MODELING TEMPORAL CHANGE IN FLOW MATRICES
Generate an AI Snapshot to get a quick, structured summary of this paper.
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
Abstract
ABSTRACT Regional analysts are often faced with the task of forecasting future flows. The ability to carry out such a task is often limited to a reliance on either a Markovian assumption or other techniques that use only one or two observed flow matrices. This paper explores alternative methods that may be used to combine the information contained in an observed temporal sequence of flow matrices. Information theoretic principles, lag structures, and constant causative matrix operators are among the techniques discussed. These methods are applied and evaluated using inter-regional migration flow data for the United States.
