Sensitivity analysis for feature selection
Sensitivity analysis for feature selection
Date
2019
item.page.datecreated
Authors
Kamalov, Firuz
Journal Title
Journal ISSN
Volume Title
Publisher
Institute of Electrical and Electronics Engineers Inc.
Abstract
Sensitivity analysis allows us to decompose the variance output into its source components. Total sensitivity index represents the effects of varying a feature on the variance of the target variable. In this paper we use total sensitivity index to evaluate features for the purpose of feature selection. We test our method on various data sets and compare its performance relative to other modern feature selection methods. The proposed method produces very robust results with high computational efficiency. © 2018 IEEE.
Description
This conference paper is not available at CUD collection. The version of scholarly record of this conference paper is published in 2018 17th IEEE International Conference on Machine Learning and Applications (ICMLA) (2019), available online at: https://doi.org/10.1109/ICMLA.2018.00238.
item.page.type
Conference Paper
item.page.format
Keywords
Big data , Feature selection , Sensitivity analysis , Total sensitivity index , Computational efficiency , Feature extraction , Machine learning , Sensitivity indices , Sensitivity analysis
Citation
Kamalov, F. (2019). Sensitivity Analysis for Feature Selection. In Proceedings - 17th IEEE International Conference on Machine Learning and Applications, ICMLA 2018 (pp. 1466–1470). https://doi.org/10.1109/ICMLA.2018.00238