Syaharuddin, Dewi Pramita, Toto Nusantara, Subanji
Many methods of smoothing for Alpha-Beta parameters in optimization of error values such as MAPE and RMSE of the Holt method. But only able to provide value from the predicted results, can’t give a description of the trend resulting from different data. Theref ore, the study attempted to display simulations and predictions by involving three types of data i.e., temperature data as cycle data (period), HDI data as data with uptrend, and poor population data as data with a downtrend. The simulation results show that the higher the alpha value given, then the higher the accuracy rate of the predicted outcome. While the beta values in the simulation data cycle are obtained smaller errors (MAPE and RMSE) if the beta values are getting smaller, whereas in data with up and downtrends indicates a small error value if the beta value is getting higher. The matchi ng alpha value ranges from 0.7 to 0.9, while the beta value is better if it approaches 0.9. © IJSTR 2019.
University of Muhammadiyah Mataram, NTB, Indonesia; State University of Malang, Indonesia
