Effective Diagnosis of Potato Plant Diseases Employing Innovative Deep Learning Methodologies

The cultivation of crops and worldwide availability of food are both greatly impacted by potato plant illnesses. For these illnesses to be effectively treated and prevented, reliable diagnosis is essential. Conventional means of detection can be laborious, liable to mistakes, and need specific knowledge. To improve the precision of illness detection, this investigation applies cutting-edge deep learning techniques which involves logistic regression and neural networks. We created a forecasting system to categorize potato illnesses into nine groups, comprising pests, fungus, pathogens, nematodes, early blight, and late blight, leveraging the “Orange” software architecture. The framework demonstrated exceptional performance in several parameters, such as the $F 1$ score, accuracy, recall, accuracy, and MCC. Neural networks surpassed logistic regression with an average proficiency of over 82% and 97% preciseness, while logistic regression still managed over $\mathbf{9 5 \%}$ accuracy in classification. The structure performed exceptionally well on several criteria, demonstrating the superiority of neural systems in the identification of intricate illness trends.