PUBLIC SENTIMENTS IN CAGAYAN VALLEY AMIDST THE COVID19 PANDEMIC: TOPIC MODELING USING THE LATENT DIRICHLET ALLOCATION (LDA) APPROACH

Melidiossa  V. Pagudpud,

PUBLIC SENTIMENTS IN CAGAYAN VALLEY AMIDST THE COVID19 PANDEMIC: TOPIC MODELING USING THE LATENT DIRICHLET ALLOCATION (LDA) APPROACH

Abstract:

COVID-19 has a significant impact on the economic, social, and health well-being of people all over the world. This article examines public attitude about the COVID-19 outbreak, notably in Cagayan Valley, Philippines. The Mallet topic modeling tool set was used to apply the Latent Dirichlet Allocation (LDA) Approach to topic modeling, a machine learning technique. The results reveal that topic modeling using the hyperparameters of number of topics (k) = 5, number of iterations = 1000, number of words = 10, and optimization interval (Alpha) = 10 gave a precision of 100%, indicating that the created cluster of words agreed with the human assessment. Similarly, the study found that public sentiment in Cagayan Valley was concentrated on five (5) latent topics: (a) COVID19 testing and case reporting; (b) relief operations; (c) contact tracing; (d) government programs addressing COVID-19; and (e) community quarantine. This examination of public sentiments in the region regarding the COVID-19 pandemic indicates the aspects on which public worries concentrated during the pandemic's struggle. This is important for the government to have a clearer and more comprehensive picture of how the region's provincial government entities can fairly respond to the pandemic.