Explainable AI and Slime Mould Algorithm for Classification of Pistachio Species

Abstract

AbstractThe safety and quality of the food are considered an essential issue in the entire world. This is due to food being the basis of human health. Nowadays, machine learning algorithms have embodied the recent technology in all stages of food processing such as food grading, food quality determination, and food classification. Pistachio nuts have an important role in the agricultural economy. To increase the efficiency of post-harvest industrial processes, there is a need to introduce technologies for classifying different species of pistachio. This study considers an automated model to separate pistachio species. The proposed pistachio species classification consists of three main phases; features selection based on slime mould algorithm phase, feature interpretation based on explainable artificial intelligence phase, and finally classification of pistachio species using logistic regression phase. The proposed pistachio species classification model obtained overall 90% classification accuracy, 90% precision, and 91% f1-score.KeywordsExplainable artificial intelligenceSwarm intelligenceFeature selectionPistachio speciesFood Industry