On Friday, March 3, 2023, A. Nafis Haikal conducted the thesis session. In the thesis session, Nafis explained the results of his final project entitled “Modification of Fuzzy Time Series (FTS) forecasting based on a combination of Interval ratio and frequency density Partition”. Besides Nafis, the thesis session was attended by Drs. Bayu Surarso, M.Sc., Ph.D. as advisor I, Dr. Susilo Hariyanto, S.Si., M.Si. as advisor II and as examiners there Farikhin, S.Si., M.Si., Ph.D. and Ratna Herdiana, M.Sc., Ph.D.
In his thesis, Nafis modified Fuzzy Time Series forecasting. Forecasting is one of the important elements in decision making. One method of forecasting that is often used is fuzzy time series. Interval division in fuzzy time series usually divides the interval into equal length, while in this study the interval division will be done using the interval ratio method and then combined with the frequency density partition method. The interval ratio method divides intervals by Ratio. The obtained intervals are then partitioned again using the frequency density partitioning method. This study focuses on the comparison between fuzzy time series method based on interval ratio, fuzzy time series method based on frequency density, and a combination of both methods. These methods are applied to the rubber production data Indonesia, Semarang city daily temperature data, and data CSI 300 index. The forecasting results show that the proposed method provides the best accuracy. The percentage of the average increase in the accuracy of the interval ratio method for rubber production data Indonesia, Semarang city daily temperature data, and data CSI 300 index respectively is 79.37%, 94.5%, and 86.9%. While the frequency density partition is 7.82%, 11.5%, and 76.15%.