Author/Authors
aksoy, bekir isparta uygulamalı bilimler üniversitesi - teknoloji fakültesi - mekatronik mühendisliği bölümü, Isparta, Türkiye , salman, osamah khaled musleh isparta uygulamalı bilimler üniversitesi - teknoloji fakültesi - mekatronik mühendisliği bölümü, Isparta, Türkiye
Title Of Article
Estimation of future climate temperature in Turkey using the ARIMA model
شماره ركورد
45221
Abstract
In recent years, global climate temperature change has been one of the most important problems of the world. Almost all countries are trying to take many prevent regarding global climate temperature change. In this study, monthly temperature data between February 1777 and September 2013 years are taken from the open-access website (Kaggle). The ARIMA model data set has been examined for its stability. Parameters of the ARIMA model are optimized by using grid search and Akaike information criterion. The training was carried out on the ARIMA model using the optimized parameters. The obtained ARIMA model has two prediction methods, static and dynamic. According to Coefficent of Determination (R2) performance evaluation criterion of the ARIMA model, 96,68% accuracy in the dynamic evaluation, and 97,19% accuracy in the static evaluation were obtained.
From Page
69
NaturalLanguageKeyword
ARIMA , Time series , Climate temperature change , Artificial intelligence
JournalTitle
Sdu International Technologic Science
To Page
76
JournalTitle
Sdu International Technologic Science
Link To Document