SVM+BiHMM基于统计方法元数据抽取混合模型.pdfVIP

ISSN 1000-9825, CODEN RUXUEW E-mail: jos@ Journal of Software , Vol.19, No.2, February 2008, pp.358−368 DOI: 10.3724/SP.J.1001.2008.00358 Tel/Fax: +86-10 © 2008 by Journal of Software . All rights reserved. ∗ SVM+BiHMM:基于统计方法的元数据抽取混合模型 + 张 铭 , 银 平, 邓志鸿, 杨冬青 (北京大学 信息科学技术学院,北京 100871) SVM+BiHMM: A Hybrid Statistic Model for Metadata Extraction + ZHANG Ming , YIN Ping, DENG Zhi-Hong, YANG Dong-Qing (School of Electronics Engineering and Computer Science, Peking University, Beijing 100871, China) + Corresponding author: Phn: +86-10 Fax: +86-10 E-mail: mzhang@, /mzhang Zhang M, Yin P, Deng ZH, Yang DQ. SVM+BiHMM: A hybrid statistic model for metadata extraction. Journal of Software, 2008,19(2):358−368. /1000-9825/19/358.htm Abstract : This paper proposes SVM+BiHMM, a hybrid statistic model of metadata extraction based on SVM (support vector machine) and BiHMM (bigram HMM (hidden Markov model)). The BiHMM model modifies the HMM model with both Bigram sequential relation and position information of words, by means of distinguishing the beginning emitting probability from the inner emitting probability. First, the rule based extractor segments documents into line-blocks. Second, the SVM classifier tags the blocks into metadata elements. Finally, the SVM+BiHMM model is built based on the BiHMM model, with the emitting probability adjusted by the Sigmoid function of SVM score, and the transition probability trained by Bigram HMM. The SVM classifier benefits from the structure patterns