Abstract

Oredr-4 Predictor-Corrector Time Series Smoothing Technique

C. I. Nwokike1 , G. O. Nwafor2 , B. B. Alhaji3 , K. M. Koko4 , C. Nwutara5 , H. E Chukwuma6 , S. N. Nwanneako7 , O. G. Onukwube8 , J. O. Onyeukwu

---

This work proposes a predictor-corrector time series smoothened values technique developed from the modification of two beautiful linear multistep numerical techniques. The study modified the 3-step explicit linear multistep technique also known as the order4 Adams-Bashforth technique and used same as an initial value predictor. For the corrector model, the study modified the 3-step implicit linear multistep technique also called the order-4 Adams-Moulton technique and designed a technique of fitting in the predicted values into the corrector model. The study also went further to develop an algorithm that performs a forward and backward smoothened values procedure before averaging these results to estimate the final smoothened value. The technique involves rigor and care but beautiful to perform and has a high performance evaluation. Keywords: Time series, Adams-Bashforth, Adam-Moulton, Oder-4, smoothing, Time Series.

PDF