ERROR-CORRECTION MODEL ANALYSIS OF GOOGLE TRENDS AND DENGUE FEVER INCIDENCES IN BANDUNG

• Main Author: Theresa Marlen Sahetapy E, Jane
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/38937
• Institution: Institut Teknologi Bandung

According to theWorld Health Organization (WHO), dengue fever has been a major health issue for people in all tropical and sub-tropical areas in the world, including Indonesia. In the past 50 years, there has been a high increase in dengue cases in many endemic countries. Surveillance of dengue cases is essential in early detection of dengue fever in order to execute the proper preventive actions necessary and avoid possible outbreaks. This paper will investigate the possibility that Google Trends can be used in the early detection of dengue fever. With Google as one of the most used online platforms in Indonesia, the popularity of a certain keyword in Google might have an affect on happenings in the real world. Google Trends also contains real time data while data on dengue cases takes time to process. If Google Trends data can be used to detect future dengue cases, then it will be useful in taking further preventive action before outbreaks occur. In this paper, the method used is cointegration. By examining the stationarity of the popularity of keywords related to dengue fever in Google and the number of weekly dengue cases in Bandung as well as the linear combination of the two variables, it is shown that the two are cointegrated. Hence, an error-correction model (ECM) can be built to predict deviance from the equilibrium condition. It is also shown that popularity of dengue fever in Google Granger-causes the number of dengue cases, but not the other way around. This is further shown by examining the impulse response function and variance decomposition. The results are expected to give further information to the health official in the preparation of dengue intervention.