Towards a probabilistic modal logic for semantic-based information retrieval

摘要: Semantic-based approaches to Information Retrieval make a query evaluation similar an inference process based on semantic relations. find out hidden relationships between document and query, but quantitative estimation of the correspondence them is often empiric. On other hand, probabilistic usually consider only statistical terms. It expected that improvement may be brought by integrating these two approaches. This paper demonstrates, using some particular models which are strongly related modal logic, such integration feasible natural. A new model developed basis extended logic. has advantages : (1) augmenting semantic-based approach with measurement, (2) finer relations than just ones. shown this verifies most conditions for absolute probability function.