网络购物行业的发展日益成熟，越来越多的消费者在购物网站发布评论信息。产品评论反映了消费者对产品的态度和意见，很有实用价值。一方面，产品评论可以影响其他消费者的购买意向；另一方面，产品评论反馈给商家产品信息，方便改进产品和提高服务。但是，想要短时间内从大量的产品评论文本数据中得到有意义的信息，是非常困难的。研究中文评论意见挖掘方法，对于提高文本信息提取效率有重要意义。本文主要研究内容如下： 以在线中文评论为研究对象，首先采用人工的方式，将主观性评论文本从产品评论中分离出来。然后利用自然语言处理技术，对中文在线评论意见文本进行预处理。 针对已有的意见挖据工作对中文产品评论信息挖掘的低查全率和低查准率问题，文中提出了一种改进的中文在线评论意见挖掘算法。该方法根据自然语言表达方式，将评论文本分为四类句式结构。然后基于副词抽取各类评论文本中的产品意见词和特征词。实验结果表明，该方法能够有效提高中文产品评论意见挖掘的查全率和查准率。 最后基于《同义词词林扩展版》，将抽取出的产品特征和意见词中的同义词合并，再根据支持度阈值剪枝，得到最终的特征词和意见词。实验表明，该方法提高了同义词合并的准确性。 本文改进的中文产品评论意见挖掘方法，可充分利用自然语言表达特点，实现对中文产品评论特征和意见词自动抽取。

The development of the online shopping industry is becoming more and more mature, and more and more consumers are making comments on the shopping site.Product reviews reflect the consumers' attitudes and views on the products. They have a very practical value.On the one hand, product reviews can affect other consumers' buying intentions;On the other hand, product reviews feedback information to businesses to facilitate improvements in products and services.However, it is very difficult to get meaningful information from a large number of texts data of product reviews in a short time.It is very important to study the extraction methods of Chinese comments mining to improve the efficiency of texts information extraction.The main contents are as follows: Take the Chinese online commentary texts as the main research object.Firstly, the subjective comment texts are separated from the product reviews by artificial way.And then use the natural language processing technology to deal with Chinese online comments. In this paper, an improved Chinese online comment mining algorithm is proposed to solve the problem of low recall and low precision in Chinese online product reviews.The method divides the comment text into four kinds of sentence structure according to the expression of natural language.And then extract the product opinion words and characteristic words in the various comment texts based on the adverbs.The experimental results show that this method can effectively improve the recall rate and Precision rate of Chinese product reviews. Finally, based on“extended version of word forest synonyms” to merge synonyms for the characteristics of the product review texts.And get the final characteristic words and opinions words. Experiments show that this method improves the validity of synonym merges. In this paper, the Chinese product comments mining method based on the adverbs, can make full use of natural language expression characteristics, and automatically extract the characteristics and opinions of Chinese product reviews.