Machine Learning-Based Classification of Infected Date Palm Leaves Caused by Dubas Insects: A Comparative Analysis of Feature Extraction Methods and Classification Algorithms

This study investigates the utilization of machine learning techniques for effectively classifying infected date palm leaves caused by Dubas insects. Three distinct feature extraction methods, namely Inceptionv3, SqueezeNet, and VGG16, are combined with five diverse machine learning algorithms: K-Nearest Neighbors (KNN), Neural Network (ANN), Random Forest (RF), Artificial Support Vector Machine (SVM), and Logistic Regression (LR). The dataset comprises a collection of images depicting infected date palm leaves, and performance evaluation metrics, including accuracy, recall, precision, and F1 score, are computed for each algorithm. The results unveil varied levels of accuracy among the feature extraction methods and machine learning algorithms. Specifically, Inceptionv3 achieved an accuracy of 80.4% for KNN, while SqueezeNet attained 75.3% and VGG16 obtained 76.6% accuracy. For SVM, the accuracy scores for Inceptionv3, SqueezeNet, and VGG16 were 72.9%, 66%, and 62.4%, respectively. ANN demonstrated promising accuracy scores of 83.8%, 80%, and 80.1% for Inceptionv3, SqueezeNet, and VGG16, respectively. Lastly, LR yielded accuracy scores of 83%, 76.2%, and 80% for Inceptionv3, SqueezeNet, and VGG16, respectively. These findings offer useful information about how various machine learning techniques perform and feature extraction methods in classifying infected date palm leaves, thereby facilitating the development of effective pest management strategies in date palm plantations.