A conceptual model for a collaborative GIS

Most current natural language based Geographical Information Systems (GIS) still interact with the user with the “question-answer” model. They do not support users effectively because the communication between the user and the system is easy to be broken down due to the ambiguity nature of the natural language. The collaborative interaction model was proposed for GIS so that GIS could simulate a GIS expert to collaborate with the user to successfully reach the user’s request. A knowledge base is necessary to such a collaborative GIS. This paper describes the process how a conceptual model for such a knowledge base was created though an interpretive approach. Three data collection methods, including participant observation, interview, and questionnaire, were conducted in order. The findings of the research shows that 1) the GIS expert and the user can collaborate with each other by many collaborative strategies; 2) There are multiple natural language expressions used by the layman user to indicate a request for a GIS command; 3) The GIS expert needs to check a GIS command and the constraints associated with it before executing it; 4) The conceptual model for the knowledge base of a collaborative GIS needs at least three components: one for controlling the reasoning and action process, one for understanding natural language and outputting response, and one for storing collaborative strategies and other knowledge. The conceptual model was iteratively developed based on the results of three data collection methods.