| 面向元信息分类的支持向量机改进技术 |
| |
| 引用本文： | 丁军平,蔡皖东.面向元信息分类的支持向量机改进技术[J].西安交通大学学报,2011,45(8):37-42. |
| |
| 作者姓名： | 丁军平 蔡皖东 |
| |
| 作者单位： | 西北工业大学计算机学院,710072,西安 |
| |
| 基金项目： | 国家高技术研究发展计划资助项目(2009AA01Z424) |
| |
| 摘 要： | 针对传统元信息分类方法的准确率不能满足主动P2P网络监测模型要求的问题,提出了一种基于改进支持向量机算法的元信息分类方法.该方法首先通过在加权最小二乘支持向量机的基础上加入对数据偏斜的处理,解决了元信息分类时关键词特征稀疏和样本高度不均衡问题,在对元信息文件名进行分词时,加入了词条之间的组合关系处理,在进行特征向量表示时,加入了对词条权值和语义属性的处理,最后使用基于粗糙集的属性规约方法进行特征向量选择,有效地降低了特征向量维度.实验结果表明,与传统方法相比,所提方法在进行元信息分类时能够大幅度提高分类准确率,准确率可达到97.8%,完全能够满足主动P2P网络监测模型的要求.
|
| 关 键 词： | 元信息分类 支持向量机 特征向量表示 粗糙集 |
An Improved Support Vector Machine Technology for Meta-Information Classification |
| |
| DING Junping,CAI Wandong.An Improved Support Vector Machine Technology for Meta-Information Classification[J].Journal of Xi'an Jiaotong University,2011,45(8):37-42. |
| |
| Authors: | DING Junping CAI Wandong |
| |
| Institution: | (School of Computer Science and Engineering,Northwestern Polytechnical University,Xi′an 710072,China) |
| |
| Abstract: | An improved support vector machine(SVM) algorithm for meta-information classification is proposed to solve the problem that the traditional classification algorithm on meta-information can't meet the requirement of initiative P2P network monitoring model.The algorithm solves the problems that the |
| |
| Keywords: | in classification are sparse and the distribution of sample is unbalanced by adding processing the data skew on LS-VSM Combination relationships among Key words are added on filename segmentation of the meta-information The weights of Key words and semantic attribute are processed for feature vector representation and the feature vector is selected using rough set specification so that the dimension of the feature vector is effectively reduced Experimental results and comparisons with the traditional algori |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
