .. What are the steps I have to do to use kmeans? .. How do I prepare the data?
thanks.

[Marius]

Hi,

do you know the video tutorials about RapidMiner on http://rapid-i.com/content/view/189/212/lang,en ? There are also some videos about text mining.

Best, Marius

[jose]

Thanks Marius. The truth is that the videos did not help me much. I did not know what attributes using the k-means to generate the classification of texts. I generate the frequency matrix and then apply the k-means, this genre I cluster. And the classification was relatively good. I wanted that I would group by topic.

My question is:
Is there another way of classifying or grouping texts? for best results.

[dudester]

There is also Tree cluster methods as well as EM (Expectation Maximization) clustering; see variable clustering methods and data mining.  I don't know if such an operator (EM) exists in DM...
You may have to get creative.