CD Skripsi
Perbandingan Algoritma K-Nn Dan C4.5 Pada Analisis Penyebab Kematian Ibu Di Indonesia
Death is something that cannot be avoided by every living thing, it can occur regardless of time and place so that no one can change these natural provisions. The population mortality rate or in this case the maternal mortality rate has a negative and statistically significant effect on economic growth. This study aims to build a classification model of the causes of maternal mortality using the KNN and C4.5 algorithms. This model uses a dataset from the 2018-2023 “Health Profile” book managed by the health department as much as 779 data, including 623 used as training data and 156 testing data. The data attributes used are bleeding, hypertension, infection, circulatory system disorders, metabolic disorders, and the number of deaths. Through calculations using confusion matrix, the KNN algorithm has an accuracy of 96% and C4.5 has an accuracy of 94.23%. In addition to using confusion matrix, the model also uses a 10-fold cross validation calculation for the generalization process, the KNN model has an accuracy of 96.78% while the C4.5 algorithm has an accuracy of 95.12%. Based on the calculation of confusion marix and 10-fold cross validation, the KNN algorithm has better performance results compared to C4.5.
Keywords: C.45, cross validation, confusion matrix, Death, KNN
Tidak tersedia versi lain