资源描述：
《自然语言处理_cu mi lab - personage dataset(剑桥大学机器智能实验室人物数据集)》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。
1、CUMILab-PERSONAGEDataset(剑桥大学机器智能实验室人物数据集)数据摘要：ThePERSONAGEdatasetcontainsatotalof580utterancesannotatedwithpersonalityratingsfromhumanjudges.TheratingswereobtainedbyusingtheTen-ItemPersonalityInventory(Goslingetal.,2003)toassessthepersonalityofanhypotheticalspeakerforeachutte
2、rance(showninawrittenform),providingratingsareonascalefrom1(low)to7(high)foreachoftheBigFivepersonalitytrait260utterancesonlyhaveextraversionratings).TheutterancesweregeneratedusingthePERSONAGEgenerator,thedatathusalsoincludesalistingofthegenerationdecisionsbeingmadetoproducet
3、heutterance,aswellastheintermediarycontentplantree,sentenceplantreeandfinalsyntacticstructureforeachutterance.ThisdatawasusedforevaluatingthestyleconveyedbythePERSONAGE-RBrule-basedgenerator,aswellasfortrainingtherankingmodelsofthePERSONAGE-OSgeneratorandtheparameterestimation
4、modelsofthePERSONAGE-PEdata-drivengenerator.中文关键词：人物,话语,等级,发生器,生成方法,评估,英文关键词：PERSONAGE,Utterances,Ratings,Generator,Generationmethods,Evaluating,数据格式：TEXT数据用途：ThisdatawasusedforevaluatingthestyleconveyedbythePERSONAGE-RBrule-basedgenerator,aswellasfortrainingtherankingmodelsof
5、thePERSONAGE-OSgeneratorandtheparameterestimationmodelsofthePERSONAGE-PEdata-drivengenerator.数据详细介绍：CUMILab-PERSONAGEDatasetThePERSONAGEdatasetcontainsatotalof580utterancesannotatedwithpersonalityratingsfromhumanjudges.TheratingswereobtainedbyusingtheTen-ItemPersonalityInvento
6、ry(Goslingetal.,2003)toassessthepersonalityofanhypotheticalspeakerforeachutterance(showninawrittenform),providingratingsareonascalefrom1(low)to7(high)foreachoftheBigFivepersonalitytrait260utterancesonlyhaveextraversionratings).TheutterancesweregeneratedusingthePERSONAGEgenerat
7、or,thedatathusalsoincludesalistingofthegenerationdecisionsbeingmadetoproducetheutterance,aswellastheintermediarycontentplantree,sentenceplantreeandfinalsyntacticstructureforeachutterance.ThisdatawasusedforevaluatingthestyleconveyedbythePERSONAGE-RBrule-basedgenerator,aswellasf
8、ortrainingtherankingmodelsofthePERSONAGE-OSgeneratorandthepar
