本文選題：詞匯功能　切入點(diǎn)：語義出版　出處：《數(shù)字圖書館論壇》2017年08期 　論文類型：期刊論文

【摘要】：為提高學(xué)術(shù)文獻(xiàn)語義出版水平,既需要在寫作和出版模式方面進(jìn)行研究,也需要探索學(xué)術(shù)文本語義理解技術(shù),以實(shí)現(xiàn)對(duì)學(xué)術(shù)文獻(xiàn),特別是存量學(xué)術(shù)文獻(xiàn)的語義化處理。本文在學(xué)術(shù)文本詞匯功能分析框架基礎(chǔ)上,提出一種基于條件隨機(jī)場(chǎng)的學(xué)術(shù)文獻(xiàn)問題和方法識(shí)別模型,該模型使用詞法特征、句法特征、組塊特征等27個(gè)特征。實(shí)驗(yàn)表明,該方法具有優(yōu)于當(dāng)前最佳的識(shí)別效果。
[Abstract]:In order to improve the level of semantic publishing of academic documents, it is necessary not only to study the writing and publishing mode, but also to explore the semantic understanding technology of academic texts in order to realize the realization of academic literature. In this paper, based on the framework of lexical functional analysis of academic texts, a problem and method recognition model of academic literature based on conditional random field is proposed. The model uses lexical features and syntactic features. The experimental results show that the proposed method is superior to the current optimal recognition effect.
【作者單位】： 武漢大學(xué)信息管理學(xué)院;武漢大學(xué)信息檢索與知識(shí)挖掘研究所;
【基金】：中國博士后科學(xué)基金項(xiàng)目(編號(hào):2016M602371) 國家自然科學(xué)基金青年項(xiàng)目“基于深度語義挖掘的引文推薦多樣化研究”(編號(hào):71704137)資助
【分類號(hào)】：G230.7;G254
，

本文編號(hào)：1562242