Predicting Cancer Types from Gene Expression Data Using Machine Learning

Genomic analysis has emerged as an essential step in biological and medical research which has accelerated with the incorporation of ML techniques allowing so much more scope in the analysis of complicated genetic information. With the vast amount and the high complexity of genomic information, it is easy to see how ML can be of assistance. Over algorithms being able to identify the structure of genetic polymorphisms, categorize them, and model phenotype features, ML provides an avenue to unlock the workings of genes, their related diseases and individual differences. This paper incorporates various supervised learning algorithms, including Logistic Regression, Support Vector Machine (SVM) and Random Forest, for genomic analysis on the Gene Expression Cancer RNA-Seq dataset, which includes approximately 1000 samples. The highest accuracy of 98.22 percentage was attained by logistic regression.