[1]陈秀平,王明文,万剑怡,等.基于Markov随机游走的渐进式半监督分类模型[J].江西师范大学学报(自然科学版),2014,(01):102-107.
 CHEN Xiu-ping,WANG Ming-wen,WAN Jian-yi,et al.The Progressively Semi-Supervised Classification Model Based on Markov Random Walk[J].,2014,(01):102-107.
点击复制

基于Markov随机游走的渐进式半监督分类模型()
分享到:

《江西师范大学学报》(自然科学版)[ISSN:1006-6977/CN:61-1281/TN]

卷:
期数:
2014年01期
页码:
102-107
栏目:
出版日期:
2014-02-28

文章信息/Info

Title:
The Progressively Semi-Supervised Classification Model Based on Markov Random Walk
作者:
陈秀平;王明文;万剑怡;左家莉
江西师范大学计算机信息工程学院,江西南昌,330022;江西师范大学初等教育学院,江西南昌,330027
Author(s):
CHEN Xiu-ping;WANG Ming-wen;WAN Jian-yi;ZUO Jia-li
关键词:
半监督分类渐进学习Markov随机游走迭代
Keywords:
semi-supervised classificationprogressive learningMarkov random walkiterating
分类号:
TP311
文献标志码:
A
摘要:
提出了一种基于Markov随机游走的渐进式半监督分类模型:在随机游走过程中,计算待标注数据到各类的迁移概率时,只考虑相应类别样本的影响,而忽略其他类别样本对随机过程的影响;并在学习过程中借鉴渐进学习思想,通过不断地“纠正”半监督学习过程中的“错误”,从而提高模型的预测精度.在20newsgroups数据集上的实验结果表明:所提出的方法能够提高半监督分类的精度.
Abstract:
The progressively semi-supervised classification model based on Markov random walk,in the random walk process has been proposed,and calculated the migration probability of samples to be marked,considering only samples of the appropriate category,while ignoring the other classes of samples; and then combined the progressive learning with semi-supervised learning.The model can improve the precision by "correcting" the errors caused in semi-supervised learning process.The results on 20newsgroups dataset in the experiment shows that the proposed method can improve the accuracy of semi-supervised classification.

参考文献/References:

[1] Zhu Xiaojin.Semi-supervised learning literature survey [ R/OL ].
[2013-03-19].http:∥www.loni.ucla.edu/~ztu/courses/2013_CS_spring/reading/ssl_survey.pdf.
[2] Zhou Zhihua,Zhan D C,Yang Q.Semi - supervised learning with very few labeled training examples [ C/OL].
[2013-03-21].Semi - supervised learning with very few labeled training Examples.
[3] 易星.半监督学习若干问题的研究 [D].北京:清华大学,2004.
[4] 董乐红,耿国华,高原.Boosting算法综述 [J].计算机应用与软件,2006,23(8):27-29.
[5] Liu Bin,Lee W S,YU P S,et al.Partially supervised classification of text documents [C/OL].
[2013-04-11].http:∥www.cs.uic.edu/~liub/S-EM/unlabelled.pdf.
[6] 郑海清,林琛,牛军钰.一种基于紧密度的半监督文本分类方法 [J].中文信息学报,2007,21(3):54-60.
[7] Bluma,Mitchell T.Combining labeled and unlabeled data with co-training [C/OL].
[2]13-04-14].http:∥dl.acm.org/citation.cfm?id=279962.
[8] Szummer M,Jaakkola T.Partially labeled classification with Markov random walks [J].Advances in Neural Information Processing Systems,2001,14(1):1-8.
[9] Arik Azran.The rendezvous algorithm:multi-class semi-supervised learning with Markov random walks [C/OL].
[2013-05-12].http:∥citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.74.2110.
[10] Robert E.Schapire,Yoram S.BoosTexter:a boosting- based system for text categorization [J].Machine Learning,2000,39(2/3):135-168.
[11] 任巨伟,杨亮,林鸿飞.情感图式构造及其在文本情感计算中的应用 [J].江西师范大学学报:自然科学版,2013,37(2):130-135.
[12]何文译,林鸿飞,杨亮.基于群体智慧的电影排序模型 [J].江西师范大学学报:自然科学版,2013,37(2):136-141.

备注/Memo

备注/Memo:
国家自然科学基金(60963014)
更新日期/Last Update: 1900-01-01