Online Group Feature Selection
Authors:
Wang Jing,
Zhao Zhong-Qiu,
Hu Xuegang,
Cheung Yiu-ming,
Wang Meng,
Wu Xindong
Abstract:
Online feature selection with dynamic features has become an active research area in recent years. However, in some real-world applications such as image analysis and email spam filtering, features may arrive by groups. Existing online feature selection methods evaluate features individually, while existing group feature selection methods cannot handle online processing. Motivated by this, we form…
▽ More
Online feature selection with dynamic features has become an active research area in recent years. However, in some real-world applications such as image analysis and email spam filtering, features may arrive by groups. Existing online feature selection methods evaluate features individually, while existing group feature selection methods cannot handle online processing. Motivated by this, we formulate the online group feature selection problem, and propose a novel selection approach for this problem. Our proposed approach consists of two stages: online intra-group selection and online inter-group selection. In the intra-group selection, we use spectral analysis to select discriminative features in each group when it arrives. In the inter-group selection, we use Lasso to select a globally optimal subset of features. This 2-stage procedure continues until there are no more features to come or some predefined stopping conditions are met. Extensive experiments conducted on benchmark and real-world data sets demonstrate that our proposed approach outperforms other state-of-the-art online feature selection methods.
△ Less
Submitted 22 October, 2014; v1 submitted 18 April, 2014;
originally announced April 2014.
