/Sentiment-Analysis

Primary LanguageJupyter Notebook

Sentiment-Analysis (Emotion Detection)

This Project has two parts:

  1. Emotion Detection from YouTube Video.
  2. Live Emotion Analysis of speech.

Detailed description:

  • Sentiment analysis of YouTube videos as well as Live speech was accomplished. The aim was to predict emotion out of a spectrum of 5 emotions. In this project we implemented Ensemble method of Vector Space Model (VSM)(Implemented from a research paper) and LSTM. Also, we designed Web-App to fulfill the aim of building an end to end application.

Internal Working of Algorithm:

Results

  • We tested the ensemble algorithm on multiple YouTube videoes.
  • Here is a sample of our test. The video was about Rape situation in India.

  • Here the prominent emotions are Anger and Disgust which reflects the nature of video.