KEYRY: A Keyword-Based Search Engine over Relational Databases Based on a Hidden Markov Model

We propose the demonstration of KEYRY, a tool for translating keyword queries over structured data sources into queries in the native language of the data source. KEYRY does not assume any prior knowledge of the source contents. This allows it to be used in situations where traditional keyword search techniques over structured data that require such a knowledge cannot be applied, i.e., sources on the hidden web or those behind wrappers in integration systems. In KEYRY the search process is modeled as a Hidden Markov Model and the List Viterbi algorithm is applied to computing the top-k queries that better represent the intended meaning of a user keyword query. We demonstrate the tool’s capabilities, and we show how the tool is able to improve its behavior over time by exploiting implicit user feedback provided through the selection among the top-k solutions generated.