CLIMATE ANALYSIS AND WEATHER FORECASTING WITH DATA MINING: THE CASE OF ADANA PROVINCE IN TURKEY

Abstract

In recent years, with the effect of the climate analysis and weather forecasting are accepted as one of the most important natural topics. Adana, located in the Mediterranean region, has a Mediterranean climate and is one of Turkey's roughest cities. Because of its Mediterranean climate, different weather conditions are observed in every season of the year in Adana. In this study, it is aimed to make a monthly average weather forecast for the next 12 months in Adana. The dataset used in this study was collected from two different weather websites since January 2019 and includes Adana's daily weather values including maximum temperature, minimum temperature, humidity, and wind speed. Weather forecasting was performed on this dataset using linear regression method with Weka data mining tool. While predicting the weather using the linear regression method, it is provided that the relationship between the variables was calculated, and an equation between them was established to make predictions. According to the results obtained, it has been observed that good results cannot be obtained when day-based forecasting is requested, but almost the same results are obtained when the monthly average are requested. In addition, the results of the estimations were approximate, and as a result of the study, the Correlation Coefficient was found to be 0.9832.Additionally, it was concluded that while forecasting the weather, linear regression, which is a data mining technique, yields positive results when the forecast is made with the monthly average.