Natural Language Generation of Museum Object Descriptions based on User Model

Natural Language Generation (NLG) techniques can be applied in generating virtual documents dynamically using information from a database (Dale et al, 1999). One of the applications of NLG techniques to generate documents dynamically is the web-based interactive virtual museum, VIGAN. NLG is used to generate the descriptions of the objects in a virtual museum dynamically based on the profile and interests of the visitor. The focus of the research is on incorporating user’s interest, age group, and visit history in the generation of museum object descriptions. The descriptions do not vary only on user’s profile but also in lexicalization. Facts are not only in describing the objects, but also in describing Ilocano personalities. User Acceptance Testing proved that object descriptions do vary based on age groups, category of interest, and lexicalization. They commented that the descriptions are easy to understand, the user interface is user friendly and the suggested objects are appropriate.