?url_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Adc&rft.title=Unsupervised+Learning+of+Semantic+Orientation+from+a+Hundred-Billion-Word+Corpus&rft.creator=Turney%2C+Peter+D.&rft.creator=Littman%2C+Michael+L.&rft.subject=Artificial+Intelligence&rft.subject=Language&rft.subject=Machine+Learning&rft.subject=Statistical+Models&rft.description=The+evaluative+character+of+a+word+is+called+its+semantic+orientation.+A+positive+semantic+orientation+implies+desirability+(e.g.%2C+%22honest%22%2C+%22intrepid%22)+and+a+negative+semantic+orientation+implies+undesirability+(e.g.%2C+%22disturbing%22%2C+%22superfluous%22).+This+paper+introduces+a+simple+algorithm+for+unsupervised+learning+of+semantic+orientation+from+extremely+large+corpora.+The+method+involves+issuing+queries+to+a+Web+search+engine+and+using+pointwise+mutual+information+to+analyse+the+results.+The+algorithm+is+empirically+evaluated+using+a+training+corpus+of+approximately+one+hundred+billion+words+%C2%97+the+subset+of+the+Web+that+is+indexed+by+the+chosen+search+engine.+Tested+with+3%2C596+words+(1%2C614+positive+and+1%2C982+negative)%2C+the+algorithm+attains+an+accuracy+of+80%25.+The+3%2C596+test+words+include+adjectives%2C+adverbs%2C+nouns%2C+and+verbs.+The+accuracy+is+comparable+with+the+results+achieved+by+Hatzivassiloglou+and+McKeown+(1997)%2C+using+a+complex+four-stage+supervised+learning+algorithm+that+is+restricted+to+determining+the+semantic+orientation+of+adjectives.+&rft.date=2002&rft.type=Departmental+Technical+Report&rft.type=NonPeerReviewed&rft.format=application%2Fpostscript&rft.identifier=http%3A%2F%2Fcogprints.org%2F2322%2F1%2FERB-1094.ps&rft.format=application%2Fpdf&rft.identifier=http%3A%2F%2Fcogprints.org%2F2322%2F5%2FERB-1094.pdf&rft.identifier=++Turney%2C+Peter+D.+and+Littman%2C+Michael+L.++(2002)+Unsupervised+Learning+of+Semantic+Orientation+from+a+Hundred-Billion-Word+Corpus.++%5BDepartmental+Technical+Report%5D++++(Unpublished)++&rft.relation=http%3A%2F%2Fcogprints.org%2F2322%2F