基于分布和逆文本类别指数的特征迁移加权算法

doi:10.11772/j.issn.1001-9081.2015.06.1643

计算机应用 ›› 2015, Vol. 35 ›› Issue (6): 1643-1648.DOI: 10.11772/j.issn.1001-9081.2015.06.1643

基于分布和逆文本类别指数的特征迁移加权算法

邱云飞¹, 刘世兴¹, 林明明¹, 邵良杉²

1. 辽宁工程技术大学软件学院, 辽宁葫芦岛 125105;
2. 辽宁工程技术大学系统工程研究所, 辽宁葫芦岛 125105

收稿日期:2014-12-22 修回日期:2015-03-17 发布日期:2015-06-12
通讯作者: 刘世兴(1990-),男,辽宁丹东人,硕士研究生,主要研究方向:数据挖掘、特征选择;494784913@qq.com
作者简介:邱云飞(1976-),男,辽宁阜新人,教授,博士,CCF会员,主要研究方向:数据挖掘、情感分析;林明明(1989-),女,辽宁大连人,硕士研究生,主要研究方向:数据挖掘、情感分析;邵良杉(1961-),男,辽宁凌源人,教授,博士,主要研究方向:数据挖掘、情感分析。
基金资助:
国家自然科学基金资助项目(70971059);辽宁省创新团队项目(2009T045);辽宁省高等学校杰出青年学者成长计划项目(LJQ2012027)。

Feature transfer weighting algorithm based on distribution and term frequency-inverse class frequency

QIU Yunfei¹, LIU Shixing¹, LIN Mingming¹, SHAO Liangshan²

1. School of Software, Liaoning Technical University, Huludao Liaoning 125105, China;
2. System Engineering Institute, Liaoning Technical University, Huludao Liaoning 125105, China

Received:2014-12-22 Revised:2015-03-17 Online:2015-06-12

摘要/Abstract

摘要：

传统机器学习面临一个难题,即当训练数据与测试数据不再服从相同分布时,由训练集得到的分类器无法对测试集文本准确分类。针对该问题,根据迁移学习原理,在源领域和目标领域的交集特征中,依据改进的特征分布相似度进行特征加权;在非交集特征中,引入语义近似度和新提出的逆文本类别指数(TF-ICF),对特征在源领域内进行加权计算,充分利用大量已标记的源领域数据和少量已标记的目标领域数据获得所需特征,以便快速构建分类器。在文本数据集20Newsgroups和非文本数据集UCI中的实验结果表明,基于分布和逆文本类别指数的特征迁移加权算法能够在保证精度的前提下对特征快速迁移并加权。

关键词: 迁移学习, 特征分布, 逆文本类别指数, 语义近似度, 特征加权

Abstract:

Traditional machine learning faces a problem: when the training data and test data no longer obey the same distribution, the classifier trained by training data can't classify test data accurately. To solve this problem, according to the transfer learning principle, the features were weighted according to the improved distribution similarity of source domain and target domain's intersection features. The semantic similarity and Term Frequency-Inverse Class Frequency (TF-ICF) were used to weight non-intersection features in source domain. Lots of labeled source domain data and a little labeled target domain were used to obtain the required features for building text classifier quickly. The experimental results on test dataset 20Newsgroups and non-text dataset UCI show that feature transfer weighting algorithm based on distribution and TF-ICF can transfer and weight features rapidly while guaranteeing precision.

Key words: transfer learning, feature distribution, Term Frequency-Inverse Class Frequency (TF-ICF), semantic similarity, feature weighting

中图分类号:

TP391.1

邱云飞, 刘世兴, 林明明, 邵良杉. 基于分布和逆文本类别指数的特征迁移加权算法[J]. 计算机应用, 2015, 35(6): 1643-1648.

QIU Yunfei, LIU Shixing, LIN Mingming, SHAO Liangshan. Feature transfer weighting algorithm based on distribution and term frequency-inverse class frequency[J]. Journal of Computer Applications, 2015, 35(6): 1643-1648.

