Feature Selection for Tandem Mass Spectrum Quality Assessment via Sparse Logistical Regression

Machine learning algorithms are widely used for quality assessment of tandem mass spectra based on a number of features. However, it is still unclear which features are most relevant to the quality of tandem mass spectra. In this paper, a sparse logistical regression method is proposed for selecting the most relevant features from those features found in the literature. To investigate the performance of the proposed method, experiments are conducted on two datasets. The results show the sparse logistical regression model can effectively select a small number of highly relevant features for tandem mass spectrum quality assessment.