通过添加Universum数据,引入了与分类样本无关的样本,并借此引入了先验域信息,构建了添加Universum数据的最小二乘投影双支持向量机(ULSPTSVM).此外,还将方法扩展到递归学习方法,用于进一步提高ULSPTSVM的分类性能.实验表明,ULSPTSVM方法可以直接减少带有Universum数据的双支持向量机(USVM)方法的训练时间,而且在多数情况下ULSPTSVM方法的测试精度优于最小二乘投影双支持向量机(LSPTSVM)方法的测试精度.
A new algorithm is constructed,called least squares projection twin support vector machine with Universum(ULSPTSVM).By adding Universum data,samples are introduced which have no relation with the samples of classification,which have a priori domain information.In addition,in order to further enhance the performance of ULSPTSVM,the method is extended to recursive learning method.Experiments show that ULSPTSVM can directly improve the training time of twin support vector machine with Universum(UTSVM),and in most cases the experimental accuracy is better than least squares projection twin support vector machine(LSPTSVM).
[1] WESTON J,COLLOBERT R,SINZ F,et al.Inference with the Universum[Z].The 23rd International Conference on Machine Learning,Pittsburgh,2006.
[2] CHERKASSKY V,DHAR S,DAI W.Practical conditions for effectiveness of the universum learning[J].IEEE Transactions on Neural Networks,2011,22(8):1241-1255.DOI:10.1109/TNN.2011.2157522.
[3] SINZ F H,CHAPELLE O,AGARWAL A,et al.An analysis of inference with the universum[Z].The 21st Annual Lonference Neural Information Processing Systems,Vancouver,2008.
[4] CHEN S,ZHANG C.Selecting informative universum sample for semisupervised learning[Z].The International Joint Conference on Artificial Intelligent,Pasadna,2009.
[5] SHEN C,WANG P,SHEN F,et al.Uboost:Boosting with the universum[J].IEEE Transaction on Pattern Analysis and Machine Intelligence,2012,34(4):825-832.DOI:10.1109/TPAMI.2011.240.
[6] JAYADEVA R,KHEMCHANDANI S,CHANDRA.Twin support vector machines for pattern classification[J].IEEE Transactions on Pattern Analysis and Machine Intellegence,2007,29(5):905-910.DOI:10.1109/TPAMI.2007.1068.
[7] KUMAR M A,GOPAL M.Application of smoothing technique on twin support vector machines[J].Pattern Recognition Letters,2008,29(13):1842-1848.DOI:10.1016/j.patrec.2008.05.016.
[8] SHAO Y H,ZHANG C H,WANG X B,et al.Improvements on twin support vector machines[J].IEEE Transactions on Neural Networks,2011,22(6):962-968.DOI:10.1109/TNN.2011.2130540.
[9] QI Z Q,TIAN Y J,SHI Y.Twin support vector machine with Universum data[J].Neural Networks,2012,36:112-119.DOI:10.1016/j.neunet.2012.09.004.
[10] SHAO Y H,DENG N Y,YANG Z M.Least square recursive projection twin support vector machine for classification[J].Pattern Recognition,2012,45:229-2307.DOI:10.1016/j.patcog.2011.11.028.
[11] DING S F,HUA X P.Recursive least square projection twin support vector machine for nonlinear classification[J].Neurocomputing,2014,134:3-9.DOI:10.1016/j.neucom.2013.02.046.
[12] MUSICANT D R.NDC:Normally Distributed Clustered Datasets[EB/OL].(1998-01-01)[2015-06-05].1998 <www.cs.wisc.edu/dmi/svm/ndc/>.
