Publication Date

Summer 6-22-2021

Degree Type

Master's Project

Degree Name

Master of Science in Computer Science (MSCS)

Department

Computer Science

First Advisor

Robert Chun

Keywords

COVID-19, Sentiment Analysis, Sentiment140, Deep Learning, Convolutional Neural Networks, Long-Short Term Memory

Abstract

Sentiment analysis is a method of understanding the user sentiment expressed in the form of text. Social media is the best place to capture the public's opinion regarding how they feel about current events. The Corona Virus Disease-2019 (COVID-19) is one of the worst pandemics we have experienced so far. An important observation is that this pandemic has not only affected the public's physical health but also took a toll on their mental health. Reddit is a social news discussion site where people discuss topics around current affairs in smaller groups called subreddits. The project's primary focus is to build a deep learning model that can classify and help analyze user sentiments about Covid-19 on Reddit. The model has been built by evaluating the performance of different classifiers on the Twitter dataset, called Sentiment140. Experiments with varying feature combinations have been evaluated on deep learning models, including Convolutional Neural Networks (CNN) and Long-Short Term Memory (LSTM). The idea is to build a model by combining the best of both architectures. LSTM excels at storing the forward information, whereas CNN can capture the local features. After reviewing these experiments, the best-performing model has been used to classify and analyze the sentiment of the Reddit users over different changes due to the Covid-19 pandemic. Overall, there have been some interesting changes in user reaction trends for posts related to Covid-19 under each subreddit over thirteen months, starting from Mar '20 to Mar '21.

Share

COinS