W14-5513 clusters resulting from previous character clustering process . These small clusters
W14-5513 separation , mainly concentrate on character clustering . First , we cluster each character
W14-5513 separation , mainly concentrate on character clustering . First , we cluster each character
P98-1108 training data using or not using character clustering . 15 Table 1 shows results obtained
W06-0120 class information generated by character clustering to impose a class limitation
P98-1108 each experiment , we used the character clustering based on MI . Each question for
N09-2069 to perform word clustering or character clustering to alleviate the data sparseness
W14-5513 PB-SMT task that using different character clustering approaches . Each training set
P98-2152 of estimating P ( RIM ) . After character clustering , ` M ' and ` a ' are clustered
W14-5513 mentioned in section 2.1 , we apply character clustering ( CC ) technique on target text
W14-5513 over time when several different character clustering approaches are applied . Next
W06-0120 apply the K-means algorithm to character clustering and develop several cluster set
W06-0120 An Improved Model Coupled with Character Clustering Automatically Generated Template
W06-0120 the character " 国 " . 2.2 Character Clustering In many cases , Chinese sentences
P98-1108 word classes but also to compute character clusterings in Japanese . The basic algorithm
P98-1108 presents our method of incorporating character clustering based on mutual information into
P98-2152 computationally efficient . 2.4 Character Clustering In general , character recognition
W00-0803 ; Rehder et al , 1998 ) or Han character clustering are the potential solutions to
W06-0130 instead of BIO tags and word and character clustering features . <title> POC-NLW Template
P98-1108 method . 3 Mutual Information-Based Character Clustering One idea is to sort words out
hide detail