International Journal on Advanced Science, Engineering and Information Technology, Vol. 11 (2021) No. 6, pages: 2328-2333, DOI:10.18517/ijaseit.11.6.14582

Biogeographical Distribution Model of Flowering Plant Capparis micracantha Using Support Vector Machine (SVM) and Generalized Linear Model (GLM) and its Ex-situ Conservation Efforts

Inggit Puji Astuti, Angga Yudaputra, Dipta Sumeru Rinandio, Ade Yusuf Yuswandi

Abstract

Capparis micracantha is a flowering plant species with a wide range of distributions. It is often used as traditional medicine by local people. This species is often overlooked due to very limited information that reveals about its conservation aspects. This study aims to update the distribution records and predict the potential current distribution of this species. Some new occurrence records were obtained through a direct field survey, plant inventory database, and reliable scientific papers. In order to gain more information about current distribution, Species Distribution Modelling (SDM) was applied to predict the potential current distribution of this species in Indonesia.  Two algorithms of machine learning (SVM and GLM) were applied to produce predictive distribution maps. The models were built from occurrence records data and environmental variables (climate and topography). Area Under Curve (AUC) was used to evaluate the model prediction. The AUC value of those models >0.80 means those models have a good performance. The AUC value of those models is stated in SVM (0.86) and GLM (0.824). SVM and GLM have almost similar resulting predictive maps of current distribution. The predictive maps would be useful to give information about the regions in Indonesia that have similar environment characteristics and climate conditions where the species are observably present. This species has also been conserved through ex-situ conservation strategies, but the seedling growth of this species still remains a challenge.

Keywords:

Biogeographical distribution; Capparis micracantha; Ex-situ conservation; GLM; SVM.

Viewed: 98 times (since abstract online)

cite this paper     download