<![CDATA[Objectively determining the quality of bananas (Musa spp.) is an important challenge in post-harvest management to ensure product standardization and minimize losses. This study aims to implement and evaluate the performance of the Random Forest Classifier (RFC) algorithm in predicting banana quality (Good or Bad) based on a combination of morphological features and organoleptic characteristics. A secondary dataset consisting of 8,000 tabular data samples was used, covering features such as Size, Weight, Sweetness, Softness, Ripeness, Acidity, and HarvestTime. The data was processed through Z-Score standardization and divided into a training:testing ratio of 80:20. Testing results showed that the RFC model achieved an exceptionally high classification accuracy of 96.62%, with balanced Precision and Recall values (0.97). Feature importance analysis revealed that Sweetness, Weight, and Size were the most dominant features and contributed significantly to quality decisions. This study proves that a data-based Machine Learning approach can provide an efficient, accurate, and non-destructive method of assessing banana quality, making it a prospective solution for automatic sorting systems in the agricultural industry.]]>
Copyrights © 2026
