Setting LDA model hyperparameters in wowpal wabbit

Question

I am an ordinary, ordinary, everyday user of Spark. The Spark LDA has hyperparameters that denote

docConcentration: Hyperparameter for previous distributions of documents by topic. Currently, there should be> 1, where higher values ​​contribute to smoother conclusions. topicConcentration: Hyperparameter for distributions on previous topics through terms (words). Currently, there should be> 1, where higher values ​​contribute to smoother conclusions.

which corresponds to the parameters $ \ alpha $ and $ \ beta $ usually specified in the literature, for which (and $ k $ is the number of topics) the LAR model logarithmic likelihood function is optimized during the convergence process.

Does anyone know if there is any option to set such arguments / parameters earlier in the vowpal wabbit LDA model?

Answer 1

Mark this description of vw lda. ! I think the parameters mentioned on slide 13 may be the ones you are looking for.