PEMODELAN PREDIKSI KONVERSI PENGGUNAAN LAHAN BERBASIS ANN-CA DI WILAYAH PERI-URBAN KABUPATEN SLEMAN

Abstract

The development in the city has caused urban areas to experience significant growth due to increased activity. One of the visible changes is the change in the use of vegetated land for built-up land. The research location is in a peri-urban area of Depok and Mlati subdistrict, Sleman Regency. This research analyzed land use conversion in 2015-2020 and modeled land change predictions for the next 20 years (2025-2045) using Artificial Neural Network - Cellular Automata (ANN-CA). The ANN method used multiple output neurons to determine the probability of land use transition. CA was used to model land use change by applying transition probabilities. The source of land use data came from extracting SPOT images, and then the modeling process used QGIS Desktop 2.18.11 on the MOLUSCE plugin. The results showed that the peri-urban area experienced a decrease in agricultural and livestock land by 152.62 ha (2.52%) while building land increased by 148.74 ha (2.46%). The 2025-2045 land use conversion prediction shows that the reduction in agricultural land, plantations, and livestock will continue, and the land area for buildings and roads will increase.