英文摘要 |
There are very few input systems supporting Taiwanese and Hakka on the market at present. Our purpose is to provide an input system supporting Mandarin and Taiwanese which is toneless and complies with multiple types of phonetic symbols, we hope that such an approach can resolve the problems resulting from tone and phonetic symbols. In this paper, we use an algorithm based on three continuous longest word first whose precision is 89.59% in one type of phonetic symbols, and 88.54% in the combination of many types of phonetic symbols. |