摘要
评论挖掘是近年来自然语言处理领域研究的热点,为此重点研究评论挖掘技术中的评论特征词提取,采用基于语法模式的抽取方法从中抽取特征词,并采用相似度计算方法去除冗余特征词,用词频统计方法去除低频词,最终生成特征词集。经实验证明,该方法能取得较好的准确率和召回率。
Opinion mining has been a hotspot in the field Natural Language Processing in recent years. This paper focused on the research of extracting feature words in opinion mining, extracted feature words based on the extraction method of grammatical patterns, removed redundant feature words using the similarity calculation method, removed the low-frequency words using word frequency statistics method, and finally generated the feature words set. Proved by the experiments, this method can achieve better recall and precision.
出处
《广东水利电力职业技术学院学报》
2014年第4期24-26,63,共4页
Journal of Guangdong Polytechnic of Water Resources and Electric Engineering
关键词
评论挖掘
语法模式
特征词提取
opinion mining
syntax pattern
feature words extraction
