G-LDA (Greek LDA)

Description

G-LDA is a project about Greek topic modeling/clustering. It applies Latent Dirichlet Allocation(LDA) extended with some modifications to work well on Greek text data. For this purpose, it uses elasticsearch-skroutz-greek-stemmer for stop-word removal and stemming. The LDA algorithm is implemented with gensim.

REST API

G-LDA also comes with a REST API in python flask that makes its use very simple and flexible. It works with such plain GET/POST requests. For example, it returns the topic of a given document just by giving
CURL -X GET -G 'http://localhost:5000/classifier' -d text=YOUR_INPUT_DOC

Deployment Instructions

Instructions for deployment will be added soon!

Contribution

G-LDA is a newly project and it may suffer from several issues. So feel free to share your ideas(and code) and contribute!