Applications of Data Analytics and Machine Learning for Digital Twin-based Precision Biodiversity: A Review

Abstract

Biodiversity projections and model evaluation are essential to inform future formulation of biodiversity policy. These could be supported by data analytics and machine learning approaches, as well as precision technologies. However, existing works are segregated by the selection of species under-study and depending on the location. This paper reviews the existing approaches for precision biodiversity covering dashboard and data analytics, deep learning and machine learning, and digital twin for precision biodiversity. We propose a framework based on interactive machine learning that could facilitate a continuous biodiversity projection modeling to facilitate incremental learning and reduce uncertainties from the complex factors that contribute to biodiversity declines. The proposed framework exploits digital twin model based on a research forest setting that pioneers this work in Malaysia. The framework comprises of short-term quick wins and long-term expectation of digitalization transformation towards precision biodiversity.