W02-1003 used the interpolated absolute discounting method . Since both the traditional
W10-3404 , 1991 ) . They compared some discounting methods by using the newspaper corpus
C00-2105 proba - bilities . The absolute discounting method ( Ney et al. , 1994 ) was adapted
W06-2601 particular , smoothing was based on the discounting method in Ney et al. ( 1994 ) combined
C02-1045 of each event , we use a simple discounting method similar to the absolute discounting
J11-4008 unigrams alone . • We specify a discounting method which facilitates better treatment
P10-1160 Ju - rafsky , we also used the discounting method suggested by Pantel and Ravichandran
P94-1038 than predicted by the uniform discounting method . Finally , the similarity-based
C04-1022 . The rst integer speci es the discounting method while second indicates the minimum
D09-1076 site b alone . Even if we use a discounting method to give a nonzero probability
C04-1022 features , backo procedures , and discounting methods . Usually , this space can not
W05-1107 This is a running theme of the discounting methods of smoothing such as absolute
D11-1064 useful as , unlike count-based discounting methods ( e.g. , Kneser - Ney ) , it
W12-3158 probabilities and a modified absolute discounting method that can be applied to fractional
P11-1009 model - ing . The simplest are discounting methods , such as additive smoothing
W02-0504 Katz ( 1987 ) that combines a discounting method along with a back-off method
D09-1078 widely-used Katz backoff and absolute discounting methods . We are not aware of any other
W07-0712 namely the Witten - Bell linear discounting method ( Witten and Bell , 1991 ) .
J11-4008 a way that is similar to other discounting methods , this formalization satisfies
P98-2157 computation in the following ways . Discounting methods for smoothing simply produce
hide detail