Classification Of Pistachio Varieties Using Machine Learning Algorithms

Abstrak

The accurate classification of pistachio varieties plays a crucial role in ensuring quality control, enhancing traceability, and improving market segmentation in the agricultural sector. This study explores the application of various machine learning algorithms—including Decision Tree, Random Forest, XGBoost, Support Vector Classifier (SVC), k-Nearest Neighbors (KNN), and Logistic Regression—for the classification of pistachio types based on morphological features. A publicly available dataset containing measurements such as kernel length, shell width, and aspect ratio was used to train and evaluate the models. The results demonstrated that ensemble methods like XGBoost and Random Forest consistently outperformed other algorithms, achieving accuracy scores of 0.86 and 0.85, respectively, with high Area Under the Curve (AUC) values in the Receiver Operating Characteristic (ROC) analysis. Furthermore, hyperparameter tuning improved model performance across the board. These findings indicate the potential of machine learning as a reliable tool for automating pistachio variety classification and supporting decision-making in agricultural practices. Future research may involve real-time classification using image-based features and integration into precision agriculture systems.