Forecasting transport development: Testing the utility of a simple regression approach

The analysis tests the utility of a simple regression approach in postdictively explaining transport development in West Malaysia from 1887 to 1968. The growth of the network is modelled as a process of contagious diffusion where transport densities of adjacent cells are used as predictor variables on a lagged basis. The partial regression coefficients provide measures of network orientation over time. The results demonstrate the importance of the contagion process in transport forecasting and provide equivalent levels of explanation when compared with a model which utilizes modernization indexes as predictor weights.I am grateful to Peter Gould for his advice and suggestions in the early stages of the research and for the comments of Peter Haggett, Mark Monmonier and J. Barry Riddell, all of whom read drafts of the paper. I also wish to acknowledge the financial support provided by a University of Vermont Faculty Fellowship.