3-依存句法分析

《3-依存句法分析》由会员上传分享，免费在线阅读，更多相关内容在行业资料-天天文库。

1、ê6í‚ð•!‹
•XåÕ•Y‡‡¡—:Ñf€/v@¬'f2012t919å,þþ;uwww.icst.pku.edu.cn/lcwm/course/sckr2012/•Iê6íì†ê6í„”(áo½Ö:hûÑÅ•¹˜‡,X•,þ„;•…¹ÍÕ•IíÍ
Í'èåÕ•ííÓ„•
•X•íI•íIÒrè2of60ÍÕ•ŸËåPfßc(æÆå‹EŸàÍÓœfß/c(/æÆ/å/‹E/ŸàÍ'èÓœfß/NNc(/ADæÆ/ADå/VV‹E/NNŸà/NN3of60åÕ•í

2、íÓ„•ííÓ„IPNPVPNNADVPADVPVPfßADADVVNPc(æÆåNNNN‹EŸà4of60åÕ••XÓ„••XÓ„rootsbjtmpobjmnrnmodfßc(æÆå‹EŸàNNADADVVNNNN5of60íI•íIÒrèŸËåPfßc(æÆå‹EŸà‹öÓ„Event:åAgent:fßTime:c(Manner:æÆPatient:‹EŸà6of60íI•íIÒrèåÕíI•Xsûþrootsbjtmpobjmnrnmodfßc(æÆå‹EŸàMannerTimePatientAgent7of60ûß

3、:chn-processor„ûߟýÍ
Í'è•X
ííÓ„•úŽííÓ„„íIÒrè}0@ÁþQÙ8of60!þúŽW{„IíÍX=f,Œ,,Û,...,u,±,Y,N,...g9of60!þúŽW{„IíÍX=f,Œ,,Û,...,u,±,Y,N,...g9of60!þúŽW{„IíÍX=f,Œ,,Û,...,u,±,Y,N,...gS=fB;I;E;Sg9of60!þúŽW{„IíÍX=f,Œ,,Û,...,u,±,Y,N,...gS=fB;I;E;Sgu+3;„ØÆå°9of60!þúŽW{„IíÍX=f,Œ,,

4、Û,...,u,±,Y,N,...gS=fB;I;E;Sgu+3;„ØÆå°BIEBESBEBE9of60!þúŽÍ{„IíÍ'èX=fu+3,‹Ë›,„™e,...,“

5、ð,“

6、›,...U.,...gS=fNN;NR;VV;VE;:::g10of60!þúŽÍ{„IíÍ'èX=fu+3,‹Ë›,„™e,...,“

7、ð,“

8、›,...U.,...gS=fNN;NR;VV;VE;:::gfßc(æÆå‹EŸà10of60!þúŽÍ{„IíÍ'èX=fu+3,‹Ë›,„™e,...,“

9、ð,“

10、›,...U.,...gS=fNN;NR;VV;VE;:::

11、gfßc(æÆå‹EŸàNNADADVVNNNN10of60!þProbabilisticSolutionsPr(X1=u;X2=+;:::;X10=°;S1=B;S2=I;:::;S10=E)Pr(X1=fß;X2=c(;:::;S1=NN;S2=AD;:::)11of60!þProbabilisticSolutionsPr(X1=u;X2=+;:::;X10=°;S1=B;S2=I;:::;S10=E)Pr(X1=fß;X2=c(;:::;S1=NN;S2=AD;:::)SequenceModelsWehavetwosequencesofrandomvaria

12、bles:X1;X2;:::;XnandS1;S2;:::;SnAssumethateachSiisinS=f1;2;:::;kg,andeachXiisinX=f1;2;:::;ogoreachXiisavector.HowdowemodelthejointdistributionPr(X1=x1;:::;Xn=xn;S1=s1;:::;Sn=sn)11of60HMMGiventheHMMandanobservationsequencex1;:::;xn,Q1:Themostprobablestatesequence?argmaxp(s1:::snjx1:::

13、xn;;)si2SQ2:Theprobabilityoftheobservationsequence?Xp(x1:::xn;;)=p(x1:::xn;s1:::sn;;)si2SQ3:Theprobabilitydistributionoverstates,foreachxi?Xp(x1:::xn;si=s;;)=p(x1:::xn;s1:::si=s:::sn;;)sj2S(j6=i)12of60HMMHowtoestimatetheparametersoftheHMM(=h;i)?Q4:Givenstatesand