Natural Language Query Processing Based on Probabilistic Context Free Grammar

The field of natural language processing (NLP) has seen a dramatic shift in both research direction and methodology in the past several years. In the past, most work in computational linguistics tended to focus on purely symbolic methods. Recently, more and more work is shifting toward hybrid methods that combine new empirical corpus-based methods, including the use of probabilistic and information theoretic techniques, with traditional symbolic methods. The main purpose of Natural Language Query Processing is for an English sentence to be interpreted by the computer and appropriate action taken. Asking questions to databases in natural language is a very convenient and easy method of data access, especially for casual users who do not understand complicated database query languages such as SQL. This paper proposes the architecture of a new NLDBI system including its probabilistic context free grammar, the inside and outside probabilities which can be used to construct the parse tree and the usage of dependency structures and verb sub categorization in analyzing the parse tree.