CCF-ADLIRTheCCFAdvancedDisciplinesLectures.pptVIP

哈工大社会计算与信息检索研究中心 哈工大社会计算与信息检索研究中心 哈工大社会计算与信息检索研究中心 Learning Sentence Representation for Emotion Classification on Microblogs Duyu Tang, Bing Qin, Ting Liu, Zhenghua Li HIT-SCIR * Outline Motivation Representation Learning for Emotion Classification Experiment Results Conclusion and Future Work * Emotion Analysis on Microblog Microblog, such as Weibo, has become a popular platform to share opinions about products or hot events. * Task Emotion Classification Classify a text as Happy, Sad, Angry or Surprise Prior Work Supervised [Pang2002, Mishne2006] Distant Supervision [Go2009, Liu2012] Unsupervised [Turney2002, Ding2008] * Outline Motivation Representation Learning for Emotion Classification Experiment Results Conclusion and Future Work * Representation Learning Traditional The bag-of-word representation can not capture the complex linguistic phenomena. Manufacturing feature engines is time-consuming. Learning based method Deep Belief Networks Text with emoticons * Deep Belief Networks Architecture * Data Collection Emotions from Weibo * Input Layer Top frequent unigrams Punctuation Lexicons Emotion word Onomatopoeia word Function word * Outline Motivation Representation Learning for Emotion Classification Experiment Results Conclusion and Future Work * Dataset Emoticon data 20,000 weibo (balanced for each category) Labeled data * Emotion Happy Sad Angry Surprise Size 548 837 905 567 Results * Results * Outline Motivation Representation Learning for Emotion Classification Experiment Results Conclusion and Future Work * Conclusion Emotion classification on Weibo Learning sentence representation on emoticon data via deep belief network For future Word representation for sentiment analysis Compositionality for sentiment analysis * Questions? * Back Up * Emoticon Selection * Detailed Params Lexicons Architecture * * 微博现已成为大众接受信息和交流情感的主要平台，比如iphone5s上市，大家会从不同角度发表个人的褒贬观点。 当现实生活中发生了某些热门事件，大家也会将自己的情绪反映在一条条微博当中。 比如右侧哈尔滨雾霾的图片，哈尔滨人很愤怒，也可能很恐惧会对身体有损害。 我们需要自动的