Reconstruction of Time Series Data with Missing Values

Missing data are a part of almüst all research and it must be decided how to dea! with it from time to time. Missing data creates several problems İn many applications which depend on good access to accurated data. Conventiona! methods for missing data, like listwİse deletion or regression imputation, are prone to three serİolis problems: Inefficİent use of the available information, leading to low power and Type II errüfS. Biased estimates of standard errüfS, leading to İncorrect p-values. Biased parameter estimates, due to failure to adjust for selectivity İn missing data. In this study, we propose a new algorithrn to predict missing values of a given time series using Radial Basis Fwıctions.