Meta-information obtaining research based on SVM Initiative learning algorithm

摘要：

Aiming at the crawler technique taken in BitTorrent Initiative monitoring model, when obtaining the torrent file information, if you didnt distinguish the link address of the webpage and torrent file, but to unify access to technique, efficiency and algorithm will encounter great difficult. Based on this requirement, this thesis improved the traditional SVM classifying algorithm; and adopted PLS-SVM to classify the link address to obtain the torrent file. When using the algorithm, firstly collect samples information, eigenvector discretize and normalize the sample information. When you get the samples eigenvector, the algorithm study from samples to get the algorithm key parameters, and put it into practical application. This method be proved that it can enhance the efficiency and speed of obtaining the torrent file.

关键词： LS-SVM,Network Crawler Meta-information P2P Monitoring

作者: Ding Junping Cai Wandong

作者单位: School of Computer Science and Engineering Northwestern Polytechnical University Xian, China

会议类型: 国际会议

会议名称: 2010 International Conference on Circuit and Signal Processing(2010年电路与信号处理国际会议 ICCSP 2010)

会议地点: 上海

会议语种:英文

页码: 289-292

在线出版日期: 2010-12-25（万方平台首次上网日期，不代表论文的发表时间）