preliminary mahout vectors and clustering

apache.sh

#!/bin/bash

MAHOUT_LOCAL=1  # skip hadoop for now
K=3
set -ex

mahout seqdirectory -i issblog -o issblog-seqfiles -ow

#-filter org.apache.lucene.analysis.en.EnglishMinimalStemFilter
#-filter org.apache.lucene.analysis.miscellaneous.StemmerOverrideFilter
#-filter org.apache.lucene.analysis.en.PorterStemFilter


#mahout seq2sparse -i issblog-seqfiles -o issblog-vectors -ow
mahout seq2sparse -i issblog-seqfiles -o issblog-vectors -ow \
	-a org.apache.lucene.analysis.en.EnglishAnalyzer

#-a org.apache.lucene.analysis.en.PorterStemFilter


mahout vectordump -i issblog-vectors/tf-vectors \
	-d issblog-vectors/dictionary.file-0 -dt sequencefile \
	-sort 1 --vectorSize 12 --printKey 1 \
	-o vectors-12

# mahout vectordump -i issblog-vectors/tfidf-vectors -d issblog-vectors/dictionary.file-0 -dt sequencefile -sort 1 --vectorSize 12


mahout kmeans -i issblog-vectors/tfidf-vectors -o issblog-kmeans-clusters -ow \
	-c issblog-init-clusters \
	-dm org.apache.mahout.common.distance.SquaredEuclideanDistanceMeasure \
	-cd 1.0 -k $K -x 20 -cl

mahout clusterdump \
	-d issblog-vectors/dictionary.file-0 \
	-dt sequencefile \
	-i issblog-kmeans-clusters
	-b 10 \
	-n 10
#	-o clusters.txt \

exit 0