example.settings

# corpus options
pickle sotu.pickle
corpus  /net/roi/okuda/state_of_the_union/speech_files_tot/*
labels  /net/roi/okuda/state_of_the_union/ankura_timestamps.data
# corpus  /net/roi/okuda/data/amazon/amazon.txt
# labels  /net/roi/okuda/data/amazon/amazon.response

# filter options
stopwords   /net/roi/okuda/data/stopwords.txt
rare    5
common  1500
smalldoc    5
pregenerate YES

# experiment options
group   slda_random
seed    531
select  random
testsize    200
startlabeled    50
endlabeled  100
increment   10
candsize    500

# model options
model   slda
cseed 531
numtopics   20
alpha   0.1
beta    0.01
var 0.01
numtrain    15
numsamplestrain 5
trainburn   50
trainlag    50
numsamplespredict   5
predictburn 100
predictlag  5