您的位置:山东大学 -> 科技期刊社 -> 《山东大学学报(工学版)》

山东大学学报(工学版) ›› 2018, Vol. 48 ›› Issue (3): 134-139.doi: 10.6040/j.issn.1672-3961.0.2017.416

• • 上一篇    下一篇

一种基于聚类的过抽样算法

王换,周忠眉   

  1. 闽南师范大学计算机学院, 福建 漳州 363000
  • 收稿日期:2017-08-24 出版日期:2018-06-20 发布日期:2017-08-24
  • 作者简介:王换(1990— ),女,河南安阳人,硕士研究生,主要研究方向为数据挖掘. E-mail:704807435@qq.com
  • 基金资助:
    国家自然科学基金资助项目(61170129)

An over sampling algorithm based on clustering

WANG Huan, ZHOU Zhongmei   

  1. School of Computer, Minnan Normal University, Zhangzhou 363000, Fujian, China
  • Received:2017-08-24 Online:2018-06-20 Published:2017-08-24

摘要: 在过抽样技术研究中,为了合成较有意义的新样本,提出一种基于聚类的过抽样算法ClusteredSMOTE-Boost。过滤小类的噪声样本,将剩余的每个小类样本作为目标样本参与合成新样本。对整个训练集聚类,根据聚类后目标样本所在簇的特点确定其权重及合成个数。将所有目标样本聚类,在目标样本所在的簇内选取K个近邻,并从中任选一个与目标样本合成新样本,使新样本与目标样本簇内的样本尽量相似,并减少由于添加样本而造成的边界复杂度。试验结果表明,ClusteredSMOTE-Boost算法在各个度量上均明显优于SMOTE-Boost、ADASYN-Boost和BorderlineSMOTE-Boost三种经典算法。

关键词: 过抽样, 样本权重, 聚类, 分类, 不平衡数据

Abstract: In the research of over sampling, in order to generate meaningful new samples, the ClusteredSMOTE-Boost was proposed, which was based on the clustering technique. The algorithm filtered the noisy of minority class samples and took the remaining minority class samples as target samples to synthesize new samples. According to characteristics of the cluster of target samples after clustering determined the weight and the number of the target samples for the whole training set. All target samples were clustered and K-nearest neighbors in the cluster of the target sample were selected, and then a sample from K-nearest neighbors was randomly chosen to synthesize new sample with target sample. Thus, new samples were similar with samples in the target cluster. This method reduced the complexity of the boundary caused by the additional new samples. The experimental results showed that the ClusteredSMOTE-Boost algorithm was superior to the three classical algorithms SMOTE-Boost, ADASYN-Boost, BorderlineSMOTE-Boost on the variety of measures.

Key words: over sampling, instance weights, classification, cluster, imbalanced data

中图分类号: 

