欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
[目的/意义]面向电子商务领域的在线评论,通过识别产品特征和评价词之间是否存在修饰关系,抽取出在线评论中的情感标签,从而帮助网购用户迅速了解某一产品的性能。[方法/过程]引入汉语组块分析对评论文本进行初始化处理,对名词性信息以及形容词性信息进行抽取。通过最大熵对初始化集合进行过滤,从而获得最终的情感标签集合。[结果/结论]该方法对评论文本的适应性较好,可以有效抽取出情感标签。[局限]需要对语料进行初始化抽取,经过过滤后才能获得最终的情感标签集合。
英文摘要：
[ Purpose/significance] For online reviews in the fields of E-commerce, rely on the existence of a modified rela- tionship between the identification and evaluation of product features, this paper extracts emotional label of online reviews to help online shoppers to quick understand the performance of a product. [ Method/process ] Chinese chunking is being used to analyze the structure of online reviews, then, the paper extracts the nominal information and adjective information from the reviews. Using the maximum entropy model to filter the candidate set, in order to get the final emotional label sets. [ Result/conclusion ] This methods has better result on reviews, it can effectively extracts the emotional label. [ Limitations ] The paper needs to be initialized corpus extraction, which can obtain a final emotional label collection after filtration.
同期刊论文项目
同项目期刊论文
期刊信息
位置：