DRIVING FACTORS OF LAND COVER CLASS CHANGES IDENTIFICATION IN BANDUNG REGION USING BINARY LOGISTIC REGRESSION BASED ON GEOGRAPHIC INFORMATION SYSTEM

• Main Author: IDHA YOVITA - NIM: 25117001 , IRENE
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/28111
• Institution: Institut Teknologi Bandung

<p align="justify">Land cover changes need to be anticipated because it can cause harm to communities and ecosystems. The negative impacts of land cover change can be mitigated if land cover change phenomenon, including factors that encourage land cover change, is well understood. Factors that cause land cover changes for each land cover class and each region will be different. Therefore, a study of the drivers of land cover change should be carried out partially per class of land cover and consider the characteristics of a region. This research is to identify the driving factors of land cover changes for the forest, agricultural land, built-up area, and wetland in Bandung Region (Bandung City, Bandung Regency, Cimahi City, and West Bandung Regency). <br /> <br /> <br /> The driving factors studied in this research are curvature, contour, slope, distance to the center of Cimahi City, distance to the center of Bandung City, distance to the center of Bandung Regency, distance to the center of West Bandung Regency, and distance to the road. The driving factors were identified using Binary Logistic Regression (BLR) method to produce a model of land cover change. BLR is a method of calculating the probability of land cover change based on the land cover interaction with the drivers of the change. Model validation is applied to determine the accuracy of the model in predicting land cover changes. <br /> <br /> <br /> The results show that land cover changes for each class are influenced by heterogent drivers. Distance to the center of Bandung City is an important factor in encouraging land cover change in all types of class. In addition, the prediction model produced has a level of accuracy that is represented by the value of Overall Accuracy for forests of 30.58%, agricultural land of 56.26%, built-up area of 77.80% and water by 61.14%. If accumulated for the entire study area, the value of model accuracy is 41.57%.<p align="justify">