You are viewing a plain text version of this content. The canonical link for it is here.

Posted to user@mahout.apache.org by Divya <di...@k2associates.com.sg> on 2010/11/01 10:51:07 UTC

K means clustering with mapreduce

Hi,

 

While running the Kmeans clustering 

I am passing the parameter 

 -xm mapreduce

 With rest of the parameters but my sequence file key and values are not in
IntWritable and WeightedVectorWritable format respectively 

 

As I could see in the method  clusterDataMR of KmeansDriver should output
file key and value should be  in IntWritable and WeightedVectorWritable
respectively.

 

Is there anything am I missing.

 

 

 

Regards,

Divya

Re: K means clustering with mapreduce

Posted by Jeff Eastman <jd...@windwardsolutions.com>.

-xm mapreduce is the default. Both the sequential and mapreduce methods 
produce the same output formats. What is the rest of your CLI 
invocation? Why do you believe you are not getting the correct output?


On 11/1/10 2:51 AM, Divya wrote:
> Hi,
>
>
>
> While running the Kmeans clustering
>
> I am passing the parameter
>
>   -xm mapreduce
>
>   With rest of the parameters but my sequence file key and values are not in
> IntWritable and WeightedVectorWritable format respectively
>
>
>
> As I could see in the method  clusterDataMR of KmeansDriver should output
> file key and value should be  in IntWritable and WeightedVectorWritable
> respectively.
>
>
>
> Is there anything am I missing.
>
>
>
>
>
>
>
> Regards,
>
> Divya
>
>