Alignment Based Similarity Measure for Grammar Learning
Xiangrui Wang, Narendra S. Chaudhari · 2006
We introduce a similarity measure, called alignment profile similarity, for the learning of context-free grammar from given language samples. Based on the alignment profile similarity, an alignment learning framework for grammatical inference is proposed. Alignment profile similarity is used to improve alignments, and therefore increase the quality of the rules identified. The experiments show that the proposed methods improve performance in terms of the percentage of correctly generated grammar rules.