CAJ | 학술논문

为了分析中文微博中海量的情感信息，文中提出了一种中文微博情感分析策略，能够有效分析出微博中的情感倾向。为了能准确分析出某领域微博情感倾向，本文构建了领域情感词典，具有自动识别、扩展等功能，减少了人工标注的繁琐。同时考虑到上下文中情感副词等影响，构建了情感副词词典，更加全面的分析情感倾向。最后通过实验表明本文提出的基于领域情感词典的分析策略有一定的可行性和准确率。
위료분석중문미박중해량적정감신식，문중제출료일충중문미박정감분석책략，능구유효분석출미박중적정감경향。위료능준학분석출모영역미박정감경향，본문구건료영역정감사전，구유자동식별、확전등공능，감소료인공표주적번쇄。동시고필도상하문중정감부사등영향，구건료정감부사사전，경가전면적분석정감경향。최후통과실험표명본문제출적기우영역정감사전적분석책략유일정적가행성화준학솔。
In order to analyze the massive emotional information in Chinese micro-blog, this article proposes a Chinese micro-blog sentiment analysis strategy, and it can analyze the emotional tendencies effectively in the micro-blog. In order to analyze the emotional tendencies of a field more accurately, this paper builds a kind of emotional dictionary through mood words with automatic identification, extended function, reduces cumbersome manual annotation. Considering the influence of emotional adverbs and expressions in the context, we build the emotional adverbs dictionary and micro-blog expressions which will analyze emotion tendency more comprehensively. In order to analyze the emotional tendencies more comprehensive, and solve the problem of more and more internet words, it constructs network vocabulary classifier to find and mark internet words. Finally, experiments show that the proposed analyze strategy which based on emotional dictionary has certain feasibility and accuracy.