Abstract

Cancer is a common severe disease today, and this type of disease has a high mortality rate. Therefore, a cancer diagnosis is an essential tool because most patients die due to a lack of early diagnosis and treatment. To better diagnose cancer, people use gene sequencing techniques and microarray techniques to replace traditional tumor morphology because cancer is a genomic disease. As people widely utilize gene sequencing technology, many gene data aggregately form a cancer gene database to help people manage these data better. Through large-scale cancer genomics datasets, people have used different Machine Learning algorithms to create cancer prediction models. The prediction model created by these algorithms has higher accuracy and a lower error rate than other cancer classification methods. Therefore, a large amount of genetic data can be obtained using different gene detection techniques to detect cancer gene expression. People will perform unique analyses and processing of these data to obtain valuable information. These large amounts of data are aggregated into specialized cancer gene databases such as TCGA that can be used to train ML algorithms to obtain the best predictive models.