Skip to main content
. 2023 May 26;9(6):e16757. doi: 10.1016/j.heliyon.2023.e16757

Table 3.

Quantitative indicators for topic modeling.

Token subset Number of tokens Diversity Coherence % of tokens
Full dataset 964,391 N/A N/A 100%
Subject 173976 0.7872 0.135 17.04%
Medium 100779 0.2613 0.531 8.45%
Technique 70690 0.423 0.481 6.33%
Genre 64421 0.357 0.289 6.68%
Mood 26328 0.4681 0.174 2.73%
Tone 14948 0.370 0.567 1.55%
Lighting 102515 0.404 0.510 5.63%
Resolution 25074 0.3769 0.623 2.60%
Artistic References 183427 0.2279 0.371 19.02%
Reception/Popularity 77248 0.1837 0.651 8.01%
Physical attributes of the subject 11862 0.593 0.116 1.23%
Emotional or psychological traits of the subject 6944 0.423 0.348 3.72%
Environment/Setting 42530 0.623 0.244 10.41%
Symmetry/Repetition 4629 0.4067 0.470 0.48%
Depth of field 15430 0.364 0.583 1.6%
Angle 27292 0.190 0.563 2.83%
Message/Meaning 14948 0.364 0.201 1.55%