Ensemble hybrid model for Hindi COVID-19 text classification with metaheuristic optimization algorithm

Area of Science:

• Natural Language Processing
• Machine Learning
• Public Health

Background:

• The COVID-19 pandemic significantly impacted global populations, leading to widespread social media discussions.
• Understanding public sentiment is crucial for public health initiatives and policy-making.

Purpose of the Study:

• To analyze the public opinion of Indian Twitter users regarding the coronavirus (COVID-19) pandemic.
• To develop and evaluate an advanced sentiment analysis model for Hindi tweets.

Main Methods:

• Utilized Hindi tweets related to COVID-19 as input data.
• Applied natural language processing for feature extraction and Grey Wolf Optimization for feature selection.
• Employed a hybrid Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) model for sentiment classification.

Main Results:

• The proposed hybrid CNN-LSTM model achieved the highest classification accuracy (95.54%), precision (91.44%), recall (89.63%), and F-score (90.87%).
• Outperformed various individual machine learning techniques including Random Forest, SVM, CNN, and LSTM.

Conclusions:

• The developed ensemble hybrid model effectively captures and classifies public sentiment towards COVID-19 from social media data.
• This approach provides valuable insights into public opinion, aiding in pandemic response and management.