PEMODELAN PERUBAHAN PENUTUPAN/PENGGUNAAN LAHAN DENGAN PENDEKATAN ARTIFICIAL NEURAL NETWORK DAN LOGISTIC REGRESSION (STUDI KASUS: DAS CITARUM, JAWA BARAT)

Abstract

Land use / cover change may occur due to factors of socio, politic, economy, culture, nature and technology. Geographical Information System can be used to analysis of landuse and connecting to driving factors. This research objectives are : to map land use/cover change in Citarum Watershed between 2000 and 2010, and to model its changing using Artificial Neural Network (ANN) and Logistic Regression (Logit). Land use/cover is interpreted using Landsat of 2010 and 2010, while the model is developed using software Idrisi Selva. Factors used as driving factors used are distance to road, distance to river, district income, slope steepness and population density. The results show that the largest decreased area is sawah (31.873 ha) and rural areas is the largest increasing area (41.574 ha), Sawah areas have been converted to to agricultural area and rural area 24.863 ha and 24.106 ha respectively. Data validation shows that land of 2012 has produce good result where Kappa value is 0,9622 for ANN and value of 0,9642 for Logit. In this case, land use projection using ANN and Logit may predict area of land use change.

Keywords: Artificial Neural Network, land use / cover, Logistic Regression, model