Trend analizi ve yapay sinir ağları: Tarımda bir uygulaması
Trend Analysis and Artificial Neural Networks: An Application in Agriculture

Summary/Abstract: The aim of this study is to show that production planning may be performed using artificial neural networks (ANN) and trend analysis in the establishment of sainfoin production amount model and in forecasting in Turkey by years. The study covers data for the period 1990-2020. In the development of ANN and trend analysis, parameter of years was used as an input parameter and production amount was used as an output parameter. Linear, quadratic and cubic models are used in trend analysis. In the ANN method, the Hyperbolic Tangent Function is used as the activation function. The efficiency of the model developed was determined using statistical parameters such as Mean Squared Error (MSE) and determination coefficient (R2). Comparing trend analysis and ANN, ANN method with lower mean square error (MSE) value gave better results. Prediction has been made according to ANN. The results foresee that sainfoin production will be in a decline in 2025 over the year 2020. The sainfoin production, which was 1 934 697 tons in 2020, is expected to be 1 860 691 tons with a decrease of 3.83% in 2025. ANN is a useful tool in terms of determining the results found in case of any changes that may occur in variables and in terms of improving the processes accordingly. It has been noted that ANN models yield better results than trend analysis in production modelling.

• Details
• Contents