Klasifikasi Tingkat Kematangan Pisang Menggunakan Metode DEEP Learning dengan CNN

This study aims to develop a banana ripeness classification system using the Convolutional Neural Network (CNN) method. The dataset includes four ripeness categories: Unripe, Half-Ripe, Ripe, and Overripe. The model was trained with variations of 10, 20, 30, 40, and 50 epochs, using a learning rate of 0.001 and a batch size of 16 to determine the optimal configuration. The experimental results show that the best performance was achieved at 30 epochs, producing a validation accuracy of 0.99 and a validation loss of 0.02, indicating high model stability and minimal error. Performance evaluation was conducted using a confusion matrix and supported by Precision, Recall, and F1-Score metrics. Two classes achieved perfect scores of 1.00, while the other two classes recorded F1-Scores above 0.95, reflecting highly accurate predictions. Misclassifications were minimal and occurred only between visually similar categories, specifically Ripe and Overripe. Overall, the CNN model demonstrates strong generalization ability and is highly suitable for automated banana ripeness classification applications