If the task is classifying documents, what is the intuitive difference between LDA, and other simple classifier such as Naive Bayes, or decision tree? Assume that we feed to all algorithms the same term document matrix.

asked May 15 '14 at 14:14

Dzung%20Nguyen's gravatar image

Dzung Nguyen
2681013

edited May 15 '14 at 14:18

Think of LDA like clustering, which means, no labels are needed. On the other hand, at LDA you should expect to be able to define what a cluster means by setting separation thresholds.

(May 15 '14 at 16:50) eder
Be the first one to answer this question!
toggle preview

powered by OSQA

User submitted content is under Creative Commons: Attribution - Share Alike; Other things copyright (C) 2010, MetaOptimize LLC.