Peningkatan Akurasi pada Metode Klasifikasi Decision Tree Menggunakan Adaptive Boosting (Adaboost)

Abstract

The Decision Tree Algorithm is an algorithm that has its main advantages compared to other algorithms, the decision tree algorithm is a classification algorithm that is commonly used. The Decision Tree C5.0 algorithm has several drawbacks, including: the C5.0 algorithm and other decision tree methods are often biased towards splitting whose features have many levels, some problems for the model may occur such as over-fit or under-fit challenges, changes Big on decision logic can result in small changes to training data, and Because C5.0 algorithm relies on parallel separation of axes, C5.0 can experience modeling inconveniences. Data imbalance causes low accuracy rate in C5.0 algorithm. The boosting algorithm is an ensemble meta-algorithm method to primarily reduce bias, and therefore variance. In each iteration, assign different weights to the distribution of the training data. Each iteration of the upgrade process changes the distribution of the training data by increasing the weight assigned to examples of incorrect classification and decreasing the weight assigned to examples of correct classification. The purpose of this research is to improve the performance of the Decision Tree C5.0 classification method using adaptive boosting (Adaboost) to predict hepatitis disease using the Confusion matrix. Tests that have been carried out with the confusion matrix use the Hepatitis dataset in the Decision Tree C5.0 classification which has an accuracy rate of 77.41% with a classification error rate of 22.58%. Whereas in the classification of Decision Tree C5.0 Adaboost has a higher accuracy rate of 83.87%, when compared to Decision Tree C5.0. The Adaboost Decision Tree C5.0 classification has a misclassification rate of 16.12%. The Heart Disease dataset has an accuracy rate of 75.92% using the C5.0 algorithm, and 29.54% error classification and increases after using Adaboost 77.77% accuracy, 24.07% error classification. Whereas in the Lung Cancer dataset an accuracy of 82.25% using the C5.0 algorithm has a classification error of 17.74%, accuracy increases 85.48% and the error classification rate decreases 14.51% after using C5.0 Adaboost. This difference is caused by the Adaboost algorithm, because the Adaboost algorithm is able to change a weak classifier into a strong classifier by increasing the weight of the observations, and Adaboost is also able to reduce the classifier error rate.