| 摘要 |
The standard data responsible for ranking of symbols in the given text are extracted and subjected to all possible permutations in their arrangements (Pr) with associated symbols, and some of text sections in selected arrangement are extracted to form supplementary text (T) for grammar message (TN). The texts (T) to be evaluated are extracted from the speech recognition signals of different natural languages (L), from which the supplemented allocations of grammatical messages (TN) are extracted using semantic contents (f), probabilities (P(TN) of grammatical messages and their complexes (Kp = (EmEm)), which are extracted using probabilities text (p(Em)). From the contents the symbol (Gs), text (Gt) which fall in a disjoint class (k1 = Gt) of characteristic function H(T) of the text, determined through tolerances are acquired. The rank orders (Ra) are assigned corresponding to text symbols (Gs) under rank symbols (Gr) are created, corresponding to arrangement (P(r)) of grammatical messages. The standard data of grammatical messages that are responsible for (Gr) are combined with all possible permutations with associated GT's and a supplement message text (T) is formed for the grammatical message (TN). An Independent claim is also included for processing device for communicating grammatical messages. |
