Penerapan Gini Index Dalam Algoritma Decision Tree Untuk Klasifikasi Edibilitas Jamur

Application of Gini Index in Decision Tree Algorithm for Mushroom Edibility Classification

Abstract

Misidentification of mushrooms can cause health risks because several poisonous mushrooms show physical characteristics similar to edible mushrooms. This study aims to apply the Gini Index in the Decision Tree algorithm to classify mushroom edibility and evaluate model performance on test data. The research used the Mushroom Dataset consisting of 8,124 samples, 22 physical feature attributes, and one target attribute labeled edible or poisonous. The research process included data inspection, one-hot encoding for categorical features, label encoding for the target class, splitting the data into training and testing sets with an 80:20 ratio, training a DecisionTreeClassifier based on the Gini Index, and evaluating the model using a confusion matrix, accuracy, precision, recall, and F1-score. The test result on 1,625 samples produced an accuracy of 99.02 percent. The model correctly classified 809 edible samples and 800 poisonous samples. Precision, recall, and F1-score for both classes were around 0.99. These findings indicate that the Gini-based Decision Tree provides accurate and interpretable classification. The model can serve as an initial decision support basis, while expert validation remains necessary for field implementation