Improving Tuberculosis Diagnosis in HIV-Positive Individuals: Microarray Data with PCA-SVM Method

Human Immunodeficiency Virus (HIV) is a virus that targets the immune system, particularly the white blood cells known as Cluster of Differentiation 4 (CD4) cells. Tuberculosis (TB) stands out as one of the most prevalent opportunistic infections and the primary cause of mortality among individuals with HIV. In 2015, the World Health Organization (WHO) reported 390,000 deaths globally due to HIV/TB co-infection. Timely detection of HIV/TB co-infection enables prompt administration of Anti-Tuberculosis Drugs (OAT) and antiretroviral drugs (ARV) to patients. Presently, TB detection in HIV patients relies on smear culture examination, involving sputum sampling. An alternative approach to TB detection in HIV patients involves microarray technology. Microarray technology allows for the simultaneous analysis of gene expression patterns across thousands of genes in a particular bodily area. Given the large dimensions of microarray data, a machine learning technique is indispensable for facilitating analysis. This study employs dimension reduction and classification processes within the framework of machine learning to develop an accurate microarray data classification model. The aim is to evaluate the performance of a TB detection model for HIV patients, employing Principal Component Analysis (PCA) for dimension reduction and Support Vector Machine (SVM) for classification. The optimal model, utilizing the RBF kernel with seven features, yielded accuracy and F1 score values of 0.86 and 0.83, respectively.