國語語音辨認中詞群雙連語言模型的解碼方法 (A Word-Class Bigram Approach to Linguistic Decoding in Mandarin Speech Recognition) [In Chinese].