Review
Huang Heyan , Chen Zhaoxiong , Song Jiping
1999, 13(5): 44-51.
It is hard for the existed single strategy machine translation approaches to solve all the problems in machine translation. In this paper ,we propose an interactive hybrid strategies machine translation approaches ,which is incorporated with rule-based parsing with multiple knowledge description ,analogy inference with heuristic search based on case translation memory ,statistic knowledge and some extend human-computer interaction. In which ,the featured case model base including various features is set up automatically based on the existed intelligent machine translation system. It is not only possible for it to use case-based heuristic analogy reasoning to produce translation result for same or similar sentence with a former translated sentence ,but also can use rule-based parsing to translate the sentences which are not similar to any model in case base. In addition ,when the knowledge is not sufficient to disambiguate the ambiguities in the current translated sentence ,it can use human interacive mode to guide the translation process ,so as to greatly enhance the translation speed and accuracy ,and made the translation system more practicable.