Brad Carter
Locally Weighted Linear Combination in a Vector Geographic Information System © 2012
This research explores the use of a local form of weighted linear combination, a multi-criteria decision analysis technique that can be used by decision-makers to select an optimal location from a collection of location alternatives. Its local form takes into account the range of attribute values within a user-defined neighbourhood in accordance with the “range sensitivity principle”. The research involved the development of an application within the ArcMap 10 Geographic Information System. The application allows the user to select and apply global weights to multiple decision criteria that are stored within a single vector (polygon) layer, and select a neighbourhood definition from a standard set including contiguity, distance, and K-nearest neighbours. Vulnerability to heat-related illness in the City of Toronto is used as a case study to examine the locally weighted linear combination technique. A combination of demographic and environmental variables was used to assess heat vulnerability at the census tract level.