Skip to main content

Table 3 Coherence, Cv, UMass, and \(C_{UCI}\) scores of topics on \(G'\) against cutting and weighting models

From: Improving topic modeling through homophily for legal documents

Model \(G', w_n, w_j\) Coherence Cv UMass \(C_{UCI}\)
LDA   0.131 0.509 \(-\) 0.457 \(-\) 4.476
RTM \(w \rightarrow \infty\) 0.159 0.611 \(-\) 0.412 \(-\) 1.078
Cutting \(w_n^P \, \ge\) 0, \(w_n^S \, \ge\) 0 0.163 0.616 \(-\) 0.394 \(-\) 1.061
  \(w_n^P \, \ge\) 5, \(w_n^S \, \ge\) 10 0.163 0.619 \(-\) 0.395 \(-\) 1.029
  \(w_j^P \, \ge\) 0.50, \(w_j^S \, \ge\) 0.25 0.167 0.613 \(-\) 0.402 \(-\) 1.057
Weighting \(w_j^P + w_j^S\) 0.165 0.620 \(-\) 0.387 \(-\) 1.031
  \(w_j^P \times w_j^S\) 0.174 0.610 \(-\) 0.419 \(-\) 1.052
  \(T_{H,0}(w_j^P, w_j^S)\) 0.173 0.607 \(-\) 0.416 \(-\) 1.083
  \(T_{Y,1}(w_j^P, w_j^S)\) 0.174 0.622 \(-\)0.371 \(-\) 1.030
  \(T_{E}(w_j^P, w_j^S)\) 0.171 0.614 \(-\) 0.395 \(-\) 1.030
  \(w_s^P + w_s^S\) 0.172 0.619 \(-\) 0.387 \(-\) 1.020
  \(w_s^P \times w_s^S\) 0.170 0.614 \(-\) 0.392 \(-\) 1.050
  \(T_{H,0}(w_s^P, w_s^S)\) 0.171 0.613 \(-\) 0.402 \(-\) 1.082
  \(T_{Y,1}(w_s^P, w_s^S)\) 0.171 0.617 \(-\) 0.399 \(-\) 1.055
  \(T_{E}(w_s^P, w_s^S)\) 0.171 0.620 \(-\) 0.387 \(-\)1.012