针对中文微博数据非结构化特点，文章从相关维、状态维、主题维、情绪维四个维度提出了一套标准的微博情绪挖掘方法，通过情绪词典生成、倾向性分析、发布者影响力测度、情绪指标构建等关键环节，能够有效地从海量微博数据中提炼用户的观点倾向性，形成量化的情绪指标。应用该方法体系对旅游市场进行实证分析，发现带有正向情绪的微博通过口碑效应对于下一期旅游客流量存在显著的积极影响，在传统时间序列模型中引入正向微博情绪指标能够显著提高预测精度；通过对负面微博数据进行分主题挖掘，能够分析不同主题下游客抱怨的原因，形成数据驱动的游园改进策略，提高旅游管理的精准性和效率。

In order to handle the unstructured characteristics of Chinese Weibo data, the paper formulated a series of standards implying four dimensions of correlation, state, topics and emotion. The authors effectively extracted and purified the opinions or tendentiousness from huge amounts of user's data within Weibo, by means of the four key links, which are the sentiment dictionary, analysis of subjectivity, measurement of publisher's influence, and the establishment of sentiment index. While using this index system to the tourism market for an empirical analysis, it indicates that positive-sentiment Weibo articles would give vigorous impact on the next group of tourists. The precision of prediction will be greatly updated when applied with our Weibo positive-sentiment index in traditional time series analysis. By detecting attitude types with negative-sentiment Weibo data, the authors also can analyze the causes of tourists complain under different topics. The accuracy and efficiency of tourism management would boost using the state-of-art strategy derived from the index.