ABSTRACT
Growing surge of misinformation among COVID-19 can post great hindrance to truth, it can magnify distrust in policy makers and/or degrade authorities' credibility, and it can even harm public health. Classification of textual context on social media data relating COVID-19 is an effective tool to combat misinformation on social media platforms. We leveraged Twitter data in developing classification methods to detect misinformation and to identify tweet sentiment. Six fusion-based classification models were built fusing three classical machine learning algorithms: multinomial nave Bayes, logistic regression, and support vector classifier. The best performing models were selected to detect misinformation and to classify sentiment on tweets that were created during early outbreak of COVID-19 pandemic and the fifth month into pandemic. We found that majority of the public held positive sentiment toward all six types of misinformation news on Twitter social media platform. Except political or biased news, general public expressed more positively toward unreliable, conspiracy, clickbait, unreliable with political/biased, and clickbait with political/biased news later in the summer month than earlier during the outbreak. The results provide decision or policy makers valuable knowledge gain in public opinion towards various types of misinformation spreading over social media.