《计算机应用》唯一官方网站 ›› 2022, Vol. 42 ›› Issue (2): 349-356.DOI: 10.11772/j.issn.1001-9081.2021071230
所属专题: 人工智能
收稿日期:
2021-07-15
修回日期:
2021-08-04
接受日期:
2021-08-10
发布日期:
2022-02-11
出版日期:
2022-02-10
通讯作者:
邢红杰
作者简介:
祁祥洲(1994—),男,河北张家口人,硕士研究生,主要研究方向:机器学习;基金资助:
Received:
2021-07-15
Revised:
2021-08-04
Accepted:
2021-08-10
Online:
2022-02-11
Published:
2022-02-10
Contact:
Hongjie XING
About author:
QI Xiangzhou, born in 1994, M. S. candidate. His research interests include machine learning.Supported by:
摘要:
多核学习(MKL)方法在分类及回归任务中均取得了优于单核学习方法的性能,但传统的MKL方法均用于处理两类或多类分类问题。为了使MKL方法适用于处理单类分类(OCC)问题,提出了基于中心核对齐(CKA)的单类支持向量机(OCSVM)。首先利用CKA计算每个核矩阵的权重,然后将所得权重用作线性组合系数,进而将不同类型的核函数加以线性组合以构造组合核函数,最后将组合核函数引入到传统OCSVM中代替单个核函数。该方法既能避免核函数的选取问题,又能提高泛化性能和抗噪声能力。在20个UCI基准数据集上与其他五种相关方法进行了实验比较,结果表明该方法在13个数据集上的几何均值(g-mean)均高于其他对比方法,而传统的单核OCSVM仅在2个数据集上的效果较好,局部多核单类支持向量机(LMKOCSVM)和基于核目标对齐的多核单类支持向量机(KTA-MKOCSVM)在5个数据集上的分类效果较好。因此,通过实验比较充分验证了所提方法的有效性。
中图分类号:
祁祥洲, 邢红杰. 基于中心核对齐的多核单类支持向量机[J]. 计算机应用, 2022, 42(2): 349-356.
Xiangzhou QI, Hongjie XING. Centered kernel alignment based multiple kernel one-class support vector machine[J]. Journal of Computer Applications, 2022, 42(2): 349-356.
数据集 | Ntar | Nnon-ta | Nfea | Ntr | Nts |
---|---|---|---|---|---|
Banana | 2 376 | 2 924 | 2 | 2 777 | 2 523 |
Blood transfusion | 178 | 570 | 4 | 313 | 435 |
Cancer | 239 | 444 | 9 | 324 | 359 |
Cleverland heart | 214 | 83 | 13 | 195 | 102 |
Flare solar | 94 | 50 | 9 | 90 | 54 |
German | 300 | 700 | 20 | 450 | 550 |
Glass | 70 | 76 | 10 | 78 | 68 |
Hill valley | 301 | 305 | 100 | 331 | 275 |
Housing | 245 | 261 | 13 | 274 | 232 |
Image | 1 188 | 898 | 18 | 1 219 | 867 |
Ionosphere | 225 | 126 | 34 | 217 | 134 |
Liver | 145 | 200 | 6 | 176 | 169 |
Ringnorm | 3 664 | 3 736 | 20 | 4 051 | 3 349 |
Sonar | 111 | 97 | 60 | 117 | 91 |
Splice | 1 344 | 1 647 | 60 | 1 569 | 1 422 |
Titanic | 14 | 10 | 3 | 10 | 14 |
Twonorm | 3 703 | 3 697 | 20 | 4 071 | 3 329 |
Waveform | 1 647 | 3 353 | 21 | 2 322 | 2 678 |
Wdbc | 212 | 357 | 9 | 276 | 293 |
Wholesale customers | 298 | 142 | 7 | 280 | 160 |
表1 实验中的数据集
Tab. 1 Datasets used in experiments
数据集 | Ntar | Nnon-ta | Nfea | Ntr | Nts |
---|---|---|---|---|---|
Banana | 2 376 | 2 924 | 2 | 2 777 | 2 523 |
Blood transfusion | 178 | 570 | 4 | 313 | 435 |
Cancer | 239 | 444 | 9 | 324 | 359 |
Cleverland heart | 214 | 83 | 13 | 195 | 102 |
Flare solar | 94 | 50 | 9 | 90 | 54 |
German | 300 | 700 | 20 | 450 | 550 |
Glass | 70 | 76 | 10 | 78 | 68 |
Hill valley | 301 | 305 | 100 | 331 | 275 |
Housing | 245 | 261 | 13 | 274 | 232 |
