Furthermore, I have learned that the idea of coherence and several metrics have already been discussed much earlier at a more abstract level and not in direct connection to text mining, etc., e.g., by Eels and Fitelson in "Symmetries and Asymmetries in Evidential Support" and several other authors before (the paper of Röder also refers to some of these authors, check the references). that includes some more detailed information on the metrics used. I'll get to #38 this summer.Īs an addition to the information Tommy has provided, here is the link to a version of the paper of Röder et al. I independently derived probabilistic coherence in 2013. They also use a measure that seems to be identical to probabilistic coherence. He also cited a paper comparing various topic coherence measures that's worth checking out: (Instead of "P(a|b) - P(b)", it should read "P(b|a) - P(b)", for example.) I've opened the issue here: did just implement a bunch of coherence measures in text2vec. However, I messed up the probabilities in the description.
to answer your question: probabilistic coherence is now documented in one of the vignettes. P(a|b) - P(b), P(a|c) - P(c), P(a|d) - P(d)Īnd all 6 differences are averaged together, giving the probabilistic coherence measure.For example, suppose you have a corpus of articles from the sports section of a newspaper.
#COHERENCE SCORE UMASS INTERPRET FULL#
Mimno's measure suffers from promoting topics full of words that are very frequent but statistically-independent of each other. That'll be sometime in the next two years.) (I haven't yet written it up, but it will be part of my PhD dissertation. ( ) Instead, it is a measure that I developed.
CalcProbCoherence does not implement the measure proposed by Mimno et al.
0 kommentar(er)
