| lr,34-5-P03-1051,bq | expanded <term> vocabulary </term> and <term> | training corpus | </term> . The resulting <term> Arabic word | #4740 To improve the segmentation accuracy, we use an unsupervised algorithm for automatically acquiring new stems from a 155 million word unsegmented corpus, and re-estimate the model parameters with the expanded vocabulary andtraining corpus. |
| other,15-7-P03-1051,bq | algorithm </term> can be used for many <term> | highly inflected languages | </term> provided that one can create a small | #4780 We believe this is a state-of-the-art performance and the algorithm can be used for manyhighly inflected languages provided that one can create a small manually segmented corpus of the language of interest. |
| other,32-5-P03-1051,bq | parameters </term> with the expanded <term> | vocabulary | </term> and <term> training corpus </term> . | #4738 To improve the segmentation accuracy, we use an unsupervised algorithm for automatically acquiring new stems from a 155 million word unsegmented corpus, and re-estimate the model parameters with the expandedvocabulary and training corpus. |
| other,11-1-P03-1051,bq | </term> by a <term> model </term> that a <term> | word | </term> consists of a sequence of <term> morphemes | #4611 We approximate Arabic's rich morphology by a model that aword consists of a sequence of morphemes in the pattern prefix*-stem-suffix* (* denotes zero or more occurrences of a morpheme). |
| other,30-7-P03-1051,bq | manually segmented corpus </term> of the <term> | language | </term> of interest . A central problem of | #4795 We believe this is a state-of-the-art performance and the algorithm can be used for many highly inflected languages provided that one can create a small manually segmented corpus of thelanguage of interest. |
| lr,15-6-P03-1051,bq | <term> exact match accuracy </term> on a <term> | test corpus | </term> containing 28,449 <term> word tokens | #4758 The resulting Arabic word segmentation system achieves around 97% exact match accuracy on atest corpus containing 28,449 word tokens. |
| other,12-3-P03-1051,bq | </term> to determine the most probable <term> | morpheme sequence | </term> for a given <term> input </term> . The | #4682 The algorithm uses a trigram language model to determine the most probablemorpheme sequence for a given input. |
| model,8-1-P03-1051,bq | Arabic 's rich morphology </term> by a <term> | model | </term> that a <term> word </term> consists of | #4608 We approximate Arabic's rich morphology by amodel that a word consists of a sequence of morphemes in the pattern prefix*-stem-suffix* (* denotes zero or more occurrences of a morpheme). |
| measure(ment),10-6-P03-1051,bq | system </term> achieves around 97 % <term> | exact match accuracy | </term> on a <term> test corpus </term> containing | #4753 The resulting Arabic word segmentation system achieves around 97%exact match accuracy on a test corpus containing 28,449 word tokens. |
| model,4-3-P03-1051,bq | </term> . The <term> algorithm </term> uses a <term> | trigram language model | </term> to determine the most probable <term> | #4674 The algorithm uses atrigram language model to determine the most probable morpheme sequence for a given input. |
| tech,1-3-P03-1051,bq | unsegmented Arabic corpus </term> . The <term> | algorithm | </term> uses a <term> trigram language model | #4671 Thealgorithm uses a trigram language model to determine the most probable morpheme sequence for a given input. |
| model,1-4-P03-1051,bq | for a given <term> input </term> . The <term> | language model | </term> is initially estimated from a small | #4690 Thelanguage model is initially estimated from a small manually segmented corpus of about 110,000 words. |
| other,15-5-P03-1051,bq | </term> for automatically acquiring new <term> | stems | </term> from a 155 million <term> word </term> | #4721 To improve the segmentation accuracy, we use an unsupervised algorithm for automatically acquiring newstems from a 155 million word unsegmented corpus, and re-estimate the model parameters with the expanded vocabulary and training corpus. |
| lr,28-2-P03-1051,bq | word segmenter </term> from a large <term> | unsegmented Arabic corpus | </term> . The <term> algorithm </term> uses a | #4666 Our method is seeded by a small manually segmented Arabic corpus and uses it to bootstrap an unsupervised algorithm to build the Arabic word segmenter from a largeunsegmented Arabic corpus. |
| lr,21-5-P03-1051,bq | from a 155 million <term> word </term><term> | unsegmented corpus | </term> , and re-estimate the <term> model | #4727 To improve the segmentation accuracy, we use an unsupervised algorithm for automatically acquiring new stems from a 155 million wordunsegmented corpus, and re-estimate the model parameters with the expanded vocabulary and training corpus. |
| lr,9-4-P03-1051,bq | is initially estimated from a small <term> | manually segmented corpus | </term> of about 110,000 <term> words </term> | #4698 The language model is initially estimated from a smallmanually segmented corpus of about 110,000 words. |
| lr,25-7-P03-1051,bq | provided that one can create a small <term> | manually segmented corpus | </term> of the <term> language </term> of interest | #4790 We believe this is a state-of-the-art performance and the algorithm can be used for many highly inflected languages provided that one can create a smallmanually segmented corpus of the language of interest. |
| tech,9-5-P03-1051,bq | </term><term> accuracy </term> , we use an <term> | unsupervised algorithm | </term> for automatically acquiring new <term> | #4715 To improve the segmentation accuracy, we use anunsupervised algorithm for automatically acquiring new stems from a 155 million word unsegmented corpus, and re-estimate the model parameters with the expanded vocabulary and training corpus. |
| other,20-5-P03-1051,bq | <term> stems </term> from a 155 million <term> | word | </term><term> unsegmented corpus </term> , | #4726 To improve the segmentation accuracy, we use an unsupervised algorithm for automatically acquiring new stems from a 155 millionword unsegmented corpus, and re-estimate the model parameters with the expanded vocabulary and training corpus. |
| other,35-1-P03-1051,bq | denotes zero or more occurrences of a <term> | morpheme | </term> ) . Our method is seeded by a small | #4635 We approximate Arabic's rich morphology by a model that a word consists of a sequence of morphemes in the pattern prefix*-stem-suffix* (* denotes zero or more occurrences of amorpheme). |
