Perplexity is defined as the inverse of the geometric mean of the log-likelihoods over all iterations of Gibbs sampling.
calculatePerplexity(completeLogLik, log = FALSE)The complete Gibbs sampling history of log-likelihoods for a single celda chain
Set log to TRUE to visualize the log(perplexity) of Celda_CG objects.
The perplexity for the provided chain as an mpfr number
Perplexity is defined in LDA as the exp of the log likelihood of the model divided by the total amount of word tokens. The corresponding perplexity for the celda models are derived in their respective model description documents. These documents will not be made publicly available until after publication of celda.