Estimation of a Noisy Discrete-Time Step Function: Bayes and Empirical Bayes Approaches

摘要: Consider the problem of estimating, in a Bayesian framework and presence additive Gaussian noise, signal which is step function. Best linear estimates Bayes are derived, evaluated compared. A characterization presented. This has an intuitive interpretation also provides way to compute with number operations order $T^3$ where $T$ fixed time span. An approximation proposed reduces total $T$. The results applied case model fails be satisfied using empirical approach.