perplexity: Optimize the hyper-parameters for LDA

In seededlda: Seeded Sequential LDA for Topic Modeling

Optimize the hyper-parameters for LDA

Description

perplexity() computes the perplexity score to help users to chose the optimal values of hyper-parameters for LDA.

Usage

perplexity(x, newdata = NULL, ...)

Arguments

x	a LDA model fitted by textmodel_seededlda() or textmodel_lda().
newdata	if provided, theta and phi are estimated through fresh Gibbs sampling.
...	additional arguments passed to textmodel_lda.

Details

perplexity() predicts the distribution of words in the dfm based on x$alpha and x$gamma and then compute the sum of disparity between their predicted and observed frequencies. The perplexity score minimizes when the chosen values of hyper-parameters such as k, alpha and gamma are optimal.

Value

Returns a singple numeric value.

See Also

divergence

seededlda documentation built on April 4, 2025, 2:33 a.m.