Hi Thang, The ranges that are specified for moses are just suggestions for random starting points for the MERT algorithm. However, it may (and often does) find weights that end up outside of these ranges. -Chris
On Thu, Jun 4, 2009 at 2:32 AM, Thang Luong Minh <[email protected]>wrote: > Dear experts, > > I noticed for several settings of my system, I obtained optimal weights > which are negative even though the default search range is [0, 0.5]. E.g. > the weight for PhraseModel_4 (lexical feature) below. > PhraseModel_1 0.045 > PhraseModel_2 0.237 > PhraseModel_3 0.048 > PhraseModel_4 -0.097 > PhraseModel_5 0.024 > > I was wondering if this is a normal behavior, and hope someone could > enlighten me on that. If it is normal, how should I strictly enforce the > range? > > Regards, > Thang > > -- > Luong Minh Thang > WING group, School of Computing, National University of Singapore > http://wing.comp.nus.edu.sg/~lmthang<http://wing.comp.nus.edu.sg/%7Elmthang> > > _______________________________________________ > Moses-support mailing list > [email protected] > http://mailman.mit.edu/mailman/listinfo/moses-support > >
_______________________________________________ Moses-support mailing list [email protected] http://mailman.mit.edu/mailman/listinfo/moses-support
