Performance of the Decision Tree Algorithm in the Classification of Edible and Poisonous Mushrooms with Information Gain Optimization

Arif Rifan Rudiyanto (Department of Informatics Engineering, Universitas Dian Nuswantoro, Indonesia)
Pujiono Pujiono (Department of Informatics Engineering, Universitas Dian Nuswantoro, Indonesia)
M. Arief Soeleman (Department of Informatics Engineering, Universitas Dian Nuswantoro, Indonesia)
Mustagfirin Mustagfirin (Department of Informatics Engineering, Universitas Wahid Hasyim, Indonesia)

Scientific Journal of Informatics
Vol 10, No 4 (2023): November 2023

Purpose: This study proposes a new mushroom classification model using a decision tree algorithm to classify edible and poisonous mushrooms by applying machine learning whose algorithm has better performance in terms of accuracy.  Methods: The information gain technique was applied at the data feature selection stage to increase the accuracy of the suggested decision tree model. This study used the same mushroom dataset as that employed in previous studies. Result: The proposed decision tree model in this study can classify edible and poisonous mushrooms with a good accuracy of 99.61%, outperforming a previous study whose final accuracy was 97.05%. Novelty: The novelty of this s is the use of information gain as a filter technique at the feature selection stage. This study aims to optimize the previous mushroom classification models with improved accuracy.