Cancer Detection Based on Microarray Data Classification Using Deep Belief Network and Mutual Information

Cancer is one of the deadliest causes of death in the world. Based on the World Health Organization (WHO) cancer is the second leading cause of death in the world with a total death of 8.8 million in 2015. Therefore, improving cancer diagnosis to prevent cancer is necessary. The technology used by researchers to improve the diagnosis of cancer is microarray data. Microarray is a research data that studies many gene expression at one time. Microarray data is used to examine various diseases, especially cancer. Each microarray sample contains DNA or RNA from an individual. The microarray data is then represented as a number table form. The microarray data that has been in the form of numerical tables can then be classified according to their respective class. The problem gained during the classification process is the amount of microarray data features are very large. In this study, used Deep Belief Network for data classification and Mutual Information as a feature selection method. The built system can classify the microarray dataset with an average of 90.84% accuracy and an average f1-score of 89.68%.