Chinese Parsing in a Phoneme-to-Character Conversion System Based on Semantic Pattern Matching
Wen‐Lian Hsu · 1994
assigned to templates so that the final outcome is not decided by a fixed rule (as in a traditional rule-based system); rather, the accumulated effect of different matched templates makes the desired outcome more likely to be obtained. The design of the CSM is intended to resolve local ambiguities using as much contextual information as possible. The underlying principle of the CSM is purely computational and bears no relation on the linguistic knowledge per se. A main feature of the CSM is its ability to accommodate all kinds of linguistic knowledge as well as computational efficiency consideration in a coherent fashion. The robustness of CSM has already shown its effectiveness in the PTC system "GOING". We shall discuss the application of the CSM in constructing a parser for the PTC system. We have recently developed a Chinese phoneme-tocharacter conversion system with an accuracy close to 96%. The underlying algorithm, called the context sensitive method, is based on semantic patte...