Author : G. Madhavi 1
Date of Publication :1st April 2018
Abstract: Classification issues in high dimensional knowledge with a little range of observations have become additional common especially in microarray knowledge. Throughout the last twenty years, voluminous economical classification models and have choice (FS) algorithms are planned for higher prediction accuracies. However, the results of associate degree FS rule supported the prediction accuracy are unstable over the variations within the coaching set, particularly in high dimensional knowledge. This paper proposes a brand new analysis live Q-statistic that comes with the steadiness of the chosen feature set additionally to the prediction accuracy. Then, we have a tendency to propose the Booster of associate degree FS rule that reinforces the worth of the Q-statistic of the rule applied. Empirical studies supported artificial knowledge and fourteen microarray knowledge sets show that Booster boosts not solely the worth of the Q-statistic however additionally the prediction accuracy of the rule applied unless the information set is in and of itself tough to predict with the given rule.
Reference :
-
- T. Abeel, T. Helleputte, Y. V. de Peer, P. Dupont, and Y. Saeys, “Robust biomarker identification for cancer diagnosis with ensemble feature selection methods,” Bioinformatics, vol. 26, no. 3, pp. 392–398, 2010.
- D. Aha and D. Kibler, “Instance-based learning algorithms,” Mach. Learn., vol. 6, no. 1, pp. 37– 66, 1991.
- S. Alelyan, “On feature selection stability: A data perspective,” PhD dissertation, Arizona State Univ., Tempe, AZ, USA, 2013.