Image | 1 188 | 898 | 18 | 1 219 | 867 |
Ionosphere | 225 | 126 | 34 | 217 | 134 |
Liver | 145 | 200 | 6 | 176 | 169 |
Ringnorm | 3 664 | 3 736 | 20 | 4 051 | 3 349 |
Sonar | 111 | 97 | 60 | 117 | 91 |
Splice | 1 344 | 1 647 | 60 | 1 569 | 1 422 |
Titanic | 14 | 10 | 3 | 10 | 14 |
Twonorm | 3 703 | 3 697 | 20 | 4 071 | 3 329 |
Waveform | 1 647 | 3 353 | 21 | 2 322 | 2 678 |
Wdbc | 212 | 357 | 9 | 276 | 293 |
Wholesale customers | 298 | 142 | 7 | 280 | 160 |
数据集 | 数据集 | ||||
---|---|---|---|---|---|
Banana | 10-1 | 0.50 | Ionosphere | 10 | 0.64 |
Blood transfusion | 102 | 0.62 | Liver | 10 | 0.65 |
Cancer | 1 | 0.05 | Ringnorm | 10-3 | 0.01 |
Cleverland heart | 10 | 0.78 | Sonar | 1 | 0.60 |
Flare solar | 10-1 | 0.40 | Splice | 10-1 | 0.67 |
German | 10 | 0.50 | Titanic | 10-2 | 0.40 |
Glass | 10-1 | 0.65 | Twonorm | 1 | 0.18 |
Hill valley | 1 | 0.50 | Waveform | 10-1 | 0.18 |
Housing | 10-1 | 0.18 | Wdbc | 10-2 | 0.15 |
Image | 1 | 0.30 | Wholesale customers | 10-1 | 0.18 |
表2 宽度参数σ和折中参数ν在数据集上的设置情况
Tab. 2 Width parameter σ and compromise parameter ν setting on UCI datasets
数据集 | 数据集 | ||||
---|---|---|---|---|---|
Banana | 10-1 | 0.50 | Ionosphere | 10 | 0.64 |
Blood transfusion | 102 | 0.62 | Liver | 10 | 0.65 |
Cancer | 1 | 0.05 | Ringnorm | 10-3 | 0.01 |
Cleverland heart | 10 | 0.78 | Sonar | 1 | 0.60 |
Flare solar | 10-1 | 0.40 | Splice | 10-1 | 0.67 |
German | 10 | 0.50 | Titanic | 10-2 | 0.40 |
Glass | 10-1 | 0.65 | Twonorm | 1 | 0.18 |
Hill valley | 1 | 0.50 | Waveform | 10-1 | 0.18 |
Housing | 10-1 | 0.18 | Wdbc | 10-2 | 0.15 |
Image | 1 | 0.30 | Wholesale customers | 10-1 | 0.18 |
数据集 | 度量准则 | OCSVM(r)[ | OCSVM(l)[ | OCSVM(p)[ | KTA-MKOCSVM[ | LMKOCSVM[ | CKA-MKOCSVM |
---|---|---|---|---|---|---|---|
Banana | g-mean±标准差 | 0.534 3±0.006 2 | 0.636 3±0.006 1 | 0.533 8±0.005 3 | 0.616 4±0.003 0 | 0.8069±0.0011 | 0.658 8±0.000 5 |
P值 | 2.22E-13 | 9.73E-07 | 4.52E-14 | 1.39E-11 | 2.18E-15 | — | |
Blood transfusion | g-mean±标准差 | 0.536 5±0.030 3 | 0.642 7±0.011 6 | 0.244 5±0.006 1 | 0.503 4±0.006 6 | 0.542 1±0.003 2 | 0.6462±0.0003 |
P值 | 1.13E-06 | 3.16E-02 | 5.58E-18 | 1.51E-13 | 3.06E-15 | — | |
Cancer | g-mean±标准差 | 0.863 2±0.007 1 | 0.695 6±0.006 4 | 0.430 2±0.015 0 | 0.534 2±0.001 9 | 0.628 2±0.002 0 | 0.9399±0.0006 |
P值 | 7.10E-11 | 5.78E-16 | 2.55E-15 | 2.42E-25 | 3.06E-15 | — | |
Cleverland heart | g-mean±标准差 | 0.487 7±0.009 2 | 0.608 1±0.003 6 | 0.256 2±0.014 1 | 0.619 2±0.008 5 | 0.6531±0.0041 | 0.519 9±0.009 9 |
P值 | 6.03E-07 | 1.76E-11 | 5.73E-19 | 7.11E-15 | 3.06E-15 | — | |
