Applying particle filter to genetic regulatory networks identification

The genetic regulatory networks (GRNs) identification problem is considered in this paper. Since it prefers to capture the characteristic behavior of GRN, which seems natural to describe as a hybrid system, a piecewise affine type model is often used as a simple model of GRNs in recent years. A particle filter is introduced to identify a piecewise ARX model, where the system mode and parameters should be estimated simultaneously. The system mode is estimated by maximum a posteriori probability and the unknown parameters are estimated by particles. A numerical simulation studies are carried out by using the carbon starvation response model of the bacterium Escherichia coli, and the proposed method is able to estimate the system mode with high accuracy.