distributeTopics: Distribute topics from LDA models onto new corpora

In Chrisss93/cnn.stories: Utilities for CNN-stories project

Description

Using an iterator from new corpora, a document-term matrix compatible with existing LDA models is created. The transform method from the LatentDirichletAllocation handles the rest

The iterator from new corpora must be transformed to take into account collocations present in the LDA models. Moreover, any terms present in the new corpora's document-term matrix, not present in the LDA models must be dropped.

Usage

distributeTopics(articles, mongo)

Arguments

articles	Used to create corpora.
mongo	A mongo connection.

Value

A list

Chrisss93/cnn.stories documentation built on May 23, 2019, 7:31 a.m.