Omicron virus emotions understanding system based on deep learning architecture.

Area of Science:

• Natural Language Processing (NLP)
• Deep Learning
• Sentiment Analysis

Background:

• Public opinion mining and sentiment analysis are crucial for understanding public attitudes in marketing and healthcare.
• The Omicron variant's rapid spread in late 2021 generated significant global attention and public discourse, particularly on social media.

Purpose of the Study:

• To develop and implement a sentiment analysis framework to explore global attitudes towards the Omicron variant.
• To categorize public sentiment as positive, neutral, or negative based on social media discussions.

Main Methods:

• Utilized natural language processing (NLP) techniques within deep learning.
• Employed a Bidirectional-Long-Short-Term-Memory (Bi-LSTM) neural network model combined with a deep neural network (DNN).
• Analyzed textual data from Twitter user tweets collected between December 11-18, 2021.

Main Results:

• The developed sentiment analysis model achieved an overall accuracy of 94.6%.
• Analysis of tweets revealed predominantly negative sentiment (42.3%) towards the Omicron variant.
• Positive sentiment accounted for 35.8% and neutral sentiment for 21.9% of the analyzed tweets.

Conclusions:

• The study successfully applied deep learning NLP techniques to gauge public sentiment regarding the Omicron variant.
• Findings indicate a significant level of public anxiety and negative sentiment associated with the Omicron variant during the specified period.
• The high accuracy of the model demonstrates its effectiveness for real-time sentiment analysis of public health issues.