Graph classification based on graph set reconstruction and graph kernel feature reduction
Graph classification based on graph set reconstruction and graph kernel feature reduction
复制标题
基于图集重构和图核特征约简的图分类
DOI：
10.1016/j.neucom.2018.03.029
复制
发表时间：
2018-06
期刊：
影响因子：
6
通讯作者：
Cao Jie
中科院分区：
文献类型：
--
作者：
Ma Tinghuai;Shao Wenye;Hao Yongsheng;Cao Jie
图作为一种结构化数据，被广泛用于对象之间复杂关系的建模，在生物信息学、网络入侵检测、社交网络等科学和工程领域得到了广泛的应用。为了预测未知图或理解不同类别之间的复杂结构，建立一种自动、高精度的图分类方法是非常必要的。核方法被认为是一种强大的图分类解决方案，它包括两个步骤，即图核映射和分类。然而，特征选择过程被忽略，那些判别能力较低的子结构导致分类精度下降。为了解决这一问题，我们提出了一种基于图集重构和图核特征约简的高效图分类算法。首先，去除最小判别频繁子图和部分非频繁子图重构原始图集；然后采用基于图核的判别分析方法对重构好的图集进行特征约简。最后介绍了图分类方法的整体框架，并给出了常用的分类器。在一系列生物信息学基准上的大量实验结果表明，与其他基于图核的分类方法相比，我们的图分类算法在预测方面有显著提高。
Graph, a kind of structured data, is widely used to model complex relationships among objects, and has been used in various of scientific and engineering fields, such as bioinformatics, network intrusion detection, social network, etc. Building an automatic and highly accurate classification method for graphs becomes quite necessary for predicting unknown graphs or understanding complex structures among different categories. The kernel method is regarded as a powerful solution to graph classification, which consists of two steps, namely, graph kernel mapping and classification. However, the feature selection process is ignored, and those sub-structures with low discriminative power result in classification accuracy decrease. In order to solve this problem, we propose an efficient graph classification algorithm based on graph set reconstruction and graph kernel feature reduction. First of all, the least discriminative frequent subgraphs and part of the infrequent subgraphs are removed to reconstruct the original graph set. Then we adopt the graph-kernel-based discriminant analysis method to perform feature reduction on the well-reconstructed graph set. At last, the whole framework of the graph classification method is introduced and any commonly used classifiers can be utilized. Extensive experimental results on a series of bioinformatics benchmarks show that our graph classification algorithm demonstrates a significant improvement of prediction comparing with other graph-kernel-based classification approaches.
登录
查看更多内容
DOI：
10.1109/dsaa.2015.7344816
发表时间：
2012-01
期刊：
2015 IEEE International Conference on Data Science and Advanced Analytics (DSAA)
影响因子：
--
作者：
Giuseppe Jurman;R. Visintainer;M. Filosi;S. Riccadonna;Cesare Furlanello
通讯作者：
Giuseppe Jurman;R. Visintainer;M. Filosi;S. Riccadonna;Cesare Furlanello
DOI：
10.1109/cvpr.2016.491
发表时间：
2016-06
期刊：
2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
影响因子：
--
作者：
E. Stumm;Christopher Mei;S. Lacroix;Juan I. Nieto;Marco Hutter;R. Siegwart
通讯作者：
E. Stumm;Christopher Mei;S. Lacroix;Juan I. Nieto;Marco Hutter;R. Siegwart
DOI：
10.1145/2556195.2556198
发表时间：
2014-02
期刊：
Proceedings of the 7th ACM international conference on Web search and data mining
影响因子：
--
作者：
U. Kang;L. Akoglu;Duen Horng Chau
通讯作者：
U. Kang;L. Akoglu;Duen Horng Chau
DOI：
10.1016/j.compmedimag.2016.04.004
发表时间：
2016-09-01
影响因子：
5.7
作者：
Du, Junqiang;Wang, Lipeng;Zhang, Daoqiang
通讯作者：
Zhang, Daoqiang
DOI：
10.1002/cpe.3437
发表时间：
2016-02
期刊：
Concurrency and Computation: Practice and Experience
影响因子：
--
作者：
Tinghuai Ma;Huan Rong;Changhong Ying;Yuan Tian;A. Al-Dhelaan;Mznah Al-Rodhaan
通讯作者：
Tinghuai Ma;Huan Rong;Changhong Ying;Yuan Tian;A. Al-Dhelaan;Mznah Al-Rodhaan