On keyword-based ad-hoc querying of hospital data stored in semistar data ontologies

This paper sketches a possible solution to the problem of the currently growing necessity in various domains for domain experts to be able to query the database of the organization in a convenient manner. The paper focuses on the domain of hospital management where the normal practice is to involve a programmer as an intermediary between the managers and the database. This is an error-prone and cumbersome solution. The decision-making process of domain experts would hugely benefit if they could retrieve the information from the database themselves. There have been attempts to develop natural language-based query languages for this exact purpose, but the ultimate goal of the simplicity of querying has not yet been reached. The approach presented in this paper involves letting users define their queries in a very weakly-controlled natural language and then to choose among the offered query translations into a more strongly-controlled natural language which already has an efficient implementation for semistar ontologies. Experiments show that the implemented query language is very readable for non-programmers, because it lacks technical details (thanks to the nature of semistar ontologies), and because it is very intuitive. This phenomenon creates a conviction that the proposed approach is highly viable.