An ANN-based risk assessment method for carbon pricing

This paper proposes an efficient method for risk assessment of carbon pricing with artificial neural network (ANN). The global warming is of main concern in the world. The power industry wants to make generation planning more flexible through the emission trading system. In this paper, an ANN-based method is proposed to predict one-step-ahead carbon pricing. As ANN, the radial base function network (RBFN) is used to approximate the nonlinear function of time-series carbon pricing. To improve the performance of RBFN, this paper makes use of preconditioned RBFN in a way that DA (deterministic annealing) clustering classifies learning data into some clusters and RBFN is constructed at each cluster. In addition, DA clustering is used to determine the center vectors of the Gaussian function in RBFN. Also, the Monte Carlo simulation is applied to the risk assessment of carbon pricing with the RBFN model. The risk of one-step-ahead carbon pricing is evaluated in probability. The proposed method is successfully applied to real data of the carbon pricing market.