An Estimation Framework for Economic Cost of Land Use Based on Artificial Neural Networks and Principal Component Analysis with R

Abstract

Aiming to estimate the cost of environmental degradation influenced by land use change in the ecosystem, an ecosystem assessment model was established in this paper to quantify the cost of environmental degradation. 12 ecosystem service indicators were selected from the World Bank website and categorized into four principal components corresponding to four subservices including provisioning, regulating, habitat and cultural services. And 7 land use change indicators are classified into 5 dimensions. Then, the principal component analysis(PCA) was applied to classify indicators and calculate the contribution of each indicator to each dimension, which was visualized using R software. Furthermore, the coefficients of indicators to each principal component were determined and each component was expressed by a linear combination of indicators. Next, 8 economic cost indicators were divided into depletion class and saving class. Then, analytic hierarchy process(AHP) and entropy weight method(EWM) were employed to determine the weight of indicators belonging to two classes respectively. To integrate three indicator systems, an Artificial Neural Networks(ANN) model are implemented to calculate the coefficients and weights between indicators in ecosystem service and ones in economic cost system. By using weights and coefficients, this model can estimate the economic cost and benefit of each land-use indicator when investing one dollar on each indicator, which is defined as cost-benefit ratio. In order to determine land-use change scales, k-means clustering algorithm was utilized to determine the optical number of clusters which is 2, indicating small-scale and large-scale.