Final Project - Aspect Based Sentiment Analysis

Description

  • Evan

  • 2020

  • Data acquisition from Twitter and Kateglo
  • Comparing various recurrent and attention base model, data augmentation, and hyperparameter optimization (e.g., word embedding, optimizer) for Aspect-based sentiment analysis
  • Achieving 72.26% test accuracy for 3 label (positive, neutral, and negative) and 81.9% for 2 label