Support Vector Machines for Eco-Environmental Quality Evaluation.

Abstract: SVM (support vector machine) is a new machine learning technique developed on statistical learning theory. Because of its excellent learning performance, this technology has become the new research hot spot. In this paper, RS and GIS based comprehensive SVM model were adopted in evaluating eco-environment of Guangzhou City. First, six notable indices were collected by considering about the principal factors of regional environment: heat, NDVI, humidity, brightness, altitude, and annual rainfall. And then, an integrated evaluation model based on SVM was adopted for calculating the eco-environmental equality level. Finally, we analyzed the spatial distribution of ecoenvironmental equality level and driving force in forming the eco-environmental quality status. Our results have contributed on the decision making in regional economic development and environment protection. [Copyright &y& Elsevier]