Multi-Class Support Vector Machine for Arabica Coffee Bean Roasting Grade Classification

The presence of computer technology based on vision machines has resolved numerous issues in various industries, most notably agriculture-related to coffee plants' development. Before being consumed, coffee beans must pass a roasting process to achieve a desirable flavor. There are three roasting levels: light, medium, and dark. This work proposed a method for classifying Arabica coffee bean roasting grades. Two primary procedures necessitated feature extraction and classification. Following the extraction of statistical data, classification was performed using the support vector machine (SVM) method. Four kernels of linear, polynomial, radial basis, and sigmoid were used in this classifier. Cross-validation was used to evaluate the proposed method's performance with k-fold values of 3, 5, and 10. The evaluation result indicates that the polynomial kernel implementation obtained a maximum accuracy of 100% using k-fold values of 5 and 10.