摘要
对抓取的苏州园林网络评论进行意见挖掘,基于词、词性和句法模式利用CRF模型抽取评论句中的评价对象,利用SVM分类算法对评论句中蕴含的情感进行分类,获得了较为理想的实验结果,表明所采用的算法较为有效,有一定的实用价值。进一步,基于评论意见挖掘结果,对苏州园林中的5个典型园林主要从整体用户情感倾向判断和寻找园林特质这两方面进行分析探讨,揭示意见挖掘的重要应用价值。
This paper mines the opinions of Suzhou garden crawled from the network.CRF model is used to extract the features based on the word,part of speech and syntactic pattern.The SVM algorithm is used to execute the sentiment classification.Experiment result is good,showing the algorithm＇s effectiveness and its practical value.Furthermore,from aspects of users sentiment analysis and gardens＇ character finding,5 typical ones from all Suzhou gardens are discussed based on the result of Suzhou garden opinion mining,which reveals the important application value of opinion mining.
出处
《微型机与应用》
2016年第13期86-89,共4页
Microcomputer & Its Applications
基金
国家社会科学基金项目(11CYY031)
关键词
意见挖掘
特征提取
情感分类
条件随机场
支持向量机
opinion mining
features extraction
sentiment classification
conditional random fields
support vector machines
