Comparison of the non-parametric smoothing estimator with the optimal estimators We consider three estimators: — Kalman estimator S k (16) with risk R K , — optimal backward interpolatioñinterpolatioñ˜ interpolatioñ˜S k (18) with the risk R OI , — non-parametric interpolation (19) with the risk R N I . Because R OI R K and R OI R N I , for convenience of comparison, let us introduce the relative errors in percentage:  

Context in source publication

Context 1

... relative error shows how much an estimator is better or worse than another one. The simulation results are presented in Fig.1 by n = 1000, σ 2 = 2, a = 0.7, b = 1, A = B = 1, τ = 1. ...