Some tests of alternative urban population density functions

Abstract

The purpose of this study is to test the usefulness of some alternative gross population density functions for urbanized areas. Two evaluation criteria are used; maximum explanatory power in standard regression analysis and accuracy in predicting total population in the urbanized area. It is concluded that the explanatory power of the negative exponential function can be improved upon in some cases by adding a quadratic distance term, but that population can be predicted more accurately if the quadratic term is omitted. Also, it is found that constraining the negative exponential function to predict population exactly reduces explanatory power by an insignificant amount. © 1976.