Neural Network Regression with Support Vector Regression for Land-Use Growth Prediction

Abstract

Growth prediction in a developing cities is important for the city planners to prevent environmental degradation, social and health problems, improper land-use location, etc. Bekasi City in Indonesia, is experiencing rapid urban growth and as a “post-suburbia” city faced many problems related to the land use management. This paper proposed a combination of Nonlinear Autoregressive with External Input (NARXNET) and the Support Vector Regression (SVR) to predict the land-use growth in the study area. The result showed the proposed method was able to predict the number of land-use in Bekasi City several years ahead.