Hierarchical unsupervised and semi-supervised topic models for sparse count data with CorEx

A workflow for CorEx-based topic modeling

A Bokeh app for exploring Corex topic models

Comparisons among NMF, LDA, BTM & Anchored CorEx models for Topic Discovery of E-commerce Customer Reviews

Project aiming to build an unsupervised topic classifier

Jupyter notebook for topic modeling on the Wikibooks dataset using LDA, LSA, Ida2Vec, and CorEx

Doc2X is a novel topic modelling technique created in June 2019, by yours truly, Pek Yun Ning. It hybridises the older Doc2Vec and Corex topic modelling algorithms to form this all-new algorithm.

NLP on customer tweets to Apple Support to uncover topics using NMF (unsupervised modeling), and classify tweets as product types based on users' initial tweets using CorEx (semi-supervised modeling)

Correlation Explanation (Corex) is a topic modelling technique that is great at identifying 'hidden' topics, or low-frequency-worded but representative topics, very well. It was originally created by Greg Ver Steeg.

This repository contains code for topic modeling of movie reviews using unsupervised (LDA), semi-supervised (CoRex) and supervised (OpenAI) models.

LDA2X is a novel topic modelling technique created in June 2019, by yours truly, Pek Yun Ning. It hybridises the older LDA and Corex topic modelling algorithms to form this all-new algorithm.

LDA2XPand is a novel topic modelling technique created in June 2019, by yours truly, Pek Yun Ning. It hybridises the older LDA, Corex, and Word2Vec topic modelling algorithms to form this all-new algorithm.