Penerapan Metode Naive Bayes Classifier dan Algoritma AdaBoost untuk Prediksi Penyakit Ginjal Kronik

Irawan, Adhi Indra (2016) Penerapan Metode Naive Bayes Classifier dan Algoritma AdaBoost untuk Prediksi Penyakit Ginjal Kronik. Other thesis, Universitas Sebelas Maret.

PDF - Published Version Download (15Kb)

Abstract

Problems that often occur in the medical dataset are many attributes that have missing values. Naïve Bayes method is known to provide good accuracy compared to other methods in dealing with missing values. However, when the results obtained are still not satisfactory then boosting with AdaBoost is used to improve its performance. This study discusses the application of Naïve Bayes method and AdaBoost algorithm to classify chronic kidney disease (CKD). From the result obtained by calculating the confusion matrix, the Naïve Bayes method achieved the accuracy of 0.95 and F1-score of 0.958 . While the combination of AdaBoost and Naïve Bayes managed to improve the accuracy of 0.98 and F1-score of 0.984. When the missing values are replaced, the accuracy of Naïve Bayes decreased to 0.945 and F1-score of 0.954, while AdaBoost successfully increased the accuracy to 0.9825 and F1- score of 0.986. This shows that the Naïve Bayes method has good ability in dealing with missing values and AdaBoost algorithm were managed to improve the Naïve Bayes performance by increasing the accuracy. Keywords: AdaBoost, CKD, Missing Value, Naïve Bayes

Actions (login required)