Integration of Statistics and Machine Learning for Combining Prediction, Knowledge Discovery and Inference.

点击翻译按钮获取中文摘要

Statistical methods to achieve scientific results conducted from genomics and omics data have been exploited. In particular, for prediction of phenotypes (disease type, treatment effect, prognosis) we enhanced to implement several proposals by a fusion between statistics and machine learning. The boosting method by maximization of Area Under Curve is applied to gene expression data. LASSO clustering is proposed to select genes for detecting subtypes in breast cancer patients.