Forest Research: Open Access
Open Access

Abstract

BP Neural Network Based Prediction of Potential Mikania micrantha Distribution in Guangzhou City

Qiu L, Zhang D, Huang H, Xiong Q and Zhang G

To predict the distribution of Mikania micrantha, one of the most harmful invasive plants in Guangzhou City, the author selected relevant environmental factors and established a feasible simple model based on BP neural network to use its strong nonlinear ability in this paper. From this model, it is concluded that the distribution possibility of Mikania micrantha in Liwan District, Yuexiu District and Haizhu District is near 0, which are classified as regions without invasion risk; the distribution possibility in Conghua District and Huadu District is 60% and 69.3% respectively, which are defined as regions with low invasion risk; the distribution possibility in Baiyun District, Panyu District, Zengcheng District and Nansha District are much higher, which are identified as regions with high invasion risk; while the distribution possibility in Luogang District, Tianhe District and Huangpu District are the highest, which are determined as regions with highest risk.