Reduksi Dimensi Data Microarray untuk Peningkatan Kinerja Algoritma Naïve Bayes dalam Pengklasifikasian

Abstract

According to Global Burden of Cancer study released by the World Health Organization (WHO) in 2020, there were 396,914 cancer cases in Indonesia with total deaths reaching 234,511 cases. To reduce these cases, early detection is still the main key. Utilizing microarray technology that can capture gene expression can be used to detect cancer. However, high dimensionality, small sample size, and noise in gene expression data are serious challenges in microarray data analysis, which can cause the curse of dimensionality problem. Therefore, dimensionality reduction is essential and sensitive to achieve satisfactory classification performance. Therefore, the authors proposed PCA method for dimensionality reduction of microarray data and Naïve Bayes was used for data classification. The proposed method shows the improved performance of Naïve Bayes in microarray data classification where the achieved accuracy value is 92.11% and f1-score is 93.88% for ovarian cancer and for prostate cancer, the accuracy is 93.55%, and f1-score reaches 92.86%.