THE KALMAN FILTER MODEL AND BAYESIAN OUTLIER DETECTION FOR TIME-SERIES ANALYSIS OF BOD DATA

The purpose of this paper is to fit a trigonometric time series model to a biochemical oxygen demand (BOD) data set using the Kalman filter approach to allow estimates of the parameters to be updated recursively with each new observation. In addition, we analyse the data set for outliers by computing the prior and posterior probabilities for all observations. The smoothing equations result in a better fit of the proposed model than the model of Papadopoulos et al. (Ecological Modelling, 55(1991): 57-65) based on the same data, in the sense that it reduces the mean squared error by more than half.