Abstract:The pandemic COVID 19 has altered individuals' daily lives across the globe. It has led to preventive measures such as physical distancing to be imposed on individuals and led to terms such as 'lockdown,' 'emergency,' or curfew' to emerge in various countries. It has affected society, not only physically and financially, but in terms of emotional wellbeing as well. This distress in the human emotional quotient results from multiple factors such as financial implications, family member's behavior and support, country-specific lockdown protocols, media influence, or fear of the pandemic. For efficient pandemic management, there is a need to understand the emotional variations among individuals, as this will provide insights into public sentiment towards various government pandemic management policies. From our investigations, it was found that individuals have increasingly used different microblogging platforms such as Twitter to remain connected and express their feelings and concerns during the pandemic. However, research in the area of expressed emotional wellbeing during COVID 19 is still growing, which motivated this team to form the aim: <i>To identify, explore and understand globally the emotions expressed during the earlier months of the pandemic COVID 19 by utilizing Deep Learning and Natural language Processing (NLP).</i> For the data collection<i>,</i> over 2 million tweets during February–June 2020 were collected and analyzed using an advanced deep learning technique of Transfer Learning and Robustly Optimized BERT Pretraining Approach (RoBERTa). A Reddit-based standard Emotion Dataset by Crowdflower was utilized for transfer learning. Using RoBERTa and the collated Twitter dataset, a multi-class emotion classifier system was formed. With the implemented methodology, a tweet classification accuracy of 80.33% and an average MCC score of 0.78 was achieved, improving the existing AI-based emotion classification methods. This study explains the novel application of the Roberta model during the pandemic that provided insights into changing emotional wellbeing over time of various citizens worldwide. It also offers novelty for data mining and analytics during this challenging, pandemic era. These insights can be beneficial for formulating effective pandemic management strategies and devising a novel, predictive strategy for the emotional well-being of an entire country's citizens when facing future unexpected exogenous shocks.

A Deep Learning–based Approach for Emotions Classification in Big Corpus of Imbalanced Tweets

Leveraging distant supervision and deep learning for twitter sentiment and emotion classification

Deep learning for emotion analysis in Arabic tweets

DeepEmotex: Classifying Emotion in Text Messages using Deep Transfer Learning

Temporal Convolutional Network-Enhanced Real-Time Implicit Emotion Recognition with an Innovative Wearable fNIRS-EEG Dual-Modal System

Emotion recognition of social media users based on deep learning

Exploring Deep Neural Networks and Transfer Learning for Analyzing Emotions in Tweets

Multi-class Emotion AI by reconstructing linguistic context of words

Applying and Understanding an Advanced, Novel Deep Learning Approach: A Covid 19, Text Based, Emotions Analysis Study

Basic and Depression Specific Emotion Identification in Tweets: Multi-label Classification Experiments

Emotion analysis of Arabic tweets using deep learning approach

EmoDNN: Understanding emotions from short texts through a deep neural network ensemble

Emotion Recognition on Social Media Using Natural Language Processing (NLP) Techniques

Emotion Detection and Analysis on Social Media

Deep learning based emotion analysis of microblog texts

A Novel Deep Learning Technique for Detecting Emotional Impact in Online Education

Sentiment Classification of Tweets with Explicit Word Negations and Emoji Using Deep Learning

An investigation into the deep learning approach in sentimental analysis using graph-based theories

Text-Based Emotion Recognition Using Deep Learning Approach

An Intelligent High Performance Automatic Sentiment Analysis Model Creation using Deep Convolution Neural Network

A Deep Neural Network Model for the Detection and Classification of Emotions from Textual Content