Topic Modelling in python

This repository contains:

• topic_model.py file which contains code for a topic modelling object.
• The pickled files directory contains all the files needed by topic_model.py
• Topic_modelling.ipynb is a notebook with the details behind the creation of the model in topic_model.py

Note: The user needs to have the spacy and gensim modules installed for this code to work.

Usage

Class: topic_model

Make a new topic model object by instatiating the topic_model

import topic_model as t
model = t.topicMod()

The 'model' object contains a trained LDA model that will find topics for a given document.

The object also has pre-loaded documents in the model.test_documents list. The list consists of 2042 documets from the 20newsgroups dataset.

The model.get_document() method gets a random document from the test documents.

The model.lda_description method takes a text document and finds the topics relevant to it.

Some examples of how to get topics for a document:

model.lda_description(model.get_document())

model.lda_description(model.test_documents[1230])

model.lda_description("Some text I want to get topics from. This text will be preferably at least 100-150 words long")