Deep Learning Based Sri Lanka Crisis Sentiment Analysis - SLCSA

Abstract

The goal of this study is to determine the public's feelings regarding the economic crisis in Sri Lanka by analyzing comments on YouTube videos. Due to the country's ongoing economic problems, Sri Lanka has become a trending topic on social media. The public attitude of Sri Lankans who are going through a difficult moment is something that people all around the world are highly interested in knowing. The suggested technology would be able to recognize public attitudes regarding Sri Lanka's economic problems and assist government organizations in making wise judgments. When a crisis strikes, the majority of people express their opinions on social media. Unless there are some solutions put forth during the pre-crisis period, crises are very difficult to overcome. As they allow information and instant contact between affected individuals, organizations, and the broader public, social media platforms have developed into a crucial part of crisis management. The lack of comprehensive research that examined popular perceptions of the Sri Lankan crisis served as the inspiration for this study's main objective. With the use of deep learning, this study uses sentiment analysis to show how individuals feel about the Sri Lankan issue. Unfortunately, the available dataset for this investigation was minimal, which made it impossible to execute the study. As a result, YouTube comments were used to generate a dataset. The suggested system was created using a CNN (convolutional neural network) along with several kinds of preprocessing methods. In order to develop the system that classifies user-provided content as positive, neutral, or negative, the author used NLP and deep learning approaches.