Predicting death risk analysis in fully vaccinated people using novel extreme regression-voting classifier

Area of Science:

• Computational biology and bioinformatics
• Public health and epidemiology
• Machine learning in healthcare

Background:

• COVID-19 vaccination campaigns face public skepticism due to misinformation regarding adverse events, illness, and mortality.
• Rumors surrounding vaccine safety can significantly undermine public trust and hinder vaccination efforts.
• Developing reliable methods to assess post-vaccination risks is crucial for informed public health strategies.

Purpose of the Study:

• To develop and evaluate machine learning models for predicting patient outcomes after COVID-19 vaccination, specifically focusing on survival status.
• To analyze the efficacy of an Extreme Regression- Voting Classifier model against other machine learning and deep learning approaches for prognosis.
• To assess model performance on both original and balanced datasets, utilizing Synthetic Minority Oversampling Technique (SMOTE) for improved generalization.

Main Methods:

• Utilized Term Frequency-Inverse Document Frequency (TF-IDF), Bag of Words (BoW), and Global Vectors (GloVe) for feature extraction.
• Compared an Extreme Regression- Voting Classifier with deep learning models including Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), and Bidirectional Long Short-Term Memory (BiLSTM).
• Performed experiments on original and SMOTE-balanced datasets to evaluate model robustness and performance.

Main Results:

• The proposed Extreme Regression- Voting Classifier combined with TF-IDF achieved an accuracy of 0.85 on the SMOTE-balanced dataset.
• For binary classification tasks, the voting classifier demonstrated state-of-the-art performance with an accuracy of 0.98.
• The study highlights the effectiveness of the proposed model in predicting post-vaccination outcomes.

Conclusions:

• Machine learning, particularly the Extreme Regression- Voting Classifier with TF-IDF, shows significant promise in predicting COVID-19 vaccine adverse event outcomes.
• Accurate prediction models can help mitigate the impact of misinformation and build public confidence in vaccination programs.
• Further research and validation are warranted to integrate these predictive tools into public health surveillance systems.