Parameter Estimation using Copula and State Dependent Particle Filter

This study aims to combine Copula and particle filter methods to estimate theparameters in the state-space model and focus on the financial model toconstruct the multi-risk factors price dynamic model for parameter estimation andempirical analysis. There are three topics in this research project. The first partcompares the performance and efficiency of the model parameters estimated byPF-EM and SD-PF methods under the given state-space model. The second partis to develop the particle filter algorithm when both the state variable and modelparameter are state-dependent. And then, this study applies the Copula toconstruct the dependency among the multivariate state variables and developsthe copula-particle filter algorithm to estimate parameters. The third part is anempirical study. Apply the proposed algorithm to the market data and multi-factorprice dynamic process. Then, use the particles representing the market riskfactors to analyze the correlation between the hidden variables and observedmarket data.