Approaches for Cancer Diagnosis Based on Microarray Data

Development on microarray technology may lead to opportunities in bioinformatics and makes it possible to diagnose cancer on the level of gene expression. Many adverse factors, such as small number of samples with high-dimensional characteristics and data class imbalances, pose challenges to traditional machine learning methods. Numerous researchers had worked on these problems and obtained significant achievements. This paper describes the data sets used in study, summarizes the approaches for cancer diagnosis based on microarray data, and provides outlook on future research direction.