| Neural network is used as an alternative for modeling land use change in coastal area. The neural net is superior to the logistic model in terms of approximation, accuracy and error as shown by this study. The results also show that neural net is more reliable for identification of spatial patterns and relationships.; Using hedonic analysis of residential properties and employing spatial econometric analysis, various spatial variables were estimated to have significant impact on the values of residential properties in urban/rural interface of Greenville County, South Carolina. The spatial-temporal model reduces the problem of error correlation with the nearest neighbors relative to those produced by the non-spatial OLS model.; Both urban-rural fringe location and lake proximity were estimated to have significant positive impact on the values of residential properties in the upstate of South Carolina. Spatial econometric models are employed to account for spatial effects in housing price analysis and to increase prediction accuracy. |
