What's New in 6.2: LDA operators

With the release of Chorus 6.2, several new operators are available in the product. To see posts about all new operators, search the forum for “What’s New in 6.2”, or check out the release notes

The LDA Trainer and LDA Predictor operators have been updated from a previously deprecated version. LDA (Latent Dirichlet Allocation) is an unsupervised text-mining algorithm used to analyze unstructured text data. The new version leverages MLLib’s LDA algorithm with the Online Variational Bayes optimizer. This updates the term-topic distribution adaptively, and is also more memory-friendly, especially on a large number of documents or vocabulary.