摘要: 以科技论文摘要句子为研究对象,提出一种两阶段的细粒度句子分类方法,通过结合摘要内各个句子的位置、关键词和上下文信息,选择部分易于分辨语境类型的句子,将其作为种子样本训练获得分类模型。利用机器学习的方法对摘要句子的背景知识、论文主题、研究方法和实验结果进行自动分类。实验结果表明,该方法中的F度量值比其他细粒度分类方法平均高3%~5%。
关键词:
细粒度,
语境,
摘要句子,
句子分类,
种子样本,
机器学习
Abstract: This paper discusses fine-grained sentence classification method with sentences from abstracts of scientific papers as its subject. This method includes two steps. It selects sentences with distinct types of context as a sample to practice so as to obtain a classification model. This is done on the base of analyzing the position, key words and context of these sentences. It classifies the rest sentences with machine learning approach to achieve automatic classification of background knowledge, paper theme, research method and experimental result. F value of the method is 3%~5% higher than that of other fine-grained methods.
Key words:
fine-grained,
context,
abstract sentence,
sentence classification,
seed sample,
machine learning
中图分类号:
华秀丽, 徐凡, 王中卿, 李培峰. 细粒度科技论文摘要句子分类方法[J]. 计算机工程, 2012, 38(14): 138-140.
HUA Xiu-Li, XU Fan, WANG Zhong-Qing, LI Pei-Feng. Fine-grained Classification Method for Abstract Sentence of Scientific Paper[J]. Computer Engineering, 2012, 38(14): 138-140.