The COVID-19 pandemic highlighted the importance of sustainable development, particularly in information and communication technology. In Nigeria, the government’s distribution of relief packages and palliatives to vulnerable citizens was met with mixed reactions on social media. This study aimed to analyze public sentiment on the Nigerian government’s COVID-19 palliatives distribution using machine learning algorithms.

Background

The pandemic reversed the United Nations’ progress in lifting people out of poverty, with 99 million people pushed into poverty in 2020. Nigeria, with 200 million people, had 136,000 infected and 1,630 deaths. The government claimed to have spent billions on relief aid, but citizens expressed concerns about transparency and accountability.

Methodology

This study constructed a Nigerian Local English Twitter Emotion dataset (Pidgin English) and experimented with six machine learning algorithms to classify aggregated citizens’ emotions. The algorithms used were Multinomial Naïve Bayes, Support Vector Machine, Random Forest, Logistics Regression, K-Nearest Neighbor, and Decision Tree.

Results

The Support Vector Machine outperformed other classifiers with 88% accuracy. The “disgust” emotion class had the highest number of counts, indicating public dissatisfaction with the palliatives’ distribution. Correlation analysis showed a significant relationship between “joy” and “fear” emotions, suggesting excitement about the palliatives but concern about inequality and transparency.

Conclusion The study revealed that public emotions on COVID-19 support and relief aid packages’ distribution in Nigeria were largely negative. The results emphasize the importance of considering public opinion in governance, particularly during emergency responses. This study contributes to filling the research gap on emergency response-related studies and provides insights for policymakers to improve transparency and accountability.