See Wikipedia, Latent Dirichlet allocation.

See Wikipedia, Hierarchical Dirichlet process.

• Compiler that supports C++11
• Boost C++ Libraries
  *NOTICE* Boost C++ Libraries shoud be built with the C++11 compiler.

See --help.

The 1st line: the number of docs
The 2nd line: the number of vocabulary
The 3rd line: the number of words (*NOTICE* not NNZ, the number of nonzero counts in the bag-of-words)
The following lines: docID wordID count

line number = wordID

UCI Machine Learning Repository: Bag of Words Data Set

MIT License
Copyright (c) 2012 Tsukasa ŌMOTO(@henry0312)

• Mr. Shuyo Nakatani(@shuyo) / Cybozu Labs Inc.
  I consulted his implementation, https://github.com/shuyo/iir/tree/master/lda.
• Mr. Hiroki Taniura(@boiled_sugar, https://github.com/boiled-sugar)
  I had my Enlgish translation corrected.
• Mr. Jan Ekström(@jeebjp, https://github.com/jeeb)
  English adviser
• Mr. Motofumi Oka(@mtfmk, https://github.com/chikuzen)
  I referred to his configure and Makefile.