Forecasting of Giresun Temperature and Pressure Data with Multilayer Artificial Neural Network

In recent years, artificial neural networks have gained significant popularity for addressing time series forecasting challenges. While traditional forecasting techniques are still widely used, they generally perform well on linear datasets. However, many real-world time series exhibit nonlinear and complex behaviours, which limits the effectiveness of classical models. To overcome this limitation, deep neural network architecture such as multilayer perceptions artificial neural network offers a promising alternative due to their capacity for capturing nonlinear patterns and modelling complex relationships through a larger number of parameters. This study focuses on forecasting weather-related variables using a multilayer perceptions artificial neural network, utilizing atmospheric pressure and temperature data collected from Giresun, Turkey. The primary goal is to identify seasonal trends and forecasting future values of these weather indicators. The performance of the proposed MLP model is evaluated and compared with several commonly adopted neural network-based approaches reported in the literature.