Optimasi Klasifikasi Bayesian Network Melalui Reduksi Attribute Menggunakan Metode Principal Component Analysis


Dimensionality reduction is a hot topic being discussed in its development has been carried out in various fields of research one of which is machine learning by reducing can reduce the capacity of dimensions without reducing (eliminating) information contained in the data. Principal Component Analysis is one of the proven reduction techniques capable of reducing data capacity without significantly eliminating the information contained in the dataset. In this research attribute reduction using principal component analysis using a dataset of factors affecting employee absence was taken from the University of California repository at Irvine (UCI). Combination with Bayesian Network to classify data as a comparison between before and after attribute reduction. This can be seen in the initial results before the reduction with an accuracy of 100% and after the fifth attribute reduction there is a decrease in accuracy by 89,7%