minhash在mahout小实验

随便写几个文本文件到examples/minhash-test/

执行以下命令：

unset HADOOP_HOME
unset HADOOP_CONF_DIR
bin/mahout seqdirectory -c UTF-8 -i examples/minhash-test/ -o minhash-test-seqfiles
bin/mahout seq2sparse -i minhash-test-seqfiles/ -o minhash-test-vectors -ow

rm -rf minhash-test-output/
bin/mahout org.apache.mahout.clustering.minhash.MinHashDriver -i minhash-test-vectors/tfidf-vectors/ -o minhash-test-output -mcs 1 -mvs 1

查看minhash-test-output发现相同文档已经聚集到一起了

参考：

mahout的官方minhash wiki -- https://cwiki.apache.org/MAHOUT/minhash-clustering.html

mahout in action第八章8.3节介绍了如何将文档表示为向量，照着做就可以了