基于互信息和众数的类间关系研究

J4 ›› 2014, Vol. 36 ›› Issue (10): 2019-2027.

基于互信息和众数的类间关系研究

陈峰，张伟娟，靳小波

（河南工业大学信息科学与工程学院，河南郑州 450001）

收稿日期:2013-03-11 修回日期:2013-06-20 出版日期:2014-10-25 发布日期:2014-10-25
基金资助:
国家自然科学基金资助项目（61203265,61103138）；河南省重点项目（122102110106）；河南工业大学博士基金资助项目（150121）

Mining inter-class relationship based on
mutual information and mode pattern

CHEN Feng,ZHANG Weijuan,JIN Xiaobo

(College of Information Science and Engineering,Henan University of Technology,Zhengzhou 450001,China)

Received:2013-03-11 Revised:2013-06-20 Online:2014-10-25 Published:2014-10-25

摘要/Abstract

摘要：

簇间关系的评估对于确定许多现实中的关键未知信息具有重要作用，可被广泛应用在犯罪侦查、进化树、冶金工业和生物嫁接等领域。提出了一种称为“众数模式+互信息”的方法对簇间关系进行排序。该方法使用众数模式从每个簇中寻找具有代表性的对象，使用互信息衡量簇间的关联程度。由于该方法利用簇间关系判断簇与簇的联系程度，所以它不同于传统的分类和聚类。在图形和癌症诊断数据上的实验表明了该算法的有效性。

关键词: 关系树, 概念簇, 众数模式, 互信息

Abstract:

The evaluation of the relationships between conceptual clusters is important to identify vital unknown information in many reallife applications, such as crime detection, evolution trees, metallurgical industry, biology engraftment and so forth. A method called "mode pattern plus mutual information" is proposed to rank the interrelationship between clusters. Mode pattern is used to find the outstanding objects from each cluster, while mutual information criterion measures the close proximity of cluster pairs. The proposed method is different from conventional algorithms of classifying and clustering due to focusing on ranking the interrelationship between clusters.Experiments are carried out on a wide range of reallife datasets, including image data and cancer diagnosis data.The experimental results demonstrate the effectiveness of the proposed algorithm.

Key words: relationship tree;conceptual clusters;mode pattern;mutual information

陈峰，张伟娟，靳小波. 基于互信息和众数的类间关系研究[J]. J4, 2014, 36(10): 2019-2027.

CHEN Feng,ZHANG Weijuan,JIN Xiaobo. Mining inter-class relationship based on
mutual information and mode pattern [J]. J4, 2014, 36(10): 2019-2027.