参考文献

[1] LIU X. Text classification research based on instance transfer learning [D]. Changchun: Jilin University, 2014. (刘晓明.基于实例迁移学习的文本分类研究[D].长春:吉林大学, 2014.)
[2] XU M, WANG S, GU X. TL-SVM: A transfer learning algorithm[J]. Control and Decision, 2014, 29(1): 141-146. (许敏, 王士同, 顾鑫.TL-SVM: 一种迁移学习算法[J].控制与决策, 2014, 29(1):141-146.)
[3] GUO Y. Research of transfer learning based on single-source and mutil-source [D]. Xi'an: Xidian University, 2013. (郭勇.基于单源及多源的迁移学习方法研究[D].西安:西安电子科技大学, 2013.)
[4] WANG X, PAN J, CHENG Y, et al. Self-adaptive transfer for decision trees based on similarity metric[J]. Acta Automatic Sinica, 2013, 39(12): 2186-2192. (王雪松, 潘杰, 程玉虎, 等.基于相似度衡量的决策树自适应迁移[J].自动化学报, 2013, 39(12):2186-2192.)
[5] KALE D, YAN L. Transfer learning based on the observation probability of each attribute [C]//Proceedings of the 2013 IEEE 13th International Conference on Data Mining. Piscataway: IEEE, 2014: 3627-3631.
[6] WANG H, FAN S, SONG J, et al. Reinforcement learning transfer based on subgoal discovery and subtask similarity [J]. IEEE/CAA Journal of Automatica Sinica, 2014, 1(3): 257-266.
[7] FANG Z, ZHANG Z. Cross domain shared subspace learning for unsupervised transfer classification [C]//Proceedings of the 2014 22nd International Conference on Pattern Recognition. Stockholm: ICPR, 2014: 3927-3932.
[8] ZHU X. Cross-domain semi-supervised learning using feature formulation [J]. IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics, 2012: 6(41): 1627-1638.
[9] RAINA R, NG A Y, KOLLER D. Constructing informative priors using transfer learning [C]//Proceedings of the 23rd International Conference on Machine Learning. New York: ACM, 2006: 713-720.
[10] DAI W, XUE G, YANG Q, et al. Transferring native Bayes classifiers for text classification [C]//Proceedings of the 22rd AAAI Conference on Artificial Intelligence. Menlo Park: AAAI Press, 2007: 540-545.
[11] DAI W, XUE G, YANG Q, et al. Co-clustering based classification for out-of-domain documents [C]//Proceedings of the 13th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York: ACM, 2007: 210-219.
[12] ARNOLD A, NALLAPATI R, COHEN W. A comparative study of methods for transductive transfer learning [C]//Proceedings of the 7th IEEE International Conference on Data Mining Workshops. Washington, DC: IEEE Computer Society, 2007: 77-82.
[13] KALE D, LIU Y. Accelerating active learning with transfer learning[C]//Proceedings of the 2013 IEEE 13th International Conference on Data Mining. Piscataway: IEEE, 2013: 1085-1090.
[14] OYEN D, LANE T. Bayesian discovery of multiple Bayesian networks via transfer learning [C]//Proceedings of the 2013 IEEE 13th International Conference on Data Mining. Piscataway: IEEE, 2013: 577-586.
[15] WUP, DIETTERIEH G. Improving svm accuracy by training on auxiliary data sources [C]//Proceedings of the 21st International Conference on Machine Learning. San Francisco: Morgan Kaufmann, 2004: 871-878.
[16] ANDO S, SUZUKI E. Unsupervised cross-domain learning by interaction information co-clustering [C]//Proceedings of the 8th IEEE International Conference on Data Mining. Piscataway: IEEE, 2008: 13-22.
[17] KAMISHIMA T, HAMASAKI M, AKAHO S. TrBagg: A simple transfer learning method and its application to personalization in collaborative tagging [C]//Proceedings of the 9th IEEE International Conference of Data Mining. Piscataway: IEEE, 2009: 219-228.
[18] ZHANG Y, YEUNG D. Transfer metric learning by learning task relationships [C]//Proceedings of the 16th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. New York: ACM, 2010: 1199-1208.
[19] CHEN B, LAM W, TSANG I, et al. Extracting discriminative concepts for domain adaptation in text mining [C]//Proceedings of the 15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York: ACM, 2009: 179-188.
[20] WANG P, DOMENICONI C, HU J. Using Wikipeda for co-clustering based cross-domain text classification [C]//Proceedings of the 8th IEEE International Conference on Data Mining. Piscataway: IEEE, 2008: 1085-1090.
[21] GAO J, FAN W, JIANG J, et al. Knowledge transfer via multiple model local structure mapping [C]//Proceedings of the 14th ACM International Conference on Knowledge Discovery and Data Mining. New York: ACM, 2008: 283-291.
[22] ARNOLD A, NALLAPATI R, COHEN W. Exploiting feature hierarchy for transfer learning in named entity recognition [C]//Proceedings of the 2008 Annual Meeting of the Association for Computational Linguistics. Stroudsberg: Association for Computational Linguistics, 2008: 245-253.
[23] RICHMAN A, SCHONE P. Mining wiki resources for multilingual named entity recognition [C]//Proceedings of the 2008 Annual Meeting of the Association for Computational Linguistic. Stroudsberg: Association for Computational Linguistics, 2008: l-9.
[24] GOLDWASSER D, ROTH D. Active sample selection for named entity transliteration [C]//Proceedings of the 46th Annual Meeting of the Association for Computational Linguistics. Stroudsberg: Association for Computational Linguistics, 2008: 53-56.
[25] PING Y. Research on clustering and text categorization based on support vector machine [D]. Beijing: Beijing University of Posts and Telecommunications, 2012. (平源. 基于支持向量机的聚类及文本分类研究[D]. 北京: 北京邮电大学, 2012.)

