Ensemble Time Series Modified Generalized Regression Neural Network Rainfall Forecasting

Toharudin, Toni dan Caraka, Rezzy Eko dan Darmawan, Gumgum dan Iskandar, Akbar dan Somantri, Oman dan Arnita, A dan Soebagyo, Joko dan Ell Goldameir, Noor dan Asmawati, S. (2018) Ensemble Time Series Modified Generalized Regression Neural Network Rainfall Forecasting. In: 2018 2nd Borneo International Conference on Applied Mathematics and Engineering (BICAME).

Text 9084630 - Published Version Download (48kB)

Abstract

Information about the predictionof the rainy season is vital for the community. The precise and accurate level of predicting the start of the season will certainly greatly assist various community activities in multiple sectors such as transportation, agriculture, forestry, health, public works, and others. SST controls the ability of the oceans to regulate heating and to regulate water distribution. The condition of local SST can be used as an indicator of the minimum amount of moisture in the atmosphere and is closely related to cloud formation in Indonesia. If the cold SST supply of water vapor in the atmosphere will be reduced, on the contrary, if the SST is warmer than average then the water vapor in the atmosphere will tend to be more. The warmer or hotter the sea surface temperature, the higher the availability of water vapor which causes cloud formation and of course the atmospheric conditions will become more humid. Time series data is a group of observations obtained at different time points with the same time interval, and the data sequence is assumed to be interconnected with each other. In this paper, we analyze the rainfall data in Sulawesi, which begins with the formation of spatial correlation and uses modified generalized regression neural network method to forecast. Get the best model with MSE testing values of 2. 77 * 10 -4 and MAD testing of 0.00017 with the number of layer units 5-5-1.

Item Type:	Conference or Workshop Item (Paper)
Subjects:	BD Speculative Philosophy BD Speculative Philosophy > BT Doctrinal Theology
Divisions:	Fakultas Teknik
Depositing User:	Dr Joko Soebagyo
Date Deposited:	17 Aug 2020 13:08
Last Modified:	17 Aug 2020 13:08
URI:	http://repository.uhamka.ac.id/id/eprint/1828

Actions (login required)

View Item