Flare solar | g-mean±标准差 | 0.476 6±0.010 7 | 0.771 4±0.004 5 | 0.229 4±0.009 1 | 0.472 5±0.010 1 | 0.7321±0.0042 | 0.532 1±0.011 7 |
P值 | 4.06E-08 | 7.29E-16 | 1.33E-20 | 7.11E-15 | 9.92E-15 | — | |
German | g-mean±标准差 | 0.468 8±0.002 0 | 0.744 4±0.008 7 | 0.515 1±0.001 2 | 0.464 4±0.009 3 | 0.8655±0.0027 | 0.649 2±0.007 5 |
P值 | 3.57E-16 | 1.58E-15 | 2.56E-13 | 7.84E-20 | 1.96E-17 | — | |
Glass | g-mean±标准差 | 0.8549±0.0001 | 0.697 8±0.013 4 | 0.510 9±0.009 5 | 0.511 8±0.007 8 | 0.707 1±0.005 5 | 0.817 3±0.002 4 |
P值 | 3.06E-12 | 1.78E-10 | 1.75E-16 | 4.91E-18 | 1.32E-16 | — | |
Hill valley | g-mean±标准差 | 0.523 1±0.000 2 | 0.533 3±0.007 5 | 0.292 9±0.010 7 | 0.483 4±0.004 3 | 0.533 2±0.001 7 | 0.6046±0.0005 |
P值 | 1.99E-24 | 2.25E-10 | 9.74E-15 | 7.58E-15 | 1.32E-16 | — | |
Housing | g-mean±标准差 | 0.573 6±0.004 2 | 0.537 7±0.005 3 | 0.631 3±0.006 9 | 0.455 4±0.008 1 | 0.652 2±0.004 5 | 0.6756±0.0011 |
P值 | 3.96E-15 | 4.05E-15 | 4.36E-09 | 8.02E-15 | 1.59E-08 | — | |
Image | g-mean±标准差 | 0.674 6±0.004 5 | 0.633 5±0.001 3 | 0.577 4±0.011 1 | 0.422 4±0.012 4 | 0.625 7±0.013 7 | 0.7781±0.0015 |
P值 | 1.18E-15 | 1.94E-32 | 5.06E-10 | 6.47E-15 | 4.03E-11 | — | |
Ionosphere | g-mean±标准差 | 0.514 7±0.006 4 | 0.674 7±0.009 5 | 0.542 7±0.008 1 | 0.7165±0.0007 | 0.562 5±0.009 1 | 0.515 7±0.009 7 |
P值 | 7.07E-01 | 1.78E-18 | 5.06E-10 | 1.71E-13 | 4.03E-11 | — | |
Liver | g-mean±标准差 | 0.511 5±0.003 2 | 0.682 4±0.008 8 | 0.632 5±0.003 7 | 0.869 6±0.000 5 | 0.813 4±0.003 9 | 0.9154±0.0011 |
P值 | 4.78E-24 | 1.19E-14 | 7.51E-21 | 1.13E-21 | 7.70E-16 | — | |
Ringnorm | g-mean±标准差 | 0.560 6±0.003 7 | 0.730 1±0.002 4 | 0.625 5±0.002 3 | 0.741 2±0.002 1 | 0.703 4±0.001 1 | 0.7501±0.0061 |
P值 | 7.54E-17 | 7.81E-10 | 3.28E-17 | 3.10E-07 | 2.25E-16 | — | |
Sonar | g-mean±标准差 | 0.589 1±0.005 6 | 0.736 2±0.006 2 | 0.9018±0.0006 | 0.503 3±0.005 6 | 0.837 5±0.003 0 | 0.775 2±0.000 5 |
P值 | 2.23E-15 | 7.81E-10 | 5.82E-39 | 5.72E-17 | 2.97E-13 | — | |
Splice | g-mean±标准差 | 0.620 1±0.008 6 | 0.663 9±0.008 5 | 0.394 5±0.016 7 | 0.664 8±0.002 6 | 0.501 5±0.010 1 | 0.7103±0.0002 |
P值 | 1.06E-10 | 7.81E-10 | 5.28E-13 | 5.72E-17 | 2.21E-13 | — | |
Titanic | g-mean±标准差 | 0.884 8±0.000 2 | 0.736 8±0.006 2 | 0.505 2±0.008 3 | 0.706 2±0.001 7 | 0.797 6±0.001 6 | 0.9156±0.0001 |
P值 | 2.95E-25 | 6.75E-31 | 8.56E-17 | 1.43E-20 | 2.43E-18 | — | |
Twonorm | g-mean±标准差 | 0.551 2±0.007 3 | 0.703 1±0.001 1 | 0.724 8±0.009 7 | 0.571 9±0.011 9 | 0.673 2±0.008 6 | 0.7393±0.0043 |
P值 | 4.90E-20 | 6.75E-31 | 9.61E-04 | 7.83E-14 | 8.09E-12 | — | |
