National Natural Science Foundation of ChinaNational Natural Science Foundation of China (NSFC) [61305075, 11401185]; China Postdoctoral Science FoundationChina Postdoctoral Science Foundation [2012M520624]; Natural Science Foundation of Shandong ProvinceNatural Science Foundation of Shandong Province [ZR2013FQ004, ZR2013DM015, ZR2015AL014]; Specialized Research Fund for the Doctoral Program of Higher Education of ChinaSpecialized Research Fund for the Doctoral Program of Higher Education (SRFDP) [20130133120014]; Fundamental Research Funds for the Central UniversitiesFundamental Research Funds for the Central Universities [14CX05042A, 15CX05053A, 15CX08011A, 15CX02064A]; University-level Undergraduate Training Program for Innovation and Entrepreneurship [20161349]
机构署名:
本校为其他机构
院系归属:
数学与统计学院
摘要:
In terms of L1/2regularization, a novel feature selection method for a neural framework model has been developed in this paper. Due to the non-convex, non-smooth and non-Lipschitz characteristics of L1/2regularizer, it is difficult to directly employ the gradient descent method in training multilayer perceptron neural networks. A smoothing technique has been considered to approximate the original L1/2regularizer. The proposed method is a two-stage updating approach. First, a multilayer network model with smoothing L1/2regularizer is trained to eliminate the unimportant features. Second, the co...