Apple (Malus domestica) Quality Evaluation Based on Analysis of Features Using Machine Learning Techniques

The use of artificial intelligence and machine learning algorithms for assessment of apple quality was evaluated in this study. Apples are renowned for containing a variety of nutritional elements. By analyzing apple characteristics, the study aimed to categorize apple quality, thus promoting apple consumption and production. The dataset used consists of 4000 data and eight features provided by an American agricultural company. There were two quality classes of apples: there were 2004 quality apples and 1996 low-quality apples. Artificial intelligence classification algorithms such as multilayer perceptron (MLP), support vector machine (SVM), random forest (RF), k‑nearest neighbor (k-NN), and decision tree (DT) have were to predict apple quality. The performance of the algorithms was evaluated on their ability to accurately predict the quality level of the apples. According to the results of the study, the MLP algorithm achieved the highest classification success with an accuracy rate of 95.63%. The accuracy values of the other algorithms were SVM with 90.75%, k‑NN with 89.75%, RF with 89.63%, and DT with 81%. Apple quality is not achieved by relying on a single feature, but rather by evaluating all the features affecting the apple together. An ideal level of acidity enriches the flavor and texture of food, whereas excessive acidity leads to a sour taste. Due to this complexity, we classified factors affecting apple quality and examined traits separately.