IMPLEMENTASI ALGORITMA DECISION TREE C4.5 DENGAN IMPROVISASI MEAN DAN MEDIAN PADA DATASET NUMERIK

Abstract

The decision tree is a method of classifying data mining. The decision tree has one type of algorithm model, namely the C4.5 algorithm. The C4.5 decision tree algorithm is easy to understand because it has a tree-like structure in general. The C4.5 algorithm in handling quantitative data is often less efficient and effective. Based on these problems, this study improvised the numerical attribute dataset using the mean and median in the preprocessing of the data. The improvisation is used to obtain a threshold value, thereby minimizing information loss and time complexity when implementing the C4.5 decision tree in predicting training data. Evaluation of the system used in this study using a confusion matrix. The confusion matrix is ​​used as a benchmark in testing the classification method using data testing. In this study, the dataset was partitioned into three scenarios. In scenario 1 with 70% training data and 20% test data, the highest accuracy is 75%. The improvisation of the mean and median on the numerical attributes in the C4.5 algorithm can be used in this scenario.