Description
Filter Methods for Parameter Estimation in Linear and Non Linear Regression Models.
Description
We present a method based on filtering algorithms to estimate the parameters of linear, i.e. the coefficients and the variance of the error term. The proposed algorithms make use of Particle Filters following Ristic, B., Arulampalam, S., Gordon, N. (2004, ISBN: 158053631X) resampling methods. Parameters of logistic regression models are also estimated using an evolutionary particle filter method.