一种基于原型学习的自适应概念漂移分类方法

doi:10.13190/j.jbupt.2017.03.005

北京邮电大学学报 ›› 2017, Vol. 40 ›› Issue (3): 43-50.doi: 10.13190/j.jbupt.2017.03.005

一种基于原型学习的自适应概念漂移分类方法

苏静¹, 裘晓峰¹, 李书芳¹, 刘道伟², 张春红¹

1. 北京邮电大学网络体系构建与融合北京市重点实验室, 北京 100876;
2. 中国电力科学研究院, 郑州 450052

收稿日期:2016-07-23 出版日期:2017-06-28 发布日期:2017-05-25
作者简介:苏静(1991-),女,硕士生,E-mail:sj199137@126.com;裘晓峰(1969-),女,副教授,硕士生导师.
基金资助:
国家电网公司科技项目（XT71-15-056）

A Prototype-Based Adaptive Concept Drift Classification Method

SU Jing¹, QIU Xiao-feng¹, LI Shu-fang¹, LIU Dao-wei², ZHANG Chun-hong¹

1. Beijing Key Laboratory of Network System Architecture and Convergence, Beijing University of Posts and Telecommunications, Beijing 100876, China;
2. China Electric Power Research Institute, Zhengzhou 450052, China

Received:2016-07-23 Online:2017-06-28 Published:2017-05-25

摘要/Abstract

摘要： 为了更准确快速地处理或适应概念漂移，提出了基于原型学习的数据流分类算法，基于发掘并优化现有方法存在的问题，提出了新的方法模型SyncPrototype，在预测方法、原型判定与更新方法等处理概念漂移问题的关键部分做出了新的尝试与优化.实验结果证明，相较于现有方法，SyncPrototype模型在分类性能、概念漂移的响应速度以及时间性能等方面都有明显提高，能够更加有效处理并适应数据流概念漂移问题.

关键词: 数据流, 概念漂移, 分类

Abstract: As a frequent problem that needs to be mainly dealt with in supervised learning scenario of streaming data, the concept drift, primarily, occurs when the data distribution or the target variable changes over time. As typical data streams, the research method which real-time solves or adapts to the concept drift of data streams can provide strong support for grid security dispatch and stable control of real-time decision-making. For accurate and quick dealing with or adapting to concept drift, a prototype-based learning algorithm of data streams classification is discussed. Based on improving the problems which have been explored in existing algorithm, a new algorithm SyncPrototype was proposed,which makes new optimization in terms of methods of classification method, prototype construction and updating. Experiment shows that SyncPrototype can outperforms the existing algorithm in terms of classification performance,time performance and response rate.

Key words: data streams, concept drift, classification

中图分类号:

TN992.53

苏静, 裘晓峰, 李书芳, 刘道伟, 张春红. 一种基于原型学习的自适应概念漂移分类方法[J]. 北京邮电大学学报, 2017, 40(3): 43-50.

SU Jing, QIU Xiao-feng, LI Shu-fang, LIU Dao-wei, ZHANG Chun-hong. A Prototype-Based Adaptive Concept Drift Classification Method[J]. JOURNAL OF BEIJING UNIVERSITY OF POSTS AND TELECOM, 2017, 40(3): 43-50.

