Experiment-data-for-UM-S-TM

Introduced 2021-10-12

0.This is experiment data for the following article: @misc{liu2021topic, title={Topic Model Supervised by Understanding Map}, author={Gangli Liu}, year={2021}, eprint={2110.06043}, archivePrefix={arXiv}, primaryClass={cs.CL} }

  1. *.txt files are the data of Table 4 of the paper.

  2. The top lines of all the *.txt files are contents of the artificial documents. Column names are : "Topic", "Distance", "Topic-len", "alpha"/"Noise" , "doc concept-length", and "Votes counter".

3.Coding of file names of *.txt files see "Table 4: Discovered SCOM of six documents". "all_topic" means the candidate topic set is all the topics in a domain.

4.For the "300docs-mentioned-in-section3.2.xlsx" file, its name tells its contents.