Some Extensions to Rushton's Spatial Preference Scaling Model

Rushton's spatial preference scaling model is extended and its computing algorithm modified in order to utilize more fully all the information contained in the input data set, and to make it a better predictor of choice behavior. The preference scale that is produced by the model measures the certainty with which generalizations can be made about choices between classes of objects. The model offers two significant advantages over commonly used spatial interaction models: freedom from a priori selection of the function relating choice to attributes of objects, and applicability to studies where the number of origin and destination zones is large relative to the populations involved.