Comparison of Machine Learning Techniques for Grade Classification of Exported Cavendish Bananas

The development of artificial intelligence technology has multiplied and penetrated all fields. One example is using artificial intelligence in a classification technique to determine the grading of bananas based on their quality. Banana grading is divided into three output classes: export quality, local quality, and defects. The banana classification input data is obtained by measuring the length and weight of the bananas and their color, which refers to their ripeness. The machine learning algorithm obtains test results by dividing three types of dataset separation ratios, namely 60% training data, 40% testing data (60:40), 70% training data, 30% testing data (70:30), and 80% training data, and 20% testing data (80:20). The Support Vector Machine (SVM) algorithm gets the best accuracy results of 0.84 at 80:20 ratio conditions. Based on the test results, the classification process with the highest accuracy uses the SVM algorithm. Thus, it is recommended for use as a classifier in classifying cavendish bananas for export purposes.