参考文献

[1] Garofalakis M, Gehrke J, Rastogi R. Querying and mining data streams:you only get one look a tutorial[C]//Proceedings of the 2002 ACM SIGMOD International Conference on Management of Data. ACM, 2002:635-635.
[2] Rosch E H. Natural categories[J]. Cognitive Psychology, 1973, 4(3):328-350.
[3] Hoens T R, Polikar R, Chawla N V. Learning from streaming data with concept drift and imbalance:an overview[J]. Progress in Artificial Intelligence, 2012, 1(1):89-101.
[4] Shao J, Ahmadi Z, Kramer S. Prototype-based learning on concept-drifting data streams[C]//Proceedings of the 20th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. ACM, 2014:412-421.
[5] Domingos P, Hulten G. Mining high-speed data streams[C]//Proceedings of the Sixth ACM SIGKDD International Conference on Knowledge Discovery and Data mining. ACM, 2000:71-80.
[6] Hulten G, Spencer L, Domingos P. Mining time-changing data streams[C]//Proceedings of the Seventh ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. ACM, 2001:97-106.
[7] Bifet A, Gavaldà R. Adaptive learning from evolving data streams[C]//International Symposium on Intelligent Data Analysis. Berlin Heidelberg:Springer, 2009:249-260.
[8] Liu G, Cheng H, Qin Z, et al. E-CVFDT:An improving CVFDT method for concept drift data stream[C]//Communications, Circuits and Systems (ICCCAS), 2013 International Conference on. IEEE, 2013, 1:315-318.
[9] Buntine W. Learning classification trees[J]. Statistics and Computing, 1992, 2(2):63-73.
[10] Pfahringer B, Holmes G, Kirkby R. New options for hoeffding trees[C]//Australasian Joint Conference on Artificial Intelligence. Berlin Heidelberg:Springer, 2007:90-99.
[11] Bifet A, Holmes G, Pfahringer B, et al. New ensemble methods for evolving data streams[C]//Proceedings of the 15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Paris:ACM, 2009:139-148.
[12] Kubát M. Floating approximation in time-varying knowledge bases[J]. Pattern Recognition Letters, 1989, 10(4):223-227.
[13] Bifet A, Gavalda R. Learning from Time-Changing Data with Adaptive Windowing[C]//SDM. 2007, 7:2007.
[14] Widmer G, Kubat M. Effective learning in dynamic environments by explicit context tracking[C]//European Conference on Machine Learning. Berlin Heidelberg:Springer, 1993:227-243.
[15] Alippi C, Boracchi G, Roveri M. Just in time classifiers:managing the slow drift case[C]//2009 International Joint Conference on Neural Networks. Atlanta:IEEE, 2009:114-120.
[16] Street W N, Kim Y S. A streaming ensemble algorithm (SEA) for large-scale classification[C]//Proceedings of the seventh ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. San Francisco:ACM, 2001:377-382.
[17] Kolter J Z, Maloof M A. Dynamic weighted majority:A new ensemble method for tracking concept drift[C]//Proceedings of the Third IEEE International Conference on Data Mining. Washington DC:IEEE, 2003:123-130.
[18] Chu F, Zaniolo C. Fast and light boosting for adaptive mining of data streams[C]//Pacific-Asia Conference on Knowledge Discovery and Data Mining. Berlin Heidelberg:Springer, 2004:282-292.
[19] Brzezinski D, Stefanowski J. Reacting to different types of concept drift:The accuracy updated ensemble algorithm[J]. IEEE Transactions on Neural Networks and Learning Systems, 2014, 25(1):81-94.
[20] Brzezinski D, Stefanowski J. Combining block-based and online methods in learning ensembles from concept drifting data streams[J]. Information Sciences, 2014, 265(5):50-67.

一种基于原型学习的自适应概念漂移分类方法

A Prototype-Based Adaptive Concept Drift Classification Method

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

[1]	张天魁蔡昌利骆晓亮朱禹涛. 基于多尺度特征Transformer的细粒度图像分类方法[J]. 北京邮电大学学报, 2023, 46(4): 70-75.
[2]	包乾辉文娟石淑珍董萌萍刘雪. 融合句法依存与 BERT-Att-BiLSTM 的属性分类[J]. 北京邮电大学学报, 2023, 46(4): 123-128.
[3]	谢丽霞刘思雨杨宏宇张良. 一种基于分类代价的软件缺陷评估方法[J]. 北京邮电大学学报, 2023, 46(4): 116-122.
[4]	赵海英王梓舟. 基于多层判别字典学习的传统服饰图像分类算法[J]. 北京邮电大学学报, 2023, 46(2): 104-108.
[5]	谢永红陶浒贾麒杨石兵韩辛亮. 基于预训练语言模型的中医症状标准化方法[J]. 北京邮电大学学报, 2022, 45(4): 14-20.
[6]	谢永红, 陶浒, 贾麒, 杨石兵, 韩辛亮. 基于预训练语言模型的中医症状标准化方法[J]. 北京邮电大学学报, 2022, 45(4): 13-18,57.
[7]	李建桂, 梁越, 高鹏飞, 刘绍华, 马应龙. 基于层次化类别嵌入的电力ICT系统故障分类[J]. 北京邮电大学学报, 2021, 44(4): 34-40.
[8]	曹毅, 黄子龙, 盛永健, 刘晨, 费鸿博. D-2-DenseNet噪音鲁棒的城市音频分类模型[J]. 北京邮电大学学报, 2021, 44(1): 86-91.
[9]	刘逸琛, 孙华志, 马春梅, 姜丽芬, 钟长鸿. 一种基于高层特征融合的网络商品分类[J]. 北京邮电大学学报, 2020, 43(5): 98-104,117.
[10]	常紫英, 王文涵, 李涛, 刘芬, 陈朋朋. 一种基于Shapelet算法的指纹定位方法[J]. 北京邮电大学学报, 2020, 43(4): 95-100.
[11]	蒋芳, 张南飞, 胡艳军, 王翊. 基于BP神经网络的CSI无源目标分类方法[J]. 北京邮电大学学报, 2020, 43(1): 40-45.
[12]	傅群超, 王枞. 用于文本分类的多探测任务语言模型微调[J]. 北京邮电大学学报, 2019, 42(6): 76-83.
[13]	刘寒. 混合采样与遗传算法相结合的垃圾网页检测[J]. 北京邮电大学学报, 2019, 42(6): 111-117.
[14]	彭聪, 孙岩, 戚鹏. 基于标签关联性的多标签Scratch分类算法[J]. 北京邮电大学学报, 2019, 42(6): 134-141.
[15]	刘寅生, 段洪涛, 范振雄, 李蓉, 李景春. 基于改进MUSIC算法的短波非规则天线阵列测向系统[J]. 北京邮电大学学报, 2019, 42(5): 42-47.