随着电子商务产业的快速发展，购物网站逐渐积累了海量的产品评论。用户通过产品评论了解产品的口碑,商家通过产品评论调整市场决策。使用计算机分析、判断产品评论的情感倾向对商家和用户都具有重要意义。本文首先介绍了产品评论情感分析的研究背景和研究现状，介绍了产品评论情感分析的核心算法和主要流程，最后以美国著名购物网站Amazon.com的英文手机产品评论为测试数据，利用Java语言和数据挖掘工具收集产品评论并对评论的情感倾向进行了分类。

With the rapid development of electronic commerce, there are mass of product reviews generated at the shopping websites. Consumers realize products’ word of mouth and producers set the market policy by understanding these reviews. Analyzing and judging the sentiment orientations of product reviews with computers is of important significance for both consumers and businesses. At the beginning, the paper introduced the research background and research actuality of sentiment analysis for product reviews, and then it introduced the core algorithm and main steps of sentiment analysis. Finally, the research collected the smartphone product reviews at amazon.com with Java application and classified the reviews by their sentiment orientations.