Optimizing Banana Type Identification: An Support Vector Machine Classification-Based Approach for Cavendish, Mas, and Tanduk Varieties

This research focuses on addressing the need for improved efficiency in the agricultural sector, particularly in banana processing in Indonesia, where the demand for bananas is consistently high. To improve the efficiency of banana processing, the research proposes the development of a machine learning based solution for automatic banana type selection. This solution uses image data of three banana types (Cavendish, Mas, and Tanduked) captured by a microscopic camera. The images are subjected to feature extraction, and a Support Vector Machine (SVM) algorithm is used to train the model. The results are implemented in a graphical user interface (GUI). The experimental results show promising results, with an accuracy of 86.67%, a precision of 87.78%, and an error rate of 13.33%, achieved with SVM parameters of C = 1000 and a linear kernel. This automated approach provides a practical and sustainable solution to the labor-intensive manual banana variety selection process, thus increasing the efficiency of the banana processing industry.