A Machine-Learning-Based Approach to Predict Deforestation Related to Oil Palm: Conceptual Framework and Experimental Evaluation

Abstract

Featured Application: This work applies machine learning to enhance the prediction of deforestation related to oil palm. This research can be used for decision makers trying to foresee and manage deforestation caused by palm oil production. Providing information about deforestation prediction can help users to make appropriate decisions about where and when they can establish new plantations to ensure a sustainable oil palm production. Deforestation is recognized as an issue that has negative effects on the ecosystem. Predicting deforestation and defining the causes of deforestation is an important process that could help monitor and prevent deforestation. Deforestation prediction has been boosted by recent advances in geospatial technologies and applications, especially remote sensing technologies and machine learning techniques. This paper highlights the issue of predicting deforestation related to oil palm, which has not been focused on in existing research studies. The paper proposes an approach that aims to enhance the prediction of deforestation related to oil palm plantations and palm oil production. The proposed approach is based on a conceptual framework and an assessment of a set of criteria related to such deforestation. The criteria are assessed and validated based on a sensitivity analysis. The framework is based on machine learning and image processing techniques. It consists of three main steps, which are data preparation, model training, and validation. The framework is implemented in a case study in the Aceh province of Indonesia to show the feasibility of our proposed approach in predicting deforestation related to oil palm. The implementation of the proposed approach shows an acceptable accuracy for predicting deforestation.