Sentiment Analysis of COVID-19 using Multimodal Fusion Neural Networks

Abstract

The purpose of this study creates a Sentiment Analysis model of COVID-19 using Multimodal Fusion Neural Networks in real time to model and visualize COVID-19 in Indonesia. This study obtained 87 percent accuracy using the Multimodal Fusion Neural Networks model, a higher 5 percent than the benchmarking model Convolutional Neural Networks. This study proves that the sentiment model built is quite promising and relevant to be implemented.