Prediksi Perkembangan Lahan Terbangun di Jabodetabek Hingga Tahun 2030 menggunakan Artificial Neural Network dan Cellular Automata

Abstract

Hybrid models that are widely used in current spatial prediction studies such as Markov chain, linear logistics and others, have weaknesses in determining and very sensitive parameters. So it requires a lot of data, time consuming, and inefficient. Therefore, a new model is needed that can be handled more easily, less time consuming and more efficiently. In this case, the researchers tried the ANN model to be applied in predicting the development of built-up land. One area that is already very densely built up is Greater Jakarta. The simulation was validated using 2020 built-up land cover and resulted in a simulation accuracy of 74% for simulations per 5 years and 85.7% for simulations per 10 years. The development of built-up land in Jabodetabek has increased and decreased from 2000-2030. The development of Jabodetabek built land provides a proof of several theories of urban geography, namely the perspective of the city as an organism and the formation of the Jakarta as a fan morphology.