基于分布和逆文本类别指数的特征迁移加权算法

Feature transfer weighting algorithm based on distribution and term frequency-inverse class frequency

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

[1]	翟飞宇, 马汉达. 基于DenseNet的经典-量子混合分类模型[J]. 《计算机应用》唯一官方网站, 2024, 44(6): 1905-1910.
[2]	时旺军, 王晶, 宁晓军, 林友芳. 小样本场景下的元迁移学习睡眠分期模型[J]. 《计算机应用》唯一官方网站, 2024, 44(5): 1445-1451.
[3]	黄天宇, 李远兴, 陈昊, 郭紫佳, 魏明军. 地空协同场景下加权模糊聚类用户簇划分方法[J]. 《计算机应用》唯一官方网站, 2024, 44(5): 1555-1561.
[4]	李鸿天, 史鑫昊, 潘卫国, 徐成, 徐冰心, 袁家政. 融合多尺度和注意力机制的小样本目标检测[J]. 《计算机应用》唯一官方网站, 2024, 44(5): 1437-1444.
[5]	王昊冉, 于丹, 杨玉丽, 马垚, 陈永乐. 面向工控系统未知攻击的域迁移入侵检测方法[J]. 《计算机应用》唯一官方网站, 2024, 44(4): 1158-1165.
[6]	刘秋杰, 万源, 吴杰. 深度双模态源域对称迁移学习的跨模态检索[J]. 《计算机应用》唯一官方网站, 2024, 44(1): 24-31.
[7]	陈克正, 郭晓然, 钟勇, 李振平. 基于负训练和迁移学习的关系抽取方法[J]. 《计算机应用》唯一官方网站, 2023, 43(8): 2426-2430.
[8]	金泽熙, 李磊, 刘继. 基于改进领域分离网络的迁移学习模型[J]. 《计算机应用》唯一官方网站, 2023, 43(8): 2382-2389.
[9]	轩勃娜, 李进, 宋亚飞, 马泽煊. 基于改进MobileNetV2的恶意代码分类方法[J]. 《计算机应用》唯一官方网站, 2023, 43(7): 2217-2225.
[10]	张慧斌, 冯丽萍, 郝耀军, 王一宁. 基于注意力机制和迁移学习的古壁画朝代识别[J]. 《计算机应用》唯一官方网站, 2023, 43(6): 1826-1832.
[11]	汪洋, 傅洪亮, 陶华伟, 杨静, 谢跃, 赵力. 基于决策边界优化域自适应的跨库语音情感识别[J]. 《计算机应用》唯一官方网站, 2023, 43(2): 374-379.
[12]	李传彪, 毕远伟. 基于跨域自适应的立体匹配算法[J]. 《计算机应用》唯一官方网站, 2023, 43(10): 3230-3235.
[13]	杨瑞杰, 郑贵林. 基于InceptionV3和特征融合的人脸活体检测[J]. 《计算机应用》唯一官方网站, 2022, 42(7): 2037-2042.
[14]	潘高峰, 樊渊, 汝玉, 郭予超. 基于点线特征融合的低纹理单目视觉同时定位与地图构建算法[J]. 《计算机应用》唯一官方网站, 2022, 42(7): 2170-2176.
[15]	陈颖, 于炯, 陈嘉颖, 杜旭升. 基于交叉层级数据共享的多任务模型[J]. 《计算机应用》唯一官方网站, 2022, 42(5): 1447-1454.