Machine translation based on templates matching and replacing
Yujian Li, Zhong Yixin · 2002
The theory of and a universal algorithm for matching and replacing templates with an arbitrary number of invariables (constants) and variables are outlined, and their remarkable advantages and attractive prospects are demonstrated for applications in machine translation. Moreover, it is indicated that the translation quality could be further improved and may theoretically achieve an accuracy of 94% when the universal algorithm is combined with some statistical methods.