摘要
摘要
随着国民经济的持续快速发展,中国电影产业的发展环境不断优化和提升。电影票房的相关研究也得到了大量的关注。本文从情感分析的角度出发,研究电影用户评论的情感倾向与电影票房之间的相关性。论文研究主要关注了两个方面的问题,一是对比分析了电影用户的情感倾向的情感指标,二是研究了电影票房与电影用户评论的情感倾向之间的关系。研究结果表明,用户评分与SnowNLP库的情感分析功能提取豆瓣短评文本的情感概率值具有一致性,均可作为用户倾向指标,且用户评分判断用户真实情感倾向的准确性更高,更能体现用户的真实情感倾向。研究进一步引入多元线性回归和随机森林回归,深入地探讨了用户评论与电影票房的相关性,结果表明用户评论中的用户评分、评价人数对票房有正向影响作用。论文最后根据研究结果对电影票房的研究提出了合理的意见和建议。关键词票房研究;情感分析;在线评论;SnowNLP;相关分析
Abstract
Ⅱ
Abstract
FengdanWei
AppliedStatisticsDirectedbyJunLu
原创力文档
预览与源文档一致,下载高清无水印
Withthesustainedandrapiddevelopmentofthenationaleconomy,thedevelopmentenvironmentofChinasfilmindustryhasbeencontinuouslyoptimizedandimproved.Therelatedresearchonmovieboxofficehasalsoreceivedalotofattention.Fromtheperspectiveofsentimentanalysis,thisthesisstudiesthecorrelationbetweentheemotionaltendenciesofmovieuserreviewsandmoviebox-office.Theresearchofthisthesismainlyfocusesontwoaspects.Oneistocompareandanalyzetheemotionalindicatorsofmovieusersemotionaltendencies,andtheotheristostudytherelationshipbetweenthemoviebox-officeandtheemotionaltendenciesofmovieuserreviews.TheresearchresultsshowthattheuserratingsareconsistentwiththesentimentprobabilityvalueoftheDoubanshortcommenttextextractedbythesentimentanalysisfunctionoftheSnowNLPlibrary,andtheycanbothbeusedasusertendencyindicators.Moreover,theuserratingismoreaccurateinjudgingtheuserstrueemotionaltendency,andcanbetterreflecttheuserstrueemotionaltendency.Theresearchfurtherintroducesmultiplelinearregressionandrandomforestregression,anddeeplyexploresthecorrelationbetweenuserreviewsandmoviebox-office.Attheendofthethesis,accordingtotheresearchresults,thepaperputsforwardreasonableopinionsandsuggestionsfortheresearchofthemovieboxoffice.
Key