A generic resampling particle filter joint parameter estimation for electricity prices with jump diffusion

In this paper, a particle filter for parameter estimation of jump diffusion models employed for modeling electricity prices [1], [2], [3] is implemented. A jump-diffusion model [4] is investigated. The jumps have the possibility to give a better explanation of the behavior of electricity prices [5]. Introduction of the jump components however complicates parameter estimation problem by the inclusion of several new parameters [4]. These parameters will describe the jump frequency and distribution. The jump models are non-Gaussian and this increases the complexity of the models further, [4], [5]. A known filtering technique for these models is particle filter [1], [2], [3]. The performance of generic particle filter to model the jump frequency and distribution parameters has been investigated. In this paper, a preliminary study is conducted. The performance of augmented generic particle filter to model the jump frequency and distribution parameters has been analyzed for a benchmark example employed in the maximum likelihood state estimator solution of [4] and favorable results were obtained. The results are compared with bench-marking closed form solution of [4] in order to once again to highlight the contribution of the paper.