Waveform | g-mean±标准差 | 0.646 5±0.003 1 | 0.517 3±0.006 5 | 0.338 6±0.014 7 | 0.591 4±0.003 8 | 0.800 5±0.001 2 | 0.8098±0.0017 |
P值 | 3.94E-24 | 3.42E-18 | 2.13E-15 | 2.90E-22 | 2.92E-10 | — | |
Wdbc | g-mean±标准差 | 0.714 9±0.004 9 | 0.666 9±0.009 6 | 0.450 3±0.005 2 | 0.542 1±0.009 9 | 0.851 3±0.000 7 | 0.8857±0.0003 |
P值 | 1.86E-15 | 8.55E-14 | 5.59E-19 | 2.18E-15 | 9.21E-22 | — | |
Wholesalecustomers | g-mean±标准差 | 0.674 6±0.004 5 | 0.633 5±0.001 3 | 0.577 4±0.011 1 | 0.422 4±0.012 4 | 0.625 7±0.013 7 | 0.7781±0.0015 |
P值 | 1.18E-15 | 1.94E-32 | 5.06E-10 | 6.47E-15 | 4.03E-11 | — |
表3 六种不同方法在20个UCI基准数据集上取得的测试结果
Tab. 3 Test results obtained by six different methods on 20 UCI datasets
数据集 | 度量准则 | OCSVM(r)[ | OCSVM(l)[ | OCSVM(p)[ | KTA-MKOCSVM[ | LMKOCSVM[ | CKA-MKOCSVM |
---|---|---|---|---|---|---|---|
Banana | g-mean±标准差 | 0.534 3±0.006 2 | 0.636 3±0.006 1 | 0.533 8±0.005 3 | 0.616 4±0.003 0 | 0.8069±0.0011 | 0.658 8±0.000 5 |
P值 | 2.22E-13 | 9.73E-07 | 4.52E-14 | 1.39E-11 | 2.18E-15 | — | |
Blood transfusion | g-mean±标准差 | 0.536 5±0.030 3 | 0.642 7±0.011 6 | 0.244 5±0.006 1 | 0.503 4±0.006 6 | 0.542 1±0.003 2 | 0.6462±0.0003 |
P值 | 1.13E-06 | 3.16E-02 | 5.58E-18 | 1.51E-13 | 3.06E-15 | — | |
Cancer | g-mean±标准差 | 0.863 2±0.007 1 | 0.695 6±0.006 4 | 0.430 2±0.015 0 | 0.534 2±0.001 9 | 0.628 2±0.002 0 | 0.9399±0.0006 |
P值 | 7.10E-11 | 5.78E-16 | 2.55E-15 | 2.42E-25 | 3.06E-15 | — | |
Cleverland heart | g-mean±标准差 | 0.487 7±0.009 2 | 0.608 1±0.003 6 | 0.256 2±0.014 1 | 0.619 2±0.008 5 | 0.6531±0.0041 | 0.519 9±0.009 9 |
P值 | 6.03E-07 | 1.76E-11 | 5.73E-19 | 7.11E-15 | 3.06E-15 | — | |
Flare solar | g-mean±标准差 | 0.476 6±0.010 7 | 0.771 4±0.004 5 | 0.229 4±0.009 1 | 0.472 5±0.010 1 | 0.7321±0.0042 | 0.532 1±0.011 7 |
P值 | 4.06E-08 | 7.29E-16 | 1.33E-20 | 7.11E-15 | 9.92E-15 | — | |
German | g-mean±标准差 | 0.468 8±0.002 0 | 0.744 4±0.008 7 | 0.515 1±0.001 2 | 0.464 4±0.009 3 | 0.8655±0.0027 | 0.649 2±0.007 5 |
P值 | 3.57E-16 | 1.58E-15 | 2.56E-13 | 7.84E-20 | 1.96E-17 | — | |
Glass | g-mean±标准差 | 0.8549±0.0001 | 0.697 8±0.013 4 | 0.510 9±0.009 5 | 0.511 8±0.007 8 | 0.707 1±0.005 5 | 0.817 3±0.002 4 |
P值 | 3.06E-12 | 1.78E-10 | 1.75E-16 | 4.91E-18 | 1.32E-16 | — | |
Hill valley | g-mean±标准差 | 0.523 1±0.000 2 | 0.533 3±0.007 5 | 0.292 9±0.010 7 | 0.483 4±0.004 3 | 0.533 2±0.001 7 | 0.6046±0.0005 |
P值 | 1.99E-24 | 2.25E-10 | 9.74E-15 | 7.58E-15 | 1.32E-16 | — | |
Housing | g-mean±标准差 | 0.573 6±0.004 2 | 0.537 7±0.005 3 | 0.631 3±0.006 9 | 0.455 4±0.008 1 | 0.652 2±0.004 5 | 0.6756±0.0011 |
