Precision Diagnosis of Wheat Powdery Mildew Using CNN and Random Forest

An in-depth investigation of a prediction model's performance that has been specially developed for predicting the intensity of wheat leaf disease across 10 different intensification levels is the focus of the study. The tabulated results demonstrate the model's success in this challenging task by incorporating crucial performance indicators like Precision, Remembering the F1-Score, Support, Support The proportion, and Accuracy. The model exhibits a remarkable level of Precision throughout all intensification levels, demonstrating its capacity to precisely pinpoint occurrences of wheat leaf disease. The model's ability to reduce inaccurate results, which is essential for disease diagnosis, is shown by precision values that vary from 86.22% to 91.91%. Similar to this, recall values, which range from 89.44% to 92.31%, continue to be consistently high. This demonstrates the model's accuracy in detecting genuine positive disease cases across a range of intensification levels, ensuring thorough disease detection. With scores spanning 88.25% to 91.78%, the F1-Score, a perfectly balanced mix of Accuracy and Recall, regularly displays strong performance. This demonstrates the model's capacity to maintain an appropriate trade-off between recall and precision, which adds to its overall effectiveness in determining the intensity of wheat leaf disease. Additionally, “Support” values show the number of occurrences connected to each intensity level, providing information about the dataset's make-up, while “Support Proportion” values make sure everybody's intensity level participates fairly in the evaluation, minimizing any potential performance biases. The model continuously achieves a remarkable ‘Accuracy’ score of 98%, demonstrating its ability to accurately classify illness intensity levels within the whole dataset.