Time-series analysis of Manila City morbidity data through autoregressive integrated moving average (ARIMA) using R Studio and XLSTAT

• Main Authors: Lenon, Guilbert Christian M., Martinez, Kathleen Jhoanne D.
• Format: text
• Language: English
• Published: Animo Repository 2019
• Subjects: Air quality management—Philippines—Manila; Air—Pollution—Physiological effect; Diseases—Reporting—Philippines—Manila; Environmental Health and Protection
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/18601
• Institution: De La Salle University

Pollution in the city of Manila is increasing and the rates for respiratory and cardiovascular diseases are rising as well. Development of preventive programs and damage controls can be attained with the guidance of forecasting through time-series analysis. Such analysis is heavily dependent on the data provided by the local agency unit, therefore, integrating health data organization with time-series analysis can be of great benefit to the community subjected in a polluted setting. Researchers used R studio and XLST AT in performing autoregressive integrated moving average (ARIMA) modeling in forecasting the respiratory and cardiovascular diseases responsible for the morbidity in Manila. The study used weeks of 2012 - 2016 for weekly and by gender forecasting, and annual record from 2003 - 2012 for yearly and optimized forecasting. The processes resulted to models which can be calibrated for future use. Upon forecasting for weekly and yearly time-series, predictions showed that longer observations with heavier weights or value provide more accurate and reliable results. Moreover, forecasting by gender illustrated that males are more susceptible to diseases than females. Results also depicted that acute respiratory infection (ARI), pneumonia, and tuberculosis are the most prevalent among the diseases selected for this study. In the course of the study, R Studio provided more reliable results compared with XLSTAT in terms of MAPE values. For future use, ARIMA models should be calibrated every now and then as the data increases in order to produce more reliable, and more accurate results. It is also recommended, especially to local agencies, to maintain a consistent profiling of morbidity and mortality in the city. Organization of health data records is vital in analysis, therefore, regular recording and census must be done.