Inside-outside reestimation from partially bracketed corpora
ACL '92 Proceedings of the 30th annual meeting on Association for Computational Linguistics
An Evaluation System for Metrical Models
Computer Music Journal
Refining the structure of a stochastic context-free grammar
IJCAI'01 Proceedings of the 17th international joint conference on Artificial intelligence - Volume 2
| Hi-index | 0.00 |
This paper describes Metrical PCFG Model that represents the metrical structure of music by its derivation. Because the basic grammar of Metrical PCFG model is too simple, we also propose a grammar expansion method that improves the grammar by duplicating a nonterminal symbol and its rules. At first, a simple PCFG model which that represent the metrical structure by a derivation just like parse tree in natural language processing. The grammar expansion operator duplicates a symbol and its rules in the PCFG model. Then the parameters of PCFG are estimated by EM algorithm. We conducted two experiments. The first one shows the expansion method specialized symbols and rules to adapt to the training data. Rhythmic patterns in a piece were represented by expanded symbols. And in the second experiment, we investigated how the expansion method improves performance of prediction for new pieces with large corpus.