Finding structure via compression
Jason L. Hutchens, Michael D. Alder · 1998
A statistical language model may be used to segment a data sequence by thresholding its instantaneous entropy. In this paper we describe how this process works, and we apply it to the problem of discovering separator symbols in a text. Our results show that language models which bootstrap themselves with structure found in this way undergo a reduction in perplexity. We conclude that these techniques may be useful in the design of generic grammatical inference systems.