Predictive Modeling of Net Blotch Disease in Spring Barley Using Artificial Neural Networks (ANN) and XGBoost: A Comparative Study

Authors : Recep Sinan Arslan, Nilüfer Akci, Dilara Çelik, Bahatdin Daşbaşı, Kadir Aytaç Özaydın, Teslima Daşbaşı

Pages : 484-503

View : 89 | Download : 45

Publication Date : 2025-08-30

Article Type : Research Paper

Abstract :The detection of plant diseases and their application to agricultural production have the potential to increase productivity and regulate pesticide use. Time-dependent meteorological climate data play an active role in plant disease control practices.The aim of this study was to develop a prediction model for the presence of net blotch disease in barley plants in Turkey. The machine learning and artificial neural network-based model was developed and evaluated using a three-year disease development period (2021-2023) as the training and testing dataset. The model was trained using nine different meteorological climate data, including temperature(min, max, avg), humidity, rainfall, wind speed, sun exposure time, actual vapor pressure, dew point temperature. The outcomes of the evaluation process yielded an 88.9% accuracy for the XGBoost classifier and a noteworthy 92.16% accuracy value for the Artificial Neural Network (ANN).The precision, recall and f-score values for the ANN model were found to be 96.10%, 89.16% and 92.50%, respectively. The findings indicated that the model exhibited superior performance in detecting the absence of the disease when compared to existing studies. Furthermore, the impact of nine distinct inputs on the manifestation of the disease and the interrelationships between these inputs were analysed. The analysis revealed that sun exposure time, temperature, wind speed and humidity factors exerted the most substantial influence on environmental variables. The high performance obtained from the model lends further credence to the hypothesis that artificial intelligence models can be successful in combating the disease and that an increase in productivity can be achieved in the product by reducing the negative effects of the disease.
Keywords : Arpa, Net Leke, Makine Öğrenmesi, Yapay Sinir ağları, Gelişmiş Özellik Analizi

ORIGINAL ARTICLE URL