| 摘要 |
PURPOSE:To improve both the efficiency and accuracy of a sentence structure analyzer by obtaining the phrase structure of a sentence from the dictionary parts-of-speech string obtained by an allocating means and specifying the relationship between the noun and predicative phrases in said phrase structure. CONSTITUTION:When the close input of a sentence is carried out through a keyboard, the sentence is displayed on a CRTB. This input sentence is divided every word. Then a dictionary is consulted for each of these divided words to obtain the grammar information on the parts of speech, the tense, etc. Based on these information and grammar rules, an analysis tree is obtained. The analysis trees showing the correct relationship or correct meaning are discriminated among plural trees for extraction of the case control relation and the deep layer structure. The polysemous analysis trees are produced through the sentence structure analysis of the form information only. Therefore, the information on the case control of declinable words to select a correct analysis tree. Then only the analysis trees having the correct relationship are produced by checking the relationship between a noun phrase and a predicative phrase.
|
