On some techniques of selecting spline smoothing parameters for a correlated dataset with autocorrelation structure in the residual

Residuals are minimized in a correlated dataset by selecting smoothing parameter with optimum performance the spline. The selection methods utilized this study include Generalized Maximum Likelihood (GML), Cross-Validation (GCV), Unbiased Risk (UBR), and Proposed Smoothing Method (PSM). aim of is to compare ability four for autocorrelation structure error term. To achieve purpose, Monte-Carlo simulation was conducted utilizing program written R-4.2.2. were evaluated using predictive Mean Squared Error (PMSE). Findings from indicated that GCV GML mostly affected presence auto correlation residual therefore had an asymptotically similar behavioural pattern. estimators conformed asymptotic properties considered; noticed all sample sizes at parameters. result also showed that; most consistent efficient among spline considered based on size autocorrelated proposed method (PSM) because it does not undersmooth relative other especially small medium 50 100.