Dynamic linear modelling of stock data

The application of the dynamic linear model to the forecasting of stock data will be discussed. Although stationarity and linearity are often assumed for convenience in the analysis of time series, nonstationarity and nonlinearity are often more realistic. The analysis of nonlinear time series models considered here is facilitated through use of the Bayesian Analysis of Time Series package. This allows the assumption of time-independent coefficients to be relaxed.