DOI: 10.4137/CIN.S0
关键词: Evidence-based practice 、 Predictive modelling 、 Biomarker (medicine) 、 MEDLINE 、 Feature selection 、 Comparative evaluation 、 Data science 、 Informatics 、 Medicine 、 Health informatics
摘要: At the University of Pittsburgh, I teach a graduate-level course ‘The Practical Analysis High-Throughput Genomic and Proteomic Data’. 50% grade is based on paper project re-analysis published data sets. The aim to encourage comparative evaluation different approaches various analytic tasks for – omic biomarker studies. students are empowered by this understand see themselves that normalization, feature selection, disease prediction model (a) exist, (b) differ in their apparent relative performance helping generate lists therapeutic targets or models. We also learn about standards, mostly from perspective formats, which critical algorithm