This archive contains software that can be used for simulating resource and
tag co-occurrence streams. Furthermore, it contains several simulated data
sets that were used in the paper 

Klaas Dellschaft and Steffen Staab. An Epistemic Dynamic Model for Tagging
Systems. Proceedings of 19th Conference on Hypertext and Hypermedia, 2008 

The paper also contains more detailed information about the
used model and the simulated data sets.




DellschaftStaabHypertext08.pdf
  Paper about the tagging model published in the proceedings of the
  Hypertext 2008 conference.


TaggingModels.jar
  A Java application that was used for simulating the tag streams in the
  subdirectories. The source code is also contained in the jar file and
  can be extracted with any zip utility. There exists a GUI and a command 
  line interface. The following commands can be used for starting the 
  simulations:

  GUI: java -Xmx512m -jar TaggingModels.jar
  Command line: java -Xmx512m -cp TaggingModels.jar de.unikold.isweb.tagging.yulesimon.TaggingSimulator


ajax/*
  This directory contains the necessary input files for simulating
  co-occurrence streams for the tag 'ajax'. Furthermore, the results of several
  simulation runs are contained in this directory.


ajax/ajax_corpusfrequencies.occ
  Word occurrence probabilities in a larger web corpus that contains
  71,525 documents about the topic 'ajax'. Required for starting simulations
  with the TopN-Model.


ajax/ajax_input.stream
  File containing the order of tags in a real co-occurrence stream from
  Delicious for the tag 'ajax'.


ajax/topn_<I>_<BK>_<n>_<h>.*
  Files containing the output of simulation runs with the TopN-Model. The
  name of the file also contains the parameter values used for generating it.
  Files ending with 'stream' contain the raw stream data that were generated.
  Files ending with 'freq' contain the frequency-rank distribution of tags in
  the stream (format: rank tagid frequency). Files ending with 'taggrowth' 
  contain the growth of the set of distinct tags in that stream (format:
  #tas #tags).


blog/*
  This directory contains the necessary input files for simulating
  co-occurrence streams for the tag 'blog'. Furthermore, the results of several
  simulation runs are contained in this directory. For more details see the
  description of the ajax-directory.


xml/*
  This directory contains the necessary input files for simulating
  co-occurrence streams for the tag 'xml'. Furthermore, the results of several
  simulation runs are contained in this directory. For more details see the
  description of the ajax-directory.