[CONF] Apache Lucene Mahout > fuzzy-k-means-commandline

[co...@apache.org]

Space: Apache Lucene Mahout (http://cwiki.apache.org/confluence/display/MAHOUT)
Page: fuzzy-k-means-commandline (http://cwiki.apache.org/confluence/display/MAHOUT/fuzzy-k-means-commandline)

Added by Isabel Drost:
---------------------------------------------------------------------
h1. Introduction

This quick start page describes how to run the [Fuzzy kMeans| http://en.wikipedia.org/wiki/Data_clustering#Fuzzy_c-means_clustering] clustering algorithm on a Hadoop cluster. More information on the implementation itself is available in JIRA issue [MAHOUT-74|https://issues.apache.org/jira/browse/MAHOUT-74]


h1. Steps

h2. Testing it on one single machine w/o cluster

In the examples directory type:

{code}mvn -q exec:java -Dexec.mainClass="org.apache.mahout.clustering.fuzzykmeans.FuzzyKMeansJob" -Dexec.args="<OPTIONS>"{code}

h2. Running it on the cluster

* In $MAHOUT_HOME/, build the jar containing the job (mvn install) The job will be generated in $MAHOUT_HOME/core/target/ and it's name will contain the Mahout version number. For example, when using Mahout 0.1 release, the job will be mahout-core-0.1.jar
* (Optional) 1 Start up Hadoop: $HADOOP_HOME/bin/start-all.sh
* Put the data: $HADOOP_HOME/bin/hadoop fs -put <PATH TO DATA> testdata
* Run the Job: $HADOOP_HOME/bin/hadoop jar $MAHOUT_HOME/core/target/mahout-core-<MAHOUT VERSION>.jar org.apache.mahout.clustering.fuzzykmeans.FuzzyKMeansJob -i testdata -c -o output -d org.apache.mahout.utils.CosineDistanceMeasure -v 0.1 -x 10 -p 3 -r 3 -a false -m 0.23 -e org.apache.mahout.matrix.SparseVector
* Get the data out of HDFS and have a look. Use bin/hadoop fs -lsr output to view all outputs.

h1. Command line options
{code}
  --input (-i) input                          The Path for input Vectors. Must be a SequenceFile of Writable, Vector.
  --clusters (-c) clusters                    The directory pathname for initial clusters.
  --output (-o) output                        The directory pathname for output points.
  --measure (-d) measure                      The classname of the DistanceMeasure.
  --convergencedelta (-v) convergenceDelta    The convergence delta value.
  --maxIter (-x) maxIter                      The maximum number of iterations.
  --maxMap (-p) maxMap                        The maximum number of maptasks.
  --maxRed (-r) maxRed                        The maximum number of reduce tasks.
  --doCanopy (-a) doCanopy                    Does canopy needed for initial clusters.
  --fuzzify (-m) fuzzify                      Param needed to fuzzify the cluster membership values.
  --vectorclass (-e) vectorclass              Class name of vector implementation to use.
  --help (-h)                                 Print out help
{code}

Change your notification preferences: http://cwiki.apache.org/confluence/users/viewnotifications.action