Improved nonlinear prediction method

Nur Hamiza Adenan, Mohd. Salmi Md. Noorani

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The analysis and prediction of time series data have been addressed by researchers. Many techniques have been developed to be applied in various areas, such as weather forecasting, financial markets and hydrological phenomena involving data that are contaminated by noise. Therefore, various techniques to improve the method have been introduced to analyze and predict time series data. In respect of the importance of analysis and the accuracy of the prediction result, a study was undertaken to test the effectiveness of the improved nonlinear prediction method for data that contain noise. The improved nonlinear prediction method involves the formation of composite serial data based on the successive differences of the time series. Then, the phase space reconstruction was performed on the composite data (one-dimensional) to reconstruct a number of space dimensions. Finally the local linear approximation method was employed to make a prediction based on the phase space. This improved method was tested with data series Logistics that contain 0%, 5%, 10%, 20% and 30% of noise. The results show that by using the improved method, the predictions were found to be in close agreement with the observed ones. The correlation coefficient was close to one when the improved method was applied on data with up to 10% noise. Thus, an improvement to analyze data with noise without involving any noise reduction method was introduced to predict the time series data.

Original languageEnglish
Title of host publicationAIP Conference Proceedings
PublisherAmerican Institute of Physics Inc.
Pages94-99
Number of pages6
Volume1602
ISBN (Print)9780735412361
DOIs
Publication statusPublished - 2014
Event3rd International Conference on Mathematical Sciences, ICMS 2013 - Kuala Lumpur
Duration: 17 Dec 201319 Dec 2013

Other

Other3rd International Conference on Mathematical Sciences, ICMS 2013
CityKuala Lumpur
Period17/12/1319/12/13

Fingerprint

predictions
weather forecasting
composite materials
logistics
noise reduction
correlation coefficients
approximation

Keywords

  • Improved method
  • Noise
  • Nonlinear prediction method
  • Time series data

ASJC Scopus subject areas

  • Physics and Astronomy(all)

Cite this

Adenan, N. H., & Md. Noorani, M. S. (2014). Improved nonlinear prediction method. In AIP Conference Proceedings (Vol. 1602, pp. 94-99). American Institute of Physics Inc.. https://doi.org/10.1063/1.4882472

Improved nonlinear prediction method. / Adenan, Nur Hamiza; Md. Noorani, Mohd. Salmi.

AIP Conference Proceedings. Vol. 1602 American Institute of Physics Inc., 2014. p. 94-99.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Adenan, NH & Md. Noorani, MS 2014, Improved nonlinear prediction method. in AIP Conference Proceedings. vol