摘要 针对当前监督学习算法在流形数据集上分类性能的缺陷，如分类精度低且稀疏性有限，本文在稀疏贝叶斯方法和流行正则化框架的基础上，提出一种稀疏流形学习算法（Manifold Learning Based on Sparse Bayesian Approach，MLSBA）.该算法是对稀疏贝叶斯模型的扩展，通过在模型的权值上定义稀疏流形先验，有效利用了样本数据的流形信息，提高了算法的分类准确率.在多种数据集上进行实验，结果表明：MLSBA不仅在流形数据集上取得良好的分类性能，而且在非流形数据集上效果也比较好；同时算法在两类数据集上均具有良好的稀疏性能.
Abstract：Aiming at the classification performance deficiencies of current supervised learning algorithms on manifold data sets,e.g.low classification accuracy and limited sparsity,a sparse manifold learning algorithm based on sparse Bayesian inference and manifold regularization framework is proposed.The algorithm is called manifold learning based on sparse Bayesian approach (MLSBA).MLSBA is an extension of sparse Bayesian model,by introducing sparse manifold priors to the weights,which can effectively employ the manifold information of sample data to improve the classification accuracy.Extensive experiments are conducted on various datasets,and the results show that MLSBA not only achieves better classification performance on manifold datasets,but also has comparable effectiveness on the non-manifold datasets,and our algorithm has good sparsity on two categories of datasets at the same time.
[1] CHEN Huanhuan,TINO Peter,YAO Xin.Probabilistic classification vector machines[J].IEEE Transactions on Neural Networks,2009,20(6):901-914.
[2] BISHOP C M.Pattern Recognition and Machine Learning[M].New York,SA:Springer Science & Business Media,2013.
[3] TIPPING Michael E.Sparse Bayesian learning and the relevance vector machine[J].Journal of Machine Learning Research,2001,1:211-244.
[4] BERGER J O.Statistical Decision Theory and Bayesian Analysis[M].New York,USA:Springer Science & Business Media,2013.
[5] LI Chang,CHEN Huanhuan.Sparse Bayesian approach for feature selection[A].IEEE Symposium on Computational Intelligence in Big Data[C].USA:IEEE Press,2014.1-7.
[6] PLATT J.Fast training of support vector machines using sequential minimal optimization[A].Advances in Kernel Methods Support Vector Learning[C].USA:MIT Press,1999.185-208.
[7] CHEN Lin,TSANG I W,XU D.Laplacian embedded regression for scalable manifold regularization[J].IEEE Transactions on Neural Networks and Learning Systems,2012,23(6):902-915.
[8] 杨国鹏,周欣,余旭初,陈伟.基于相关向量机的高光谱影混合像元分解[J].电子学报,2010,38(12):2751-2756. YANG Guopeng,ZHOU Xin,YU Xuchu,CHEN Wei.Relevance vector machine for hyperspectral imagery unmixing[J].Acta Electronica Sinica,2010,38(12):2751-2756.(in Chinese)
[9] MIKHAIL Belkin,PARTHA Niyogi,VIKAS Sindhwani.Manifold regularization:A geometric framework for learning from labeled and unlabeled examples[J].Journal of Machine Learning Research,2006,7:2399-2434.
[10] TZIKAS D G,LIKAS C L,GALATSANOS N P.Sparse Bayesian modeling with adaptive kernel learning[J].IEEE Transactions on Neural Networks,2009,20(6):926-937.
[11] 成萍,司锡才,姜义成,徐荣庆.基于稀疏贝叶斯学习的稀疏信号表示ISAR成像方法[J].电子学报,2008,36(3):547-550. CHENG Ping,SI Xicai,JIANG Yicheng,XU Rongqing.Sparse signal representation ISAR imaging method based on sparse Bayesian learning[J].Acta Electronica Sinica,2008,36(3):547-550.(in Chinese)
[12] TIPPING Michael E,FAUL A.Analysis of sparse Bayesian learning[A].Proceedings of the Conference on Neural Information Processing Systems[C].Vancouver,British Columbia,Canada,2002.383-389.
[13] 王天云,陆新飞,丁丽,尹治平,陈卫东.基于贝叶斯压缩感知的FD-MIMO雷达Off-Grid目标稀疏成像[J].电子学报,2016,44(6):1314-1321. WANG Tianyun,LU Xinfei,DING Li,YIN Zhiping,CHEN Weidong.Bayesian compressive sensing-based sparse imaging for off-grid target in frequency diverse MIMO radar[J].Acta Electronica Sinica,2016,44(6):1314-1321.(in Chinese)
[14] ZHU Xiaojin,LAFFERTY John,ROSENFELD Ronald.Semi-Supervised Learning with Graphs[M].Diss Carnegie Mellon University,Language Technologies Institute,School of Computer Science,2005.
[15] HUANG Guangbin,ZHU Q Y,SIEW C K.Extreme learning machine:theory and applications[J].Neurocomputing,2006,70(1):489-501.
[16] CHEN Huanhuan,TI O Peter,YAO Xin.Efficient probabilistic classification vector machine with incremental basis function selection[J].IEEE Transactions on Neural Networks and Learning Systems,2014,25(2):356-369.
2018, Vol. 46 
