1. Sunflower Crop Yield Prediction Using Machine Learning Methods.
- Author
-
Gökler, Seda Hatice
- Abstract
Sunflower, one of the most important crops, is produced in many countries to meet especially for edible oil demand. Since the sunflower plant is affected by many factors, such as the amount of rain and air temperature, the yield changes from year to year, which has adverse effects on the balance between demand and supply. Because of the product produced in many countries is not enough; it has to be imported. Turkey is one of the world's leading sunflower importers. The yield must be accurately estimated for the imported quantity to be correct. Importing in large quantities causes inventories, while small quantities cause the sunflower oil demand to not be met. It is used methods such as the direct method, simulation, and remote sensing to estimate sunflower yield. However, these methods have some shortcomings. In this article, machine learning methods, such as decision tree (DT), support vector machine (SVM) and random forest (RF), are used for production prediction. In order to increase the effectiveness of the methods, the values of the hyperparameters are determined by Halving Grid Search (HGS) method that is tuning method. The methods were implemented in Edirne, which is among the province with the highest sunflower yield in Turkey. The results were evaluated with ANOVA method and performance evaluation criteria, MAE, MAPE, RMSE, and R2. The R2 values obtained for the test data were determined as 0.92, 0.68 and 0.80 for the DT, SVM and RF methods, respectively. In addition, the number of combinations and execution times were compared using the grid search method and the HGS method for the DT method that gave the best results. While 644204 combinations were solved in 4608 seconds with grid search, 5324 combinations were solved in 23 seconds with HGS. Thus, DT method, providing the prediction with the lowest error, is determined a suitable method for sunflower yield prediction and then accurate buying decision making. [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF