TY - GEN
T1 - Stable L2-regularized ensemble feature weighting
AU - Li, Yun
AU - Huang, Shasha
AU - Chen, Songcan
AU - Si, Jennie
PY - 2013
Y1 - 2013
N2 - When selecting features for knowledge discovery applications, stability is a highly desired property. By stability of feature selection, here it means that the feature selection outcomes vary only insignificantly if the respective data change slightly. Several stable feature selection methods have been proposed, but only with empirical evaluation of the stability. In this paper, we aim at providing a try to give an analysis for the stability of our ensemble feature weighting algorithm. As an example, a feature weighting method based on L2-regularized logistic loss and its ensembles using linear aggregation is introduced. Moreover, the detailed analysis for uniform stability and rotation invariance of the ensemble feature weighting method is presented. Additionally, some experiments were conducted using real-world microarray data sets. Results show that the proposed ensemble feature weighting methods preserved stability property while performing satisfactory classification. In most cases, at least one of them actually provided better or similar tradeoff between stability and classification when compared with other methods designed for boosting the stability.
AB - When selecting features for knowledge discovery applications, stability is a highly desired property. By stability of feature selection, here it means that the feature selection outcomes vary only insignificantly if the respective data change slightly. Several stable feature selection methods have been proposed, but only with empirical evaluation of the stability. In this paper, we aim at providing a try to give an analysis for the stability of our ensemble feature weighting algorithm. As an example, a feature weighting method based on L2-regularized logistic loss and its ensembles using linear aggregation is introduced. Moreover, the detailed analysis for uniform stability and rotation invariance of the ensemble feature weighting method is presented. Additionally, some experiments were conducted using real-world microarray data sets. Results show that the proposed ensemble feature weighting methods preserved stability property while performing satisfactory classification. In most cases, at least one of them actually provided better or similar tradeoff between stability and classification when compared with other methods designed for boosting the stability.
UR - http://www.scopus.com/inward/record.url?scp=84892907071&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84892907071&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-38067-9_15
DO - 10.1007/978-3-642-38067-9_15
M3 - Conference contribution
AN - SCOPUS:84892907071
SN - 9783642380662
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 167
EP - 178
BT - Multiple Classifier Systems - 11th International Workshop, MCS 2013, Proceedings
T2 - 11th International Workshop on Multiple Classifier Systems, MCS 2013
Y2 - 15 May 2013 through 17 May 2013
ER -