参考文献［16］

［1］	Gordon A D.‘Hierarchical Classification’ in clustering and classification［M］. NJ:World Scientific, 1996.
［2］	Mohebi E,Hassan S.Clustering of high dimensional data based on an optimized kohonen maps［C］∥Proc of 2011 the 4th IEEE International Conference on Computer Science and Information Technology(ICCSIT 2011),2011:201208.
［3］	Zhou Yantao,Tang Jianbo,Wang Jiaqin.Improved TFIDF feature selection algorithm based on information entropy［J］. Computer Engineering and Applications, 2007, 43(35):156158.(in Chinese)
［4］	Zhang Shichao,Chen Feng,Wu Xindong,et al. Identifying bridging rules between conceptual clusters［C］∥Proc of the 12th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining(KDD’06), 2006:815820.
［5］	Kumar D, Ramakrishnan N, Helm R F, et al. Algorithms for storytelling［C］∥Proc of the 12th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD’06), 2006:604610.
［6］	Huda A H, Martone M E, David C. Challenges and opportunities in mining neuroscience data［J］. Science,2011,331(6018):708712.
［7］	Alhammady H,Ramamohanarao K.Using emerging patterns to construct weighted decision trees［J］. IEEE Transactions on Knowledge and Data Engineering, 2006,18(7):865876.
［8］	Leon S J. Linear algebra:With application［M］. New Jersey:Pearson Prentice Hall, 2006.
［9］	Zhu H, Li L, Zhou H. Nonlinear dimension reduction with WrightFisher kernel for genotype aggregation and association mapping［J］.Bioinformatics,2012, 28(18):i375i381.
［10］	Higgins J P, Thompson S G. Quantifying heterogeneity in a metaanalysis ［J］.Statistics in Medicine, 2002,21(11):15391558.
［11］	Yiu M, Mamoulis N. Reverse nearest neighbors search in ad_hoc subspaces［J］. IEEE Transactions on Knowledge and Data Engineering, 2007,19(3):412426.
［12］	Yan X,Zhou X,Han J.Mining closed relational graphs with connectivity constraints［C］∥Proc of the 21st International Conference on Data Engineering (ICDE’05), 2005:357358.
［13］	Connor M, Kumar P. Fast construction of knearest neighbor graphs for point clouds［J］. IEEE Transactions on Visualization and Computer Graphics, 2012,16(4):599608.
［14］	Cilibrasi R,Vitanyi P.Clustering by compression［J］. IEEE Transactions on Information Theory, 2005,51(4):15231545.
［15］	Damani S, Bacconi A, Libiger O. Characterization of circulating endothelial cells in acute myocardial infarction［J］. Science, 2012,4(126):126133.
［16］	Yu Feng,Yang Chengwu.σentropy,σdistance measure and σsimilarity measure of intuitionistic intervalvalued fuzzy sets［J］. Computer Engineering and Applications,2008,44(9):7072.(in Chinese)
［17］	Fizzotti M, Chen E Y, Link M P. Simultaneous expression of RBTN2 and BCRABL oncogenes in a TALL with a t(11;14)(p13;q11) and a lateappearing Philadelphia chromosome［J］. Leukemia ,1994,8(7):11241130.
［18］	Gurbuxani S, Lacorte J M, Raina V. Detection of BCRABL transcripts in acute lymphoblastic leukemia in Indian patients［J］. Leukemia Research ,1998, 22(1):7780.
	附中文参考文献：
［3］	周炎涛,唐剑波,王家琴.基于信息熵的改进TFIDF特征选择算法［J］. 计算机工程与应用, 2007, 43(35):156158.
［16］	俞峰，杨成梧.直觉区间值模糊集的σ熵、σ距离测度和σ相似测度［J］.计算机工程与应用，2008,44(9):7072.

[1]	李雨晨, 魏巍, 白伟明, 王达. 基于标签共现关系的多标签特征选择[J]. 计算机工程与科学, 2021, 43(11): 2049-2055.
[2]	王鑫, 陈建凯, 翟俊海, . 区间值属性单调决策树算法的扩展[J]. 计算机工程与科学, 2020, 42(03): 557-563.
[3]	孙燕, 孙茂松, 赵海兴, 冶忠林, . 表示学习中句子与随机游走序列等价性的一种新证明[J]. 计算机工程与科学, 2020, 42(02): 373-380.
[4]	朱天宇，全惠敏，刘国才. 联合弯曲能量和标志点对应约束的非刚性医学图像配准[J]. 计算机工程与科学, 2019, 41(05): 851-857.
[5]	张育培，刘树慧. 基于特征漂移的数据流集成分类方法[J]. J4, 2014, 36(05): 977-985.
[6]	涂伟1，甘丽新2，黄乐辉1，谢志华2. 基于改进互信息的信息检索扩展模型[J]. J4, 2013, 35(3): 150-154.
[7]	王小伟1, 郭红涛2, 王中锋3. 抗干扰词攻击的免疫垃圾邮件过滤模型[J]. J4, 2013, 35(12): 173-177.
[8]	李玲1，刘华文1,2，马宗杰1，赵建民1. 基于特征选择的集成多标签分类算法[J]. J4, 2013, 35(10): 137-143.
[9]	冯雪芳，吴锡生. 基于归一化边缘互信息与自适应加速粒子群的图像配准方法[J]. J4, 2013, 35(1): 119-123.
[10]	刘君1,2，武和雷3. 基于形状和灰度信息的断层图像二次配准算法[J]. J4, 2012, 34(12): 110-114.
[11]	王柯朱启兵. 一种基于差别矩阵的启发式属性约简算法[J]. J4, 2008, 30(6): 73-75.
[12]	马政德杜云飞周海芳刘衡竹. 遥感图像配准中相似性测度的比较和分析[J]. J4, 2008, 30(2): 45-48.
[13]	卢振泰张明慧陈武凡. 基于FCM算法与互信息量的图像自动分割[J]. J4, 2007, 29(6): 36-38.
[14]	冯国柱[1] 李超[1,2] 吴翊[1]. RSA公钥密码的信息论分析[J]. J4, 2007, 29(2): 4-6.