| W04-1102 | course , the second question in the | Chinese tokenization | process . Therefore , we will |
| P15-1004 | Arabic word tokenization . For | Chinese tokenization | , we use a simple longest-match |
| E14-1072 | together . In particular , we applied | Chinese tokenization | ( Chang et al. , 2008 ) , and |
| P14-1129 | tokenizer for the NIST condition . For | Chinese tokenization | , we use a simple longest-match-first |
| W95-0114 | on our results . 7.1 Effect of | Chinese Tokenization | We used a statistically augmented |
| W06-3601 | our results are independent of | Chinese tokenizations | ( although our language models |
| W06-0139 | contrast , after correcting the | Chinese tokenization | rules as well as SIGHAN official |
| W95-0114 | 1994 ; Wu & Fung 1994 ) . | Chinese tokenization | is a difficult problem and tokenizers |
| A94-1030 | acquisition tools . <title> IMPROVING | CHINESE TOKENIZATION | WITH LINGUISTIC FILTERS ON STATISTICAL |
| A94-1000 | Pazienza 174 Posters Improving | Chinese Tokenization | with Linguistic Filters on Statistical |
