?url_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Adc&rft.title=Solving+Multiple-Instance+Problem%3A+A+Lazy+Learning+Approach&rft.creator=Wang%2C+Jun&rft.creator=Zucker%2C+Jean-Daniel&rft.subject=Artificial+Intelligence&rft.subject=Machine+Learning&rft.description=As+opposed+to+traditional+supervised+learning%2C+multiple-instance+learning+%0A++++concerns+the+problem+of+classifying+a+bag+of+instances%2C+given+bags+that+are+%0A++++labeled+by+a+teacher+as+being+overall+positive+or+negative.+Current+research+%0A++++mainly+concentrates+on+adapting+traditional+concept+learning+to+solve+this+%0A++++problem.+In+this+paper+we+investigate+the+use+of+lazy+learning+and+Hausdorff+%0A++++distance+to+approach+the+multiple-instance+problem.+We+present+two+variants+of+%0A++++the+K-nearest+neighbor+algorithm%2C+called+Bayesian-KNN+and+Citation-KNN%2C+solving+%0A++++the+multiple-instance+problem.+Experiments+on+the+Drug+discovery+benchmark+data+%0A++++show+that+both+algorithms+are+competitive+with+the+best+ones+conceived+in+the+%0A++++concept+learning+framework.+Further+work+includes+exploring+of+a+combination+of+%0A++++lazy+and+eager+multiple-instance+problem+classifiers.&rft.publisher=Morgan+Kaufmann&rft.contributor=Langley%2C+Pat&rft.date=2000&rft.type=Conference+Paper&rft.type=PeerReviewed&rft.format=application%2Fpdf&rft.identifier=http%3A%2F%2Fcogprints.org%2F2124%2F3%2Fwang_ICML2000.pdf&rft.identifier=++Wang%2C+Jun+and+Zucker%2C+Jean-Daniel++(2000)+Solving+Multiple-Instance+Problem%3A+A+Lazy+Learning+Approach.++%5BConference+Paper%5D+++++&rft.relation=http%3A%2F%2Fcogprints.org%2F2124%2F