Semiparametric regression based on fourier series for longitudinal data with Weighted Lest Square (WLS) optimization

Abstract

Regression modelling is one Statistical methods that be used to investigate the relationship between predictor variable and response variable. In regression modelling can be estimated with three approaches, such as parametric, nonparametric and semiparametric regression. In this case, we concentrated to elaborate semiparametric regression. Semiparametric regression consists of parametric and nonparametric component. This research examined semiparametric regression model with Fourier series estimator for longitudinal data. By minimizing Weighted Least Square (WLS), the Fourier series estimator depends on the oscillation parameter. The result is the estimator for parameter and curve regression, that be used to model with real data. The optimal model is selected based on minimum Generalized Cross Validation (GCV) which affects the small value of Mean Square Error (MSE) and high determination coefficient so that the model can be used further as estimation and prediction.