Measuring Stability of Feature Selection in Biomedical Datasets

摘要: An important step in the analysis of high-dimensional biomedical data is feature selection. Typically, a subset selected by selection method evaluated for relevance towards task such as prediction or classification. Another property stability that refers to robustness features perturbations data. In biomarker discovery, example, domain experts prefer parsimonious are relatively robust slight changes We present measure called adjusted computes with respect random This useful comparing methods and superior similar measures do not account demonstrate application this on dataset.