摘要

在许多实际应用中出现了大量的冗余数据,这些数据可能是高维的,这时进行回归预测将会出现过拟合的现象,并且还会出现预测精度偏低等问题。另外,大多数回归方法都是基于向量的,忽略了矩阵数据原始位置之间的关系。为此,文中提出了一种基于样本特征核矩阵的稀疏双线性回归(Kernel Matrix-based Sparse Bilinear Regression, KMSBR)方法。该方法直接将数据矩阵作为输入,其是通过左右回归系数矩阵而建立的,利用样本的特征核矩阵和L2,1范数,能够同时实现对样本及样本特征的选择,且考虑了数据的原始位置,提高了算法的性能。在若干数据集上的实验结果表明,KMSBR能有效地选择相对重要的样本和特征,从而提高算法的运行效率,且其预测精度优于已有的几种回归模型。

访问全文
The relationships of school-based sexuality education, sexual knowledge and sexual behaviors-a study of 18,000 Chinese college students
Li, Chunyan*; Cheng, Zixi; Wu, Taiwen; Liang, Xiao; Gaoshan, Junjian; Li, Lihe; Hong, Ping; Tang, Kun*
Reproductive Health, 2017, 14(1): 103.

31

被引

访问全文
Childhood adversities and mental disorders in first-year college students: results from the World Mental Health International College Student Initiative
Husky, Mathilde M.; Sadikova, Ekaterina; Lee, Sue; Alonso, Jordi; Auerbach, Randy P.; Bantjes, Jason; Bruffaerts, Ronny; Cuijpers, Pim; Ebert, David D.; Garcia, Raul Gutierrez; Hasking, Penelope; Mak, Arthur; McLafferty, Margaret; Sampson, Nancy A.; Stein, Dan J.; Kessler, Ronald C.*
Psychological Medicine, 2023, 53(7): 2963-2973.
访问全文
The revised international technical guidance on sexuality education - a powerful tool at an important crossroads for sexuality education
Herat, Joanna; Plesons, Marina*; Castle, Chris; Babb, Jenelle; Chandra-Mouli, Venkatraman
Reproductive Health, 2018, 15(1): 185.

3

被引

访问全文
Rethinking Gender and Sexuality in Education
Amir Biglarbeigi Ghajarieh; Mohammad Amin Mozaheb
International Journal of Higher Education, 2012.
访问全文
Sexuality Education: A Critical Need
Shindel Alan W*
The Journal of Sexual Medicine, 2015, 12(7): 1519-1521.

3

被引

6

被引

访问全文
Sexuality Education in Brazilian Medical Schools
Rufino Andrea Cronemberger*; Madeiro Alberto; Batista Castello Girao Manoel Joao
The Journal of Sexual Medicine, 2014, 11(5): 1110-1117.

20

被引

推荐论文
更多