Ontology acquisition and semantic retrieval from semantic annotated Chinese poetry

We aim to utilize semantic Web [James Hendler et al., (2002)] technology to the semantic annotation of classical Chinese poetry. We investigate the feasibilities and advantages of semantic retrieval and automated ontology acquisition from semantically annotated poems based on a Chinese thesaurus. We have induced a set of semantic composition rules for pair-wise character (word) patterns that can be used to parse the poem sentences and recursively generate RDF [(1999)] triple relations among the pair of characters (words). We have also defined a scoring scheme to assess semantic similarity for semantic retrieval. We showed that the semantic retrieval method significantly outperformed the keyword-based retrieval method.