A Parsing Model Based on Ordered Tree Inclusion Matching

In order to integrate the context and linguistic information into the parsing model in a holistic and strict way, and to describe the complex structure more precisely, a parsing model is proposed to visually take the above information into account. Considering that the functional words and the words with special usage can have a profound impact on the syntactic structure, some patterns are built especially for them, and the probabilities of patterns are pre-computed and stored. The matching between partial parse and pattern is based on the ordered tree inclusion, and the probability of parse tree is computed on the basis of the matched patterns. It is plausible to match and retrieve the probability from the pattern bank for there are many algorithms with linear space and time spent for ordered tree inclusion problem. The method can provide a vivid, strict and somewhat fuzzy restriction for the complex structures modeling.