Skip to content

Single node topic model learning and inference via method of moments using tensor decomposition. Alternating least squares with pre-processing (a whitening step consists of orthogonalization and dimensionality reduction) is implemented.

License

Notifications You must be signed in to change notification settings

sonichi/TopicModeling

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

TopicModeling

citation:

@article{DBLP:journals/corr/HuangNHVA13, author = {Furong Huang and Niranjan U. N and Mohammad Umar Hakeem and Prateek Verma and Animashree Anandkumar}, title = {Fast Detection of Overlapping Communities via Online Tensor Methods on GPUs}, journal = {CoRR}, year = {2013}, volume = {abs/1309.0787}, url = {http://arxiv.org/abs/1309.0787}, timestamp = {Sat, 25 Oct 2014 03:19:58 +0200}, biburl = {http://dblp.uni-trier.de/rec/bib/journals/corr/HuangNHVA13}, bibsource = {dblp computer science bibliography, http://dblp.org} }

================ Single node topic model learning and inference via method of moments using tensor decomposition. Alternating least squares with pre-processing (a whitening step consists of orthogonalization and dimensionality reduction) is implemented.

Synthetic Data Generator: TopicModeling/SyntheticDataGenerator.m

Data folder is: $(SolutionDir)\datasets\

Input Arguments:

//========================================================================= // User Manual: // (1) Data specs InputArgument 1: NX is the training sample size InputArgument 2: NX_test is the test sample size InputArgument 3: NA is the vocabulary size InputArgument 4: KHID is the number of topics you want to learn InputArgument 5: alpha0 is the mixing parameter, usually set to < 1 InputArgument 6: DATATYPE denotes the index convention. // -> DATATYPE == 1 assumes MATLAB index which starts from 1,DATATYPE ==0 assumes C++ index which starts from 0 . // e.g. 10000 100 500 3 0.01 1 const char* FILE_GA = argv[7]; const char* FILE_GA_test = argv[8]; // (2) Input files InputArgument 7: $(SolutionDir)\datasets$(CorpusName)\samples_train.txt InputArgument 8: $(SolutionDir)\datasets$(CorpusName)\samples_test.txt // e.g. $(SolutionDir)datasets\synthetic\samples_train.txt $(SolutionDir)datasets\synthetic\samples_test.txt const char* FILE_alpha_WRITE = argv[9]; const char* FILE_beta_WRITE = argv[10]; const char* FILE_hi_WRITE = argv[11]; // (3) Output files InputArgument 9: FILE_alpha_WRITE denotes the filename for estimated topic marginal distribution InputArgument 10: FILE_beta_WRITE denotes the filename for estimated topic-word probability matrix InputArgument 11: FILE_hi_WRITE denote the estimation of topics per document for the test data. // The format is: // $(SolutionDir)\datasets$(CorpusName)\result\alpha.txt // $(SolutionDir)\datasets$(CorpusName)\result\beta.txt // $(SolutionDir)\datasets$(CorpusName)\result\hi.txt // e.g. $(SolutionDir)datasets\synthetic\result\alpha.txt $(SolutionDir)datasets\synthetic\result\beta.txt $(SolutionDir)datasets\synthetic\result\hi.txt //=====================================================================

About

Single node topic model learning and inference via method of moments using tensor decomposition. Alternating least squares with pre-processing (a whitening step consists of orthogonalization and dimensionality reduction) is implemented.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • C++ 98.5%
  • C 1.4%
  • Other 0.1%