Abstract
Median smoother of even window size is obtained by averaging the two middle points of arranged sequence using arithmetic mean. Some adjustments were proposed by substituting the algorithm of arithmetic mean using quadratic, geometric, harmonic and contraharmonic mean. The modified running median was appended in 4253HT algorithm. This paper is mainly to assess the performances of proposed adjustment via simulation study. The results show that modified 4253HT using contra harmonic mean performed the best in extracting signal from heavy noise. The extracted signal was then used for forecasting by applying seasonal ARIMA algorithm. Forecasting by smoothed values was found to produce better forecast than forecasting involving actual values which contained heavy noise.
Metadata
Item Type: | Article |
---|---|
Creators: | Creators Email / ID Num. Khairol Azmi, Nurul Nisa’ nisa@tmsk.uitm.edu.my Adam, Mohd Bakri UNSPECIFIED Ali, Norhaslinda UNSPECIFIED |
Subjects: | Q Science > QA Mathematics > Instruments and machines > Electronic Computers. Computer Science > Algorithms |
Divisions: | Universiti Teknologi MARA, Terengganu > Dungun Campus |
Journal or Publication Title: | e-Academia Journal |
UiTM Journal Collections: | UiTM Journal > e-Academia Journal (e-AJ) |
ISSN: | 2289 - 6589 |
Volume: | 8 |
Page Range: | pp. 78-84 |
Keywords: | Compound Smoother, Running Median, Non-Linear, 4253HT, Signal, Noise |
Date: | 2018 |
URI: | https://ir.uitm.edu.my/id/eprint/86788 |