머하웃 완벽 가이드) - 9장 LDA 알고리즘 돌리기

9장 LDA 돌리는 순서

 

+ $MAHOUT_HOME/bin/mahout seqdirectory 
+ -i reuters 
+ -o reuters-seqdir 
+ -c UTF-8
+ -chunk 64
+ -xm sequential

 

mahout seq2sparse \
-i reuters-seqdir \
-o reuters-cvb -wt tf -seq -nv 

mahout rowid
-i reuters-cvb/tf-vectors
-o reuters-cvb

mahout cvb
-dict reuters-cvb/dictionary.file-0
-ow -i reuters-cvb/matrix/
-o reuters-topics
-k 10 -x 20 -dt topics-output -mt topics-model

mahout vectordump -i topics-output/part-m-00000
-d reuters-cvb/dictionary.file-*
-dt sequencefile
-o path/articles_topics.txt

mahout vectordump -i reuters-topics/part-m-00000
-d reuters-cvb/dictionary.file-*
-dt sequencefile
-o path/topic_word1.txt

아... 이게 뭐여