ACL
2006
15 years 4 months ago
2006
We propose an unsupervised segmentation method based on an assumption about language data: that the increasing point of entropy of successive characters is the location of a word ...