Dead Link Prediction Model Based on Double SVM

DOI

10.2991/meic-15.2015.60How to use a DOI?

Keywords

dead link; Support Vector Machine; the effective attributes;Dead Link Prediction Model;the independent model

Abstract

All the search engines are facing the problem of dead link. Dead link prediction model, which is able to quickly distinguish between normal links and dead links, filter out dead links, ensure the validity of the search results. This paper proposed to construct the SVM dead link prediction model by the effective attributes of the links because the fast update of the links in engine library, this model can quickly identify dead links. Due to the attributes related to the types of web site, independent training prediction model was proposed according to the different types of web site to improve the precision rate, experiments proved that the precision and recall rate of the independent model higher than the uniform model. It is unrealistic that training samples marked completely rely on manual because the independent training samples’ amount is large, so using SVM dead link prediction model based on web content to prepare the sample because it’s high accuracy, and applicable to all links. It constituted a dead link prediction model based on double SVM by using SVM dead link prediction model twice, which greatly improved the precision rate of dead link prediction, and reduces the prediction time.

Copyright

© 2015, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Honglan Liu
AU  - Xiaona Qin
AU  - Yong He
PY  - 2015/04
DA  - 2015/04
TI  - Dead Link Prediction Model Based on Double SVM
BT  - Proceedings of the 2015 International Conference on Mechatronics, Electronic, Industrial and Control Engineering
PB  - Atlantis Press
SP  - 252
EP  - 256
SN  - 2352-5401
UR  - https://doi.org/10.2991/meic-15.2015.60
DO  - 10.2991/meic-15.2015.60
ID  - Liu2015/04
ER  -