Forecasting the Amount of Pneumonia Patients in Jakarta with Weighted High Order Fuzzy Time Series

Abstract

Forecasting the amount of Pneumonia patients could help medical practitioners to prepare the required medicines, aid-workers, or even prevent it by sharing knowledge to parents, elders, and smokers. This problem poses great concerns on the lives of many people, therefore, adequate accuracy is required in forecasting. Fuzzy Time Series (FTS) is an alternative way to forecast data. By using ARIMA and Holt's Exponential Smoothing, there are some problems that are difficult to obtain the best model. Using our FTS method, we modified the Cheng algorithm by using higher order (using two or more historical data) to make the accuracy better by seeing the Mean Absolute Percentage Error (MAPE). Data was selected from the amount of Pneumonia Patients in Jakarta from 2008 to 2018. We use R to carryout ARIMA and Holt's Exponential Smoothing. Forecasting's accuracy will decrease if the timeframe between these occurrences is lengthy. As a result of this, we made use of 5 periods which are January until May 2019. The result obtained was compared against ARIMA and Holt's Exponential Smoothing, as well as the MAPE are 9.70%, 16.85%, and 18.55% respectively.