  • TP311
[1] WANG S, YAO X. Multi-class imbalance problems: analysis and potential solutions[J]. IEEE Transactions on Systems Man & Cybernetics: Part B, 2012, 42(4):1119-1130.
[2] HE H, GARCIA E A. Learning from imbalanced data[J]. IEEE Transactions on Knowledge & Data Engineering, 2009, 21(9):1263-1284.
[3] KUMAR M, BHUTANI K, AGGARWAL S. Hybrid model for medical diagnosis using neutrosophic cognitive maps with genetic algorithms[C] //IEEE International Conference on Fuzzy Systems. Istanbul, Turkey: IEEE, 2015:1-7.
[4] SRIVASTAVA A, KUNDU A, SURAL S, et al. Credit card fraud detection using hidden Markov model[J]. IEEE Transactions on Dependable & Secure Computing, 2008, 5(1):37-48.
[5] LI J, FONG S, MOHAMMED S, et al. Improving the classification performance of biological imbalanced datasets by swarm optimization algorithms[J]. Journal of Supercomputing, 2016, 72(10): 3708-3728.
[6] 杨明, 尹军梅, 吉根林. 不平衡数据分类方法综述[J]. 南京师范大学学报(工程技术版), 2008, 8(4): 7-12. YANG Ming, YIN Junmei, JI Genlin. Classification methods on imbalanced data: a survey[J]. Journal of Nanjing Normal University(Engineering and Technology Edition), 2008, 8(4): 7-12.
[7] SUN Z, SONG Q, ZHU X, et al. A novel ensemble method for classifying imbalanced data[J]. Pattern Recognition, 2015, 48(5):1623-1637.
[8] SAEZ J A, LUENGO J, STEFANWSKI J, et al. SMOTE—IPF: addressing the noisy and borderline examples problem in imbalanced classification by a re-sampling method with filtering[J]. Information Sciences, 2015, 291(5):184-203.
[9] RAMENTOL E, CABALLERO Y, BELLO R, et al. SMOTE-RSB*: a hybrid preprocessing approach based on oversampling and under sampling for high imbalanced data-sets using SMOTE and rough sets theory[J]. Knowledge & Information Systems, 2012, 33(2):245-265.
[10] BARUA S, ISLAM M M, YAO X, et al. MWMOTE: majority weighted minority over sampling technique for imbalanced data set learning[J]. IEEE Transactions on Knowledge & Data Engineering, 2014, 26(2):405-425.
[11] BORAL A, CYGAN M, KOCIUMAKA T, et al. A fast branching algorithm for cluster vertex deletion[J]. Theory of Computing Systems, 2016, 58(2):357-376.
[12] FOMIN S, GRIGORIEV D, KOSHEVOY G. Subtraction-free complexity, cluster transformations, and spanning trees[J]. Foundations of Computational Mathematics, 2016, 16(1):1-31.
[13] DAVIES D L, BOULDIN D W. A cluster separation measure[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 1979, 1(2):224.
[14] ZENG H J, HE Q C, CHEN Z, et al. Learning to cluster web search results[C] //International ACM SIGIR Conference on Research and Development in Information Retrieval. Sheffield, UK: ACM, 2004:210-217.
[15] 胡小生, 张润晶, 钟勇. 一种基于聚类提升的不平衡数据分类算法[J]. 集成技术, 2014(2):35-41. HU Xiaosheng, ZHANG Runjing, ZHONG Yong. A clustering-based enhanced classification algorithm for imbalanced data[J]. Journal of Integration Technology, 2014(2):35-41.
[16] CHAWLA N V, BOWYER K W, HALL L O, et al. SMOTE: synthetic minority over-sampling technique[J]. Journal of Artificial Intelligence Research, 2011, 16(1):321-357.
[17] HE H, BAI Y, GARCIA E A, et al. ADASYN: adaptive synthetic sampling approach for imbalanced learning[C] //IEEE International Joint Conference on Neural Networks. Hoboken, USA: IEEE, 2008:1322-1328.
[18] HAN H, WANG W Y, MAO B H. Borderline-SMOTE: a new over-sampling method in imbalanced data sets learning[J]. Lecture Notes in Computer Science, 2005, 3644(5):878-887.
[19] CHAWLA N V, LAZAREVIC A, HALL L O, et al. SMOTEBoost: improving prediction of the minority class in boosting[J]. Lecture Notes in Computer Science, 2003, 2838:107-119.
[20] BUNKHUMPORNPAT C, SINAPIROMSARAN K, LURSINSAP C. Safe-Level-SMOTE: safe-level-synthetic minority over-sampling technique for handling the class imbalanced problem[C] //Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining. Berlin, Germany: Springer, 2009:475-482.
[1] 张璞,刘畅,王永. 基于特征融合和集成学习的建议语句分类模型[J]. 山东大学学报(工学版), 2018, 48(5): 47-54.
[2] 曹雅,邓赵红,王士同. 基于单调约束的径向基函数神经网络模型[J]. 山东大学学报(工学版), 2018, 48(3): 127-133.
[3] 龙柏,曾宪宇,李徵,刘淇. 电商商品嵌入表示分类方法[J]. 山东大学学报(工学版), 2018, 48(3): 17-24.
[4] 谢志峰,吴佳萍,马利庄. 基于卷积神经网络的中文财经新闻分类方法[J]. 山东大学学报(工学版), 2018, 48(3): 34-39.
[5] 张佩瑞,杨燕,邢焕来,喻琇瑛. 基于核K-means的增量多视图聚类算法[J]. 山东大学学报(工学版), 2018, 48(3): 48-53.
[6] 王婷婷,翟俊海,张明阳,郝璞. 基于HBase和SimHash的大数据K-近邻算法[J]. 山东大学学报(工学版), 2018, 48(3): 54-59.
[7] 陈嘉杰,王金凤. 基于蚁群算法求解Choquet模糊积分模型[J]. 山东大学学报(工学版), 2018, 48(3): 81-87.
[8] 读习习,刘华锋,景丽萍. 一种融合社交网络的叠加联合聚类推荐模型[J]. 山东大学学报(工学版), 2018, 48(3): 96-102.
[9] 杨天鹏,徐鲲鹏,陈黎飞. 非均匀数据的变异系数聚类算法[J]. 山东大学学报(工学版), 2018, 48(3): 140-145.
[10] 叶明全,高凌云,万春圆. 基于人工蜂群和SVM的基因表达数据分类[J]. 山东大学学报(工学版), 2018, 48(3): 10-16.
[11] 庞人铭,王波,叶昊,张海峰,李明亮. 基于PCA相似度和谱聚类相结合的高炉历史数据聚类[J]. 山东大学学报(工学版), 2017, 47(5): 143-149.
[12] 王磊,邓晓刚,曹玉苹,田学民. 基于MLFDA的化工过程故障模式分类方法[J]. 山东大学学报(工学版), 2017, 47(5): 179-186.
[13] 李素姝,王士同,李滔. 基于LS-SVM与模糊补准则的特征选择方法[J]. 山东大学学报(工学版), 2017, 47(3): 34-42.
[14] 何其佳,刘振丙,徐涛,蒋淑洁. 基于LBP和极限学习机的脑部MR图像分类[J]. 山东大学学报(工学版), 2017, 47(2): 86-93.
[15] 郭超,杨燕,江永全,宋祎. 基于多视图分类集成的高铁工况识别[J]. 山东大学学报(工学版), 2017, 47(1): 7-14.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!