Weather Forecasting over Iraq Using Machine Learning
Israa Jasim Mohammed *1, Bashar Talib Al-Nuaimi 2, Ther Intisar Baker 3
1 College of Science, University of Diyala, Baqubah, Iraq
2 Computer Science Department, University of Diyala, Diyala 32001, Iraq
3 College of Science, University of Diyala, Baqubah, Iraq
Emails: scicompms2211@uodiyala.edu.iq; alnuaimi_bashar@uodiyala.edu.iq; dher@uodiyala.edu.iq
Abstract
The weather generally comprises various factors, such as wind speed, precipitation, and rainfall. Environmental weather forecasting is a demanding task for researchers, and in recent years it has attracted much study attention. Our assessment considers a wide range of weather conditions across Iraq utilizing information gathered from NASA's estimate of the world's energy resources for the years 1981 to 2021. Therefore, the correct forecast of meteorological parameters is a difficult challenge due to their changing environmental conditions. Random forest, decision tree, and GBR algorithms are used for weather forecasting. A comparison among used methods is performed and the RF is achieved the best results with accuracy, MAE, MSE, R2 of 92%, 0.5, 2.45, and 0.92, respectively.
Keywords: Weather forecasting; random forest; decision tree; GBR.