Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
926769 | Cognition | 2013 | 15 Pages |
Abstract
⺠We test effects of a Zipfian frequency distribution on statistical word segmentation. ⺠Two large-scale experiments are conducted over the web. ⺠Identifying words is not more difficult within a Zipfian language. ⺠Zipfian input distributions facilitate accurate segmentation if test items are presented in context. ⺠We suggest chunking models better simulate the human performance than transition finding models.
Related Topics
Life Sciences
Neuroscience
Cognitive Neuroscience
Authors
Chigusa Kurumada, Stephan C. Meylan, Michael C. Frank,