Skip to content

Scorpi35/Message_Emotion_Detector

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

40 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Emotion detection from Messages

Project Description

Sentiment Analysis tends to evaluate 3 basic emotions:- Positive, Neutral & Negative. I tended to take things little farther and classify the text respective to 7 emotions:- Anger, Disgust, Fear, Gult, Joy, Sadness and Shame. This repository contains the dataset for training a model, basic sequential training procedure and testing of the model. The model has been illutratively implemented in a website where emotios behind messages are detected.

Website

http://emotion-detection-messenger.herokuapp.com/ (Note:- Since the model has to be loaded, it might take some time.)

Key Concepts

  • Sentiment Analysis
  • Emotion Detection

Dataset

Tools and libraries

  • Language:- Python
  • IDE:- PyCharm
  • Libraries:- nltk, numpy, tensorflow, numpy, keras, sklearn, pandas

Algorithm

  1. Pre-processing
  2. Tokenization
  3. Generate word embeddings
  4. Training LSTM Model with bidirectional layers

Training Description

LSTM

  • Model:- Sequential, 1 Embbedding layer, 3 Bidirectional layers, 2 Dense layers
  • Optimizer:- Adam
  • Epochs:- 150
  • Batch Size:- 128
  • Validation Split:- 0.2

Future Extension

  • Integration of Plutchik's Wheel of Emotions
  • Analyzing context of conversation
  • Introducing time feature

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages