计算机科学 ›› 2017, Vol. 44 ›› Issue (Z11): 88-91.doi: 10.11896/j.issn.1002-137X.2017.11A.017
姚明海,林宣民,王宪保
YAO Ming-hai, LIN Xuan-min and WANG Xian-bao
摘要： 为了提高大规模高维度数据的训练速度和分类精度,提出了一种基于局部敏感哈希的SVM快速增量学习方法。算法首先利用局部敏感哈希能快速查找相似数据的特性,在SVM算法的基础上筛选出增量中可能成为SV的样本,然后将这些样本与已有SV一起作为后续训练的基础。使用多个数据集对该算法进行了验证。实验表明,在大规模增量数据样本中,提出的SVM快速增量学习算法能有效地提高训练学习的速度,并能保持有效的准确率。
| [1] ATTAR V,SINHA P,WANKHADE K.A fast and light classifer for data streams[J].Evol Syst,2010,1(3):199-207. [2] VAPNIK V.The Nature of Statistical Learning Theory[M].NewYork:Springer Verlag,1995. [3] SYED N A,SUNG K.Handling Concept Drifts in Incremental Learning with Support Vector Machines[C]∥Proc.of the 5th ACM SIGKDD International Confe-rence.1999:316-321. [4] DIEHL C P,CAUWENBERGHS G.SVM Incremental Learning Adaptation and Optimization[C]∥ International Joint Con-ference on Neural Networks.IEEE,2003:2685-2691. [5] 茅嫣蕾,魏赟.一种基于KKT条件和壳向量的 SVM 增量学习算法[J].电子科技,2016,29(2):38-40. [6] 李妍坊,苏波,刘功申.一种基于组合保留集的SVM增量学习算法[J].上海交通大学学报,2016,50(7):1054-1059. [7] 曹健,孙世宇,段修生,等.基于 KKT 条件的 SVM 增量学习算法[J].火力与指挥控制,2014(7):139-143. [8] LUO J,PRONOBIS A,CAPUTO B.Incremental Learning forPlace Recognition in Dynamic Environments[C]∥IEEE International Conference on Intelligent Robots and Systems.2007:721-729. [9] 张灿淋,姚明海,童小龙,等.一种新的基于 KKT 条件的错误驱动 SVM 增量学习算法[J].计算机系统与应用,2014,23(1):144-148. [10] JHALA I S,DALAL P.Optimized Incremental SVM based Clas-sifier for Spam Filtering using Internet Acronyms[J].International Journal for Innovative Research in Science & Technology,2015,2(1):2349-6010. [11] CHAKROUN M,WALI A,ARIBI Y,et al.Video event detection using auto-associative neural network and incremental SVM models[C]∥International Conference on Intelligent System Design & Application.2015:563-568. [12] JAGTAP R V,POTEY M A.Recognition of Human Activityusing Incremental SVM[J].Imperial Journal of Interdisciplinary Research,2016,7(2):2454-2462. [13] ANDONI A,INDYK P,NGUYEN H L.Beyond locality-sensitive hash [C]∥Proceedings of the Twenty-Fifth Annual ACM-SIAM Symposium on Discrete Algorithms.2014:1018-1028. [14] GIONIS A,INDYK P,MOTWANI R.Similarity search in high dimensions via hashing[C]∥Proceedings of the International Conference on Very Large Databases.1999:25-38. [15] DATAR M,IMMORLICA N,INDYK P,et al.Locality sensitive hashing scheme based on p-stable distributions[C]∥Procee-dings of the ACM Symposium on Computational Geometry.2004:23-36. |
| No related articles found! |
|
||
首页