Optimasi Metode Single Exponential Smoothing dengan Fuzzy Time Series dalam Prediksi Angka Jumlah Pasien Covid-19

Abstract

The Fuzzy Time Series is a way of capturing patterns from something past to get a chance at future predictions. In making predictions, it can be done by taking data and making a decision. To improve the performance of the Fuzzy Time Series Method, additional methods are needed, currently researchers combine the performance of Single Exponential Smoothing and Fuzzy Time Series in producing better predictions. The research results show that the process of processing datasets using the Single Exponential Smoothing Method speeds up training time to increase accuracy and reduce error values. Due to the repeated process that the Single Exponential Smoothing method has. The calculation results of the Single Exponential Smoothing Method with an alpha value of 0.1 can produce an error value (MAPE) of 6.057739865 and recalculated using the Fuzzy Time Series Method resulting in a higher error value (MAPE), which reaches the number 13.86804695. The results of calculations using the fuzzy time series method also show that the number of recovered patients decreases in the next period.