A Study on Durian Selection Based on Decision Tree Algorithm

Non-destructive classification of fruits based on the maturity stage is beneficial to the consumer and fruit industry. Improper ripening can lead to low eating quality and economic loss for the producers. Durian is known as the "King of Fruits" and is one of the rarest and sweetest fruits in South Asia. This study used decision tree algorithm in machine learning to analyze the dataset by assigning it to leaf nodes, thus enabling the prediction of selection of uncut durian. The results show the following conclusions: 1) The selection prediction of uncut durian was achieved by the decision tree algorithm model, which provides a rational decision-making reference for consumers. 2) The inference logic of the ID3 decision tree algorithm is intuitive, the interpretability is clear, and it is convenient for the visualization of the model. 3) The ID3 decision tree algorithm can not deal with attributes that have consecutive values, missing values, and it is inconvenient to deal with the consecutive features, and it is inclined to select the features with more values as split attributes, which may lead to overfitting.