(numerically robust) Dimension reduction via Jensen-Shannon Divergence & Principal Components
This function is largely a copy of the repsective function in https://github.com/cpsievert/LDAvis/blob/master/R/createJSON.R, however, with a fix to avoid log(0) proposed by Maren-Eckhoff in https://github.com/cpsievert/LDAvis/issues/56
jsPCA_robust(phi)
phi |
matrix, with each row containing the distribution over terms for a topic, with as many rows as there are topics in the model, and as many columns as there are terms in the vocabulary. |
Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.