Utilization of the Batch Training Method for Predicting Natural Disasters and Their Impacts

Autor: M A Hanafiah, Anjar Wanto, Riki Winanjaya, Harly Okprana, Ni Luh Wiwik Sri Rahayu Ginantra
Rok vydání: 2021
Předmět:
Zdroj: IOP Conference Series: Materials Science and Engineering. 1071:012022
ISSN: 1757-899X
1757-8981
DOI: 10.1088/1757-899x/1071/1/012022
Popis: Indonesia is one of the countries that often experiences natural disasters, including earthquakes, floods, tsunamis, etc. All of this causes losses, both casualties, Broken, and Anguishing for the population. Based on this, this paper is proposed, which aims to predict natural disasters in the coming years in Indonesia, casualties, Broken, and their consequences. This paper is an extension of previous research, which is still an architectural model to predict Indonesia’s natural disasters and their impacts. Model 4-10-1 is the best in this study, which produces 91% accuracy. Based on this architectural model, this paper will predict natural disasters that occur and their impacts for the years to come in Indonesia. The research dataset and algorithms used remain the same, namely the natural disaster dataset for 2008-2019. Resourced from its National Emergency Management Department and the Batch Training algorithm. Specifically, the results of this proposed paper are in the form of a prediction of natural disasters that will occur, dead and disappear, injured, Anguishing and displaced, houses severely Broken, moderately Broken, lightly Broken to submerged, and Broken to facilities and infrastructure such as health facilities, facilities. worship and educational facilities.
Databáze: OpenAIRE