Hybrid CNN & Random Forest Model for Effective Ginger Leaf Disease Diagnosis

This study uses a method based on machine learning to conduct a thorough analysis into the automatic categorization of illnesses affecting ginger leaves. The study focuses on eight distinct classes of ginger leaf diseases: Leaf Spot (Phyllosticta zingiberi), Yellow Leaf Roll (Ginger Yellows), Leaves Blight (Colletotrichum gloeosporioides), Bacterial Wilt (Ralstonia solanacearum), Rhizome Rot (Pythium spp.), Fusarium Wilt (Fusarium oxysporum), The powdery Mildew (Oidium spp.), and Anthracnose (Colletotrichum spp.). The system of classification is evaluated using important performance indicators such as precision, recall, or F1-Score, which provide insight into the model's capacity to accurately identify every illness class. The total amount of instances for each class is indicated by the support values, ensuring a diverse as well as representative dataset for testing and training purposes. Precision ratings that range from 91.39% ability 94.44% demonstrate the model's ability to reliably categorise each condition. Furthermore, recall rates greater than 90% highlight the model's responsiveness in capturing cases of each disease type. The resulting F1-Score, which balances precision and recall, regularly averages 92.50%. Weighted average metrics, which account for class imbalances, show a total precision of 92.51%, validating the framework's robustness across several disease categories. The micro average, which treats all instances equally, is consistent with the general precision, confirming the model's dependability. According to the study, the suggested model for machine learning has a high level of reliability and generalizability. This makes it a useful tool for automated identification and categorization of ginger leaf diseases. This study advances agricultural and botanic practices by providing a realistic approach for early identification of diseases in ginger plants.