A Behavioural Analysis of US Election 2020 Using Deep Learning

EasyChair Preprint 8915

Abstract

Technology has evolved into a pivotal driving force for global communication. Twitter, Facebook, and Instagram are the most essential channels for expressing opinions on the daily developments that occur in and around the world. In this work, the tweets of US Election 2020 have been collected from twitter. After pre-processing the tweets, Sentimental Analysis using textblob and Behaviour Analysis using text2emotion are applied. The generation of feature vector is done using n-grams (unigram + bigram) for Machine Learning and neural networks keras for Deep Learning. Finally, the performance of the model is evaluated by using Deep Learning algorithms and then compared with Machine Learning algorithm. The model has achieved an accuracy of 84% using Bi-LSTM.

Keyphrases: Bi-LSTM, GloVe, Sentiments Analysis, behaviour analysis, deep learning, machine learning

@booklet{EasyChair:8915,
  author    = {Mohit Mathur and Ansh Agarwal and Shilpi Gupta},
  title     = {A Behavioural Analysis of US Election 2020 Using Deep Learning},
  howpublished = {EasyChair Preprint 8915},
  year      = {EasyChair, 2022}}