P值 | 3.96E-15 | 4.05E-15 | 4.36E-09 | 8.02E-15 | 1.59E-08 | — | |
Image | g-mean±标准差 | 0.674 6±0.004 5 | 0.633 5±0.001 3 | 0.577 4±0.011 1 | 0.422 4±0.012 4 | 0.625 7±0.013 7 | 0.7781±0.0015 |
P值 | 1.18E-15 | 1.94E-32 | 5.06E-10 | 6.47E-15 | 4.03E-11 | — | |
Ionosphere | g-mean±标准差 | 0.514 7±0.006 4 | 0.674 7±0.009 5 | 0.542 7±0.008 1 | 0.7165±0.0007 | 0.562 5±0.009 1 | 0.515 7±0.009 7 |
P值 | 7.07E-01 | 1.78E-18 | 5.06E-10 | 1.71E-13 | 4.03E-11 | — | |
Liver | g-mean±标准差 | 0.511 5±0.003 2 | 0.682 4±0.008 8 | 0.632 5±0.003 7 | 0.869 6±0.000 5 | 0.813 4±0.003 9 | 0.9154±0.0011 |
P值 | 4.78E-24 | 1.19E-14 | 7.51E-21 | 1.13E-21 | 7.70E-16 | — | |
Ringnorm | g-mean±标准差 | 0.560 6±0.003 7 | 0.730 1±0.002 4 | 0.625 5±0.002 3 | 0.741 2±0.002 1 | 0.703 4±0.001 1 | 0.7501±0.0061 |
P值 | 7.54E-17 | 7.81E-10 | 3.28E-17 | 3.10E-07 | 2.25E-16 | — | |
Sonar | g-mean±标准差 | 0.589 1±0.005 6 | 0.736 2±0.006 2 | 0.9018±0.0006 | 0.503 3±0.005 6 | 0.837 5±0.003 0 | 0.775 2±0.000 5 |
P值 | 2.23E-15 | 7.81E-10 | 5.82E-39 | 5.72E-17 | 2.97E-13 | — | |
Splice | g-mean±标准差 | 0.620 1±0.008 6 | 0.663 9±0.008 5 | 0.394 5±0.016 7 | 0.664 8±0.002 6 | 0.501 5±0.010 1 | 0.7103±0.0002 |
P值 | 1.06E-10 | 7.81E-10 | 5.28E-13 | 5.72E-17 | 2.21E-13 | — | |
Titanic | g-mean±标准差 | 0.884 8±0.000 2 | 0.736 8±0.006 2 | 0.505 2±0.008 3 | 0.706 2±0.001 7 | 0.797 6±0.001 6 | 0.9156±0.0001 |
P值 | 2.95E-25 | 6.75E-31 | 8.56E-17 | 1.43E-20 | 2.43E-18 | — | |
Twonorm | g-mean±标准差 | 0.551 2±0.007 3 | 0.703 1±0.001 1 | 0.724 8±0.009 7 | 0.571 9±0.011 9 | 0.673 2±0.008 6 | 0.7393±0.0043 |
P值 | 4.90E-20 | 6.75E-31 | 9.61E-04 | 7.83E-14 | 8.09E-12 | — | |
Waveform | g-mean±标准差 | 0.646 5±0.003 1 | 0.517 3±0.006 5 | 0.338 6±0.014 7 | 0.591 4±0.003 8 | 0.800 5±0.001 2 | 0.8098±0.0017 |
P值 | 3.94E-24 | 3.42E-18 | 2.13E-15 | 2.90E-22 | 2.92E-10 | — | |
Wdbc | g-mean±标准差 | 0.714 9±0.004 9 | 0.666 9±0.009 6 | 0.450 3±0.005 2 | 0.542 1±0.009 9 | 0.851 3±0.000 7 | 0.8857±0.0003 |
P值 | 1.86E-15 | 8.55E-14 | 5.59E-19 | 2.18E-15 | 9.21E-22 | — | |
Wholesalecustomers | g-mean±标准差 | 0.674 6±0.004 5 | 0.633 5±0.001 3 | 0.577 4±0.011 1 | 0.422 4±0.012 4 | 0.625 7±0.013 7 | 0.7781±0.0015 |
P值 | 1.18E-15 | 1.94E-32 | 5.06E-10 | 6.47E-15 | 4.03E-11 | — |
1 | PANTAZI X E, MOSHOU D, TAMOURIDOU A A. Automated leaf disease detection in different crop species through image features analysis and one class classifiers[J]. Computers and Electronics in Agriculture, 2019, 156: 96-104. 10.1016/j.compag.2018.11.005 |
