Hanning’s Performance in 4253HT Smoother on Four Signals of Different Noise Levels

Abstract

Smoothing is a method of data analysis that aims to provide well-defined patterns or signals by removing noise or unstructured patterns from data sets. To obtain clean and smooth data, Hanning is applied as one of the important components in smoothing. However, Hanning is not resistant to outliers. Therefore, this study aims to determine the best type of Hanning that is able to obtain the greatest performance of 4253HT smoother in signal recovery. Functions of Linear, Complex Sinusoidal, Custom Pulse Train, and Sawtooth signals corrupted with five levels of contaminated normal noise were used as signals in the smoothing process. All signals were applied to assess three different Hanning types, which were Tukey, Husain and Shitan. Besides, a Root Mean Square Error (RMSE) was used as an evaluator to determine and assess the performance of 4253HT smoother when utilizing an alternative Hanning. Based on the overall performance of 4253HT smoother, Husain Hanning presented the best outcome and worked most efficiently at all levels of noise except at the lowest noise (10%), which Tukey Hanning executed better. The findings of this study could benefit other researchers to decide the best Hanning to be used before performing forecasting and further analysis to improve the accuracy of predicting.