IDENTIFICATION OF THE MAIN DRIVING FACTORS OF CHANGE IN COVER AND LAND USE BASED ON THE RANDOM FOREST METHOD IN EAST JAVA PROVINCE

• Main Author: Andhika Pratama, Muhammad
• Format: Final Project
• Language: Indonesia
• Subjects: Teknik (Rekayasa, enjinering dan kegiatan berkaitan)
• Online Access: https://digilib.itb.ac.id/gdl/view/75136
• Institution: Institut Teknologi Bandung

Change and land use is an important issue in the management of natural resources and sustainable development. Understanding the factors contributing to land change can provide valuable insights for regional decision-making and planning. In the last decade, East Java has experienced a significant phenomenon of land change which has a negative impact on the environment and human life, so it is necessary to carry out land control efforts by identifying the driving factors for land change which represent physical aspects such as elevation, slope, and soil type. social in the form of population growth and economic in the form of accessibility. The purpose of this research is to analyse the pattern of land cover change in 2000 – 2018 and identify the main driving factors for land cover change in East Java. The random forest regression method is a type of machine learning used in this study because it can overcome complexity in data and has the ability to identify and assess the influence of factors driving land change, while for land change patterns, the analysis is carried out by dividing into 3 periods with interval of 6 years. The data used in this research includes land cover data from the Ministry of Environment and Forestry, DEM data, GRIP data, Soil Map of the World data, East Java administration data and Worldpop data. The results of the study show that in 3 periods of land change, the forest land class has the potential to become dry land and the inundated vegetation class has the potential to become ponds. Built-up areas had the largest proportion of increase, namely 33.24%, while open closures had the largest proportion of decline, namely 22.37%. The results of the random forest regression model with the independent variables in the form of 7 driving factors and the dependent variable using the area of land change obtained a coefficient of determination of 87.26% which indicates the strong influence of the driving factor on the level of area change, with the variable distance to the economic centre as the main driving factor.