2 | STEIN R A, JAQUES P A, VALIATI J F. An analysis of hierarchical text classification using word embeddings[J]. Information Sciences, 2019, 471: 216-232. 10.1016/j.ins.2018.09.001 |
3 | YIN X, ZHANG Y Z, CHEN X H. A binary-classification method based on dictionary learning and ADMM for network intrusion detection [C]// Proceedings of the 2017 International Conference on Cyber-Enabled Distributed Computing and Knowledge Discovery. Piscataway: IEEE, 2017: 326-333. 10.1109/cyberc.2017.88 |
4 | SCHÖLKOPF B, WILLIAMSON R C, SMOLA A, et al. Support vector method for novelty detection [C]// Proceedings of the 12th International Conference on Neural Information Processing Systems. Cambridge: MIT Press, 1999: 582-588. 10.1162/089976600300015565 |
5 | TAX D M J, DUIN R P W. Support vector data description[J]. Machine Learning, 2004, 54(1): 45-66. 10.1023/b:mach.0000008084.60811.49 |
6 | VAPNIK V N. The Nature of Statistical Learning Theory[M]. New York: Springer, 1995: 131-133. 10.1007/978-1-4757-2440-0_7 |
7 | SCHÖLKOPF B, PLATT J C, SHAWE-TAYLOR J. Estimating the support of a high dimensional distribution[J]. Neural Computation, 2001, 13(7): 1443-1471. 10.1162/089976601750264965 |
8 | 汪洪桥,孙富春,蔡艳宁,等.多核学习方法[J].自动化学报, 2010, 36(8): 1037-1050. 10.3724/SP.J.1004.2010.01037 |
WANG H Q, SUN F C, CAI Y N, et al. On multiple kernel learning methods[J]. Acta Automatica Sinica, 2010, 36(8): 1037-1050. 10.3724/SP.J.1004.2010.01037 | |
9 | BENNETT K P, MOMMA M, EMBRECHTS M J. MARK: a boosting algorithm for heterogeneous kernel models [C]// Proceedings of the 8th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York: ACM, 2002: 24-31. 10.1145/775047.775051 |
10 | LANCKRIET G R G, CRISTIANINI N, BARTLETT P, et al. Learning the kernel matrix with semidefinite programming[J]. Journal of Machine Learning Research, 2004, 5: 27-72. |
11 | BACH F R, LANCKRIET G R G, JORDAN M I. Multiple kernel learning, conic duality, and the SMO algorithm [C]// Proceedings of the 21st International Conference on Machine Learning. New York: ACM, 2004: No.6. 10.1145/1015330.1015424 |
12 | SONNENBURG S, RÄTSCH G, SCHÄFER C. A general and efficient multiple kernel learning algorithm [C]// Proceedings of the 12th International Conference on Neural Information Processing Systems. Cambridge: MIT Press, 2005: 1273-1280. 10.1007/11415770_30 |
13 | RAKOTOMAMONJY A, BACH F R, CANU S, et al. SimpleMKL[J]. Journal of Machine Learning Research, 2008, 9: 2491-2521. |
14 | GÖNEN M, ALPAYDIN E. Localized multiple kernel learning [C]// Proceedings of the 25th International Conference on Machine Learning. New York: ACM, 2008: 352-359. 10.1145/1390156.1390201 |
