基于网络用户情感分析的预测方法研究

Research on Predicting Methods Based on Network User Sentiment Analysis

网络用户情感分析领域的研究为特定领域社会行为的预测提供了新的方法和工具。本文分析了基于情感分析进行预测的逻辑基础、典型预测方法、关键技术以及当前存在的问题和发展趋势。研究发现:研究基于网络用户情感分析预测社会活动趋势的方法在政治、财经等多个领域具备应用条件;典型预测方法可归纳为以情感分析结果作为辅助依据的预测方法和以情感分析结果作为主要依据的预测方法;预测过程涉及情感分析源的选择、预测时间提前量的确定以及情感词统计处理三个关键环节;当前研究还存在网络用户情感的代表性,待分析语料的全面和正确获取,以及网络用户情感的正确分析和统计等问题,有待深入研究。

The research on field of network user sentiment analysis provides new methods and tools for predicting social ac- tivities in special domain. The paper analyses the logical basis, typical methods, key technologies, problems and development tendency of predicting methods based on sentiment analysis. The paper concludes that the predicting methods can be used in many domains, such as palitics and finance; the typical methods include predicting methods that take sentiment analysis as aux- iliary mad predicting methods that take sentiment analysis as primary; the predicting process involves three key points which are the choice of sentiment analysis resources, the determination of lead time and the compute of sentiment words; current researches also have some problems that need to be researched further, which include the representativeness of network users sentiment, the comprehensive and correct acquisition of corpus, and the correct analysis and statistics of network users sentiment. 2 figs. 47 refs.