SENTIMENT DYNAMIC ANALYSIS REGARDING MENTAL HEALTH ISSUES DURING THE COVID-19 PANDEMIC USING K-NEAREST NEIGHBOR METHOD AND ISING MODEL

• Main Author: Kartika Rahmadhiyanty, Destira
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/67155
• Institution: Institut Teknologi Bandung

Complex systems are systems that organically formed by agents whose behavior is unpredictable. However, when these agents are gathered in large numbers, we can observe the collective behavior. This phenomenon can be explained by using sociophysics. One of the social phenomena that has begun to be widely discussed by public since the COVID-19 pandemic started is the issue of mental health. The WHO survey shows an increase in the use of mental health services in many countries during the pandemic. Some changes in aspects of life due to the pandemic have affected people's mental health conditions. This is indicated by the various sentiments regarding mental health issues that appear on social media which trigger interactions between individuals which results the dynamics of sentiment. This interaction resembles a complex system. In this study, sentiment analysis will be carried out by using K-Nearest Neighbor algorithm. Furthermore, we aim to analyze the sentiment or opinion dynamic by using Ising model. The results show that 52% of the data are negative sentiments with the popular keywords “pandemi”, “dunia”, “buruk” and “diganggu”. This shows the high frequency of negative sentiment regarding mental health issues during the pandemic. In addition, it is concluded that the number of population and the sentiment ratio will affect the time required to achieve equilibrium, while extreme events will affect the sentiment changes.