Struktur Peramalan System Multi-Model untuk pemodelan matematika pada Forecast Indeks Pembangunan Manusia Provinsi Bali

The purpose of this research is to develop a product was called Forecasting System Multi-Model (FSM) determine the best method in forecasting system by constructing several methods form Graphical User Interface (GUI) Matlab. It done all indicator accuration find mathematical model time series data certain period. In simulation phase, used Human Development Index (HDI) Bali Province 2010 - 2017 predict HDI 2018. tested were Moving Average (SMA, WMA and EMA), Exponential Smoothing Method (SES, Brown, Holt, Winter), Naive Method, Interpolation (Newton Gregory), Artificial Neural Network (Back Propagation). Then models/methods evaluated see level accuracy each based on value MAD, MSE, MAPE. Based result from 10 known that Holt most accurate with prediction 2018 equal 67,45 MAPE respectively 0.22654, 0.075955 0.34829.