Knowledge Representations for Constructing Chains of Contexts in Geographic Information Systems

Abstract

Solving complex informal problems using spatial data is often used in industry and business. In the absence of a solution algorithm, analyst resorts to a heuristic search for a solution, which is based on an interactive dialogue with a geographic information system (GIS). The analyst builds a cartographic workspace by sending GIS queries. The workspace is explored visually using software tools for spatial, statistical, and other special types of analysis. During the analysis, the user raises his situational awareness. It is important to organize the dialogue wisely as there is a danger of cognitive overload. In the latter case, there is a possibility of making the wrong decision. This paper proposes a method for representing knowledge about the process of analyzing spatial situations. The essence of the method is that knowledge is represented by chains of contexts and patterns of appropriate user behavior in visual analysis. A model of an image of a chain of contexts is proposed, which consists of the center of the image and its admissible transformations. A method of organizing a dialogue is described, which allows a GIS to develop recommendations for choosing a context in a chain of contexts. Patterns of level, tendency, and rhythm are proposed to represent knowledge about appropriate behavior. Application of the proposed method of knowledge representation in recommender GIS improves the quality of spatial data analysis.

Keywords

• Geographic information systems
• Context
• Spatial analysis
• Context-aware recommender systems