<![CDATA[Forecasting plays an important role in strategic decision-making in various fields, ranging from economics to public management. However, traditional forecasting methods such as ARIMA and linear regression have limitations in handling the characteristics of time series data that are often non-linearly patterned, non-stationary, and experience sudden fluctuations. This research proposes a fuzzy time series (FTS) approach that is more flexible in modeling uncertainty and complex patterns without relying on the assumption of linearity. The purpose of this research is to forecast the population in Sleman Regency using the Yolcu-Egrioglu-Aladag (YEA) higher order fuzzy time series. The data used is annual data on the population of Sleman Regency from 1998 to 2024, a total of 27 observational data divided into 80% training data and 20% testing data. The YEA FTS uses a slice operation to simplify inputs in higher-order models, fuzzification with Fuzzy C-Means and forming fuzzy relations through a Feedforward Neural Network with the Levenberg-Marquardt algorithm, and defuzzification. The results show that the YEA model produces a small Mean Absolute Percentage Error (MAPE) value on the testing data, so it is good for forecasting the population in Sleman Regency. The MAPE value of the YEA FTS is 2.22% (training) and 1.00% (testing). Keywords: : Forecasting, Fuzzy Time Series, High Order Yolcu-Egrioglu-Aladag, Fuzzy C-Means, Feedforward Neural Network, Population]]>