15 | 丁跃. 一种改进的局部多核学习算法[J].重庆工商大学学报(自然科学版), 2014, 31(11): 56-61. |
DING Y. A kind of improved localized multiple kernel learning algorithm[J]. Journal of Chongqing Technology and Business University (Natural Science Edition), 2014, 31(11): 56-61. | |
16 | 何强,张娇阳.核对齐多核模糊支持向量机[J].智能系统学报, 2019, 14(6): 1163-1169. 10.11992/tis.201904050 |
HE Q, ZHANG J Y. Kernel-target alignment multi-kernel fuzzy support vector machine[J]. CAAI Transactions on Intelligent Systems, 2019, 14(6): 1163-1169. 10.11992/tis.201904050 | |
17 | CRISTIANINI N, SHAWE-TAYLOR J, ELISSEEFF A, et al. On kernel-target alignment [C]// Proceedings of the 14th International Conference on Neural Information Processing Systems: Natural and Synthetic. Cambridge: MIT Press, 2001: 367-373. 10.7551/mitpress/1120.003.0052 |
18 | LU Y T, WANG L T, LU J F, et al. Multiple kernel clustering based on centered kernel alignment[J]. Pattern Recognition, 2014, 47(11): 3656-3664. 10.1016/j.patcog.2014.05.005 |
19 | XUE H, SONG Y, XU H M. Multiple indefinite kernel learning for feature selection[J]. Knowledge-Based Systems, 2020, 191: No.105272. 10.1016/j.knosys.2019.105272 |
20 | WANG Z, HONG S S, YAO L J, et al. Multiple universum empirical kernel learning[J]. Engineering Applications of Artificial Intelligence, 2020, 89: No.103461. 10.1016/j.engappai.2019.103461 |
21 | OIKONOMOU V P, NIKOLOPOULOS S, KOMPATSIARIS I. A Bayesian multiple kernel learning algorithm for SSVEP BCI detection[J]. IEEE Journal of Biomedical and Health Informatics, 2019, 23(5): 1990-2001. 10.1109/jbhi.2018.2878048 |
22 | WANG T H, ZHANG L, HU W Y. Bridging deep and multiple kernel learning: a review[J]. Information Fusion, 2021, 67: 3-13. 10.1016/j.inffus.2020.10.002 |
23 | REBAI I, BenAYED Y, MAHDI W. Deep multilayer multiple kernel learning[J]. Neural Computing and Applications, 2016, 27(8): 2305-2314. 10.1007/s00521-015-2066-x |
24 | GAUTAM C, BALAJI R, SUDHARSAN K, et al. Localized multiple kernel learning for anomaly detection: one-class classification[J]. Knowledge-Based Systems, 2019, 165: 241-252. 10.1016/j.knosys.2018.11.030 |
25 | HE Q, ZHANG Q S, WANG H Y. Kernel-target alignment based multiple kernel one-class support vector machine [C]// Proceedings of the 2019 IEEE International Conference on Systems. Piscataway: IEEE, 2019: 228-235. 10.1109/smc.2019.8914503 |
26 | MEILǍ M. Data centering in feature space [C]// Proceedings of the 9th International Workshop on Artificial Intelligence and Statistics. New York: JMLR.org, 2003: 209-216. |
27 | CORTES C, MOHRI M, ROSTAMIZADEH A. Algorithms for learning kernels based on centered alignment[J]. Journal of Machine Learning Research, 2012, 13: 795-828. |
28 | BEN-HUR A, NOBLE W S. Kernel methods for predicting protein-protein interactions[J]. Bioinformatics, 2005, 21(S1): i38-i46. 10.1093/bioinformatics/bti1016 |
29 | CORTES C, MOHRI M, ROSTAMIZADEH A. Two-stage learning kernel algorithms [C]// Proceedings of the 27th International Conference on Machine Learning. Madison, WI: Omnipress, 2010: 239-246. |
30 | DAMOULAS T, GIROLAMI M A. Pattern recognition with a Bayesian kernel combination machine[J]. Pattern Recognition Letters, 2009, 30(1): 46-54. 10.1016/j.patrec.2008.08.016 |
31 | DUA D, GRAFF C. UCI machine learning repository[DS/OL]. [2021-05-08]. . |
32 | WU M R, YE J P. A small sphere and large margin approach for novelty detection using training data with outliers[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2009, 31(11): 2088-2092. 10.1109/tpami.2009.24 |
33 | LEWIS D P, JEBARA T, NOBLE W S. Nonstationary kernel combination [C]// Proceedings of the 23rd International Conference on Machine Learning. New York: ACM, 2006: 553-560. 10.1145/1143844.1143914 |
[1] | 周佳航, 邢红杰. 基于双自编码器和Transformer网络的异常检测方法[J]. 《计算机应用》唯一官方网站, 2023, 43(1): 22-29. |
[2] | 邱云志, 汪廷华, 戴小路. 双重特征加权模糊支持向量机[J]. 《计算机应用》唯一官方网站, 2022, 42(3): 683-687. |
[3] | 王梅, 宋晓晖, 刘勇, 许传海. 神经正切核K‑Means聚类[J]. 《计算机应用》唯一官方网站, 2022, 42(11): 3330-3336. |
[4] | 杨悦, 王士同. 基于随机特征映射的四层多核学习方法[J]. 《计算机应用》唯一官方网站, 2022, 42(1): 16-25. |
[5] | 王梅, 许传海, 刘勇. 基于神经正切核的多核学习方法[J]. 《计算机应用》唯一官方网站, 2021, 41(12): 3462-3467. |
[6] | 効琦, 尹增山, 高爽. 基于检测与跟踪相互迭代的极暗弱目标搜索算法[J]. 计算机应用, 2021, 41(10): 3017-3024. |
[7] | 李瑶, 赵云芃, 李欣芸, 刘志芬, 陈俊杰, 郭浩. 基于sparse group Lasso方法的脑功能超网络构建与特征融合分析[J]. 计算机应用, 2020, 40(1): 62-70. |
[8] | 白东颖, 易亚星, 王庆超, 余志勇. 面向概念漂移问题的渐进多核学习方法[J]. 计算机应用, 2019, 39(9): 2494-2498. |
[9] | 李飞, 杜亮, 任超宏. 基于全局融合的多核概念分解算法[J]. 计算机应用, 2019, 39(4): 1021-1026. |
[10] | 尚方信, 郭浩, 李钢, 张玲. 基于One-class SVM的噪声图像分割方法[J]. 计算机应用, 2019, 39(3): 874-881. |
[11] | 孙石磊, 王超, 赵元棣. 基于轮廓系数的参数无关空中交通轨迹聚类方法[J]. 计算机应用, 2019, 39(11): 3293-3297. |
[12] | 范君, 王新, 徐慧. 粒子群优化混合核极限学习机的构造煤厚度预测方法[J]. 计算机应用, 2018, 38(6): 1820-1825. |
[13] | 刘万军, 秦济韬, 曲海成. 基于改进单类支持向量机的工业控制网络入侵检测方法[J]. 计算机应用, 2018, 38(5): 1360-1365. |
[14] | 忽丽莎, 王素贞, 陈益强, 胡春雨, 蒋鑫龙, 陈振宇, 高兴宇. 基于目标均衡度量的核增量学习跌倒检测方法[J]. 计算机应用, 2018, 38(4): 928-934. |
[15] | 张乐园, 李佳烨, 李鹏清. 低秩约束的非线性属性选择算法[J]. 计算机应用, 2018, 38(12): 3444-3449. |
阅读次数 | ||||||
全文 |
|
|||||
摘要 |
|
|||||