[1]邓廷权,王强.半监督类保持局部线性嵌入方法[J].智能系统学报,2021,16(1):98-107.[doi:10.11992/tis.202003007]
DENG Tingquan,WANG Qiang.Semi-supervised class preserving locally linear embedding[J].CAAI Transactions on Intelligent Systems,2021,16(1):98-107.[doi:10.11992/tis.202003007]
点击复制
DENG Tingquan,WANG Qiang.Semi-supervised class preserving locally linear embedding[J].CAAI Transactions on Intelligent Systems,2021,16(1):98-107.[doi:10.11992/tis.202003007]
半监督类保持局部线性嵌入方法
《智能系统学报》[ISSN 1673-4785/CN 23-1538/TP] 卷:
16
期数:
2021年第1期
页码:
98-107
栏目:
学术论文—知识工程
出版日期:
2021-01-05
- Title:
- Semi-supervised class preserving locally linear embedding
- Keywords:
- nonlinear feature extraction; manifold learning; semi-supervised; labeled information; clustering; visualization
- 分类号:
- TP181
- 摘要:
- 为使局部线性嵌入（local linear embedding, LLE）这一无监督高维数据的非线性特征提取方法提取出的特征在分类或聚类学习上更优，提出一种半监督类保持局部线性嵌入(semi-supervised class preserving local linear embedding, SSCLLE)的非线性特征提取方法。该方法将半监督信息融入到LLE中，首先对标记样本近邻赋予伪标签，增大标记样本数量。其次，对标记样本之间的距离进行局部调整，缩小同类样本间距，扩大异类样本间距。同时在局部线性嵌入优化目标函数中增加全局同类样本间距和异类样本间距的约束项，使得提取出的低维特征可以确保同类样本点互相靠近，而异类样本点彼此分离。在一系列实验中，其聚类精确度以及可视化效果明显高于无监督LLE和现有半监督流特征提取方法，表明该方法提取出的特征具有很好的类保持特性。
- Abstract:
- To make local linear embedding (LLE), the nonlinear feature extraction method for unsupervised high-dimensional data, more optimal in classification or clustering learning, we propose a nonlinear semi-supervised class preserving local linear embedding (SSCLLE) feature extraction method. This method integrates semi-supervised information into LLE. First, pseudo-labels are assigned to the nearby neighbors of the labeled samples to increase the number of labeled samples. Second, the distance between the labeled samples is partially adjusted to reduce the distance between similar samples and expand the distance between heterogeneous samples. Simultaneously, the constraints of the globally same sample spacing and heterogeneous sample spacing are added in the local linear embedding optimization objective function so that the extracted low-dimensional features can ensure that the same sample points are near each other, whereas the heterogeneous sample points are separated from each other. In a series of experiments, the clustering accuracy and visualization effect of the proposed method are significantly higher than those of unsupervised LLE and the existing semi-supervised flow feature extraction methods, indicating that the features extracted by this method have good class retention characteristics.
- 参考文献/References:
-
[1] LIU Feng, ZHANG Weijie, GU Suicheng. Local linear laplacian eigenmaps:a direct extension of LLE[J]. Pattern recognition letters, 2016, 75:30-35.
[2] JIANG Bo, DING C, LUO Bin. Robust data representation using locally linear embedding guided PCA[J]. Neurocomputing, 2018, 275:523-532.
[3] WANG Qian, WANG Weiguo, NIAN Rui, et al. Manifold learning in local tangent space via extreme learning machine[J]. Neurocomputing, 2016, 174:18-30.
[4] TANG Z, LAO H. Robust image hashing via DCT and LLE[J]. Computers and security, 2016, 62:133-148.
[5] ZHANG Yan, ZHANG Zhao, QIN Jie, et al. Semi-supervised local multi-manifold Isomap by linear embedding for feature extraction[J]. Pattern recognition, 2018, 76:622-678.
[6] LIU Zhonghua, WANG Xiaohong, PU Jiexin, et a. Nonnegative low-rank representation based manifold embedding for semi-supervised learning[J]. Knowledge-based systems, 2017, 136:121-129.
[7] CHEN Lin, YANG Meng. Semi-supervised dictionary learning with label propagation for image classification[J]. Computational visual media, 2017, 3(1):83-94.
[8] MIKALSEN K O, SOGUERO-RUIZ C, BIANCHI F M, et al. Noisy multi-label semi-supervised dimensionality reduction[J]. Pattern recognition, 2019, 90:257-270.
[9] PARK S H, KIM S B. Active semi-supervised learning with multiple complementary information[J]. Expert systems with applications, 2019, 126:30-40.
[10] ZHENG Feng, SONG Zhan, SHAO Ling, et al. A semi-supervised approach for dimensionality reduction with distributional similarity[J]. Neurocomputing, 2013, 103:210-221.
[11] SUN Shiliang, HUSSAIN Z, SHAWE-TAYLOR J. Manifold-preserving graph reduction for sparse semi-supervised learning[J]. Neurocomputing, 2014, 124:13-21.
[12] KIM K. An improved semi-supervised dimensionality reduction using feature weighting:application to sentiment analysis[J]. Expert systems with applications, 2018, 109:49-65.
[13] ROWEIS S T, SAUL L J. Nonlinear dimensionality reduction by locally linear embedding[J]. Science, 2000, 290(5500):2323-2326.
[14] LANGLEY P, PAZZANI M J, FISHER D H. Concept formation:knowledge and experience in unsupervised learning[M]. San Mateo:Morgan Kaufmann Publishers, 1991.
[15] YANG Bo, XIANG Ming, ZHANG Yupei. Multi-manifold discriminant isomap for visualization and classification[J]. Pattern recognition, 2016:215-230.
[16] 张长帅, 周大可, 杨欣. 一种基于核的半监督局部线性嵌入方法[J]. 计算机工程, 2011, 37(20):157-159
ZHANG Changshuai, ZHOU Dake, YANG Xin. Method of kernel-based semi-supervised local linear embedding[J]. Computer engineering, 2011, 37(20):157-159
[17] KIM K, LEE J. Sentiment visualization and classification via semi-supervised nonlinear dimensionality reduction[J]. Pattern recognition, 2014, 47(2):758-768.
[18] COSTA J A, HERO III A O. Classification constrained dimensionality reduction[C]//Proceedings of IEEE International Conference on Acoustics, Speech, and Signal Processing. Philadelphia, USA, 2005:1077-1080.
[19] MARCILLA A, REYES-LABARTA J A, OLAYA M M. Should we trust all the published LLE correlation parameters in phase equilibria? Necessity of their assessment prior to publication[J]. Fluid phase equilibria, 2017, 433:243-252.
[20] BEZDEK J C. Pattern recognition with fuzzy objective function algorithms[M]. New York:Plenum, 1981:35-36.
- 相似文献/References:
-
[1]徐 蓉,姜 峰,姚鸿勋.流形学习概述[J].智能系统学报,2006,1(1):44.
XU Rong,JIANG Feng,YAO Hong-xun.Overview of manifold learning[J].CAAI Transactions on Intelligent Systems,2006,1():44.
[2]谢朝霞,穆志纯,谢建军.基于LLE的多姿态人耳识别[J].智能系统学报,2008,3(4):321.
XIE Zhao-xia,MU Zhi-chun,XIE J ian-jun.Multi-pose ear recogn ition based on locally linear embedding[J].CAAI Transactions on Intelligent Systems,2008,3():321.
[3]文贵华,江丽君,文 军.局部测地距离估计的Hessian局部线性嵌入[J].智能系统学报,2008,3(5):429.
WEN Gui-hua,J IANG L i-jun,WEN Jun.Using locally estimated geodesic distances to improve Hessian local linear embedding[J].CAAI Transactions on Intelligent Systems,2008,3():429.
[4]刘? 琚,乔建苹.基于学习的超分辨率重建技术[J].智能系统学报,2009,4(3):199.
LIU Ju,QIAO Jian-ping.Learningbased superresolution reconstruction[J].CAAI Transactions on Intelligent Systems,2009,4():199.
[5]谈超,关佶红,周水庚.增量与演化流形学习综述[J].智能系统学报,2012,7(5):377.
TAN Chao,GUAN Jihong,ZHOU Shuigeng.Incremental and evolutionary manifold learning: a survey[J].CAAI Transactions on Intelligent Systems,2012,7():377.
[6]练浩,曾宪华,李淑芳.有监督全局流形排序的图像检索算法[J].智能系统学报,2014,9(1):92.[doi:10.3969/j.issn.1673-4785.201303021]
LIAN Hao,ZENG Xianhua,LI Shufang.Supervised global manifold ranking based image retrieval algorithm[J].CAAI Transactions on Intelligent Systems,2014,9():92.[doi:10.3969/j.issn.1673-4785.201303021]
[7]张钢,谢晓珊,黄英,等.面向大数据流的半监督在线多核学习算法[J].智能系统学报,2014,9(3):355.[doi:10.3969/j.issn.1673-4785.201403067]
ZHANG Gang,XIE Xiaoshan,HUANG Ying,et al.An online multi-kernel learning algorithm for big data[J].CAAI Transactions on Intelligent Systems,2014,9():355.[doi:10.3969/j.issn.1673-4785.201403067]
[8]程旸,王士同.基于局部保留投影的多可选聚类发掘算法[J].智能系统学报,2016,11(5):600.[doi:10.11992/tis.201508022]
CHENG Yang,WANG Shitong.A multiple alternative clusterings mining algorithm using locality preserving projections[J].CAAI Transactions on Intelligent Systems,2016,11():600.[doi:10.11992/tis.201508022]
[9]徐慧敏,陈秀宏.图正则化稀疏判别非负矩阵分解[J].智能系统学报,2019,14(6):1217.[doi:10.11992/tis.201811021]
XU Huimin,CHEN Xiuhong.Graph-regularized, sparse discriminant, non-negative matrix factorization[J].CAAI Transactions on Intelligent Systems,2019,14():1217.[doi:10.11992/tis.201811021]
[10]余沁茹,卢桂馥,李华.自适应图正则化的低秩非负矩阵分解算法[J].智能系统学报,2022,17(2):325.[doi:10.11992/tis.202102007]
YU Qinru,LU Guifu,LI Hua.Nonnegative low-rank matrix factorization with adaptive graph neighbors[J].CAAI Transactions on Intelligent Systems,2022,17():325.[doi:10.11992/tis.202102007]
备注/Memo
收稿日期:2020-03-04。
基金项目:国家自然科学基金项目(11471001，61872104)
作者简介:邓廷权，教授，博士生导师，中国人工智能学会粒计算与知识发现专业委员会委员、黑龙江省工业与应用数学学会副理事长，主要研究方向为不确定性信息分析理论与方法、机器学习与数据挖掘、模式识别与人工智能。主持和参与国家自然科学基金面上项目各2项、主持多项省部级、国家重点实验室基金和横向项目。发表学术论文100余篇;王强，硕士研究生，主要研究方向为数据分析理论与方法.
通讯作者:王强. E-mail: 1005834631@qq.com
更新日期/Last Update:
2021-02-25