CODEWITHSUNDEEP
pythonnlpnltk
Chathuri Fernando
Chathuri Fernando

Reputation: 972

Building a Training Classifier Python with NLTK

Currently I'am working on a project which is emotions detecting (happy,sad etc) from text in a chat application using Python and NLTK. I'am not much familiar with NLP and Python. As a basic way, I hope to use keyword based method. In that case I have to make a emotional keyword list under each emotional state and need to find whether there is any emotional keyword in given sentence and identify the relevant emotional state accordingly. So what I need here to know, do I need to create a training data set and feature list to do that task, If yes, how can I do it. Please help me.

Upvotes: 0

Views: 537

Answers (1)

Brian O'Donnell
Brian O'Donnell

Reputation: 1876

You will need a set of words that have been labeled. One place to start is the AFINN sentiment dictionary which is a large set of words that have been manually labeled. The slides by Wei-Ting Kuo shows how to use the AFINN word set.

Laurent Luce's blog walks through the entire sentiment analysis process using Tweets although he starts with a labeled training set.

enter image description here

Also take a look at NLTK's 'How To' on sentiment analysis

There are a number of emotion data sets that may help at https://www.w3.org/community/sentiment/wiki/Datasets#Emotions_datasets_by_Media_Core_.40_UFL.

Upvotes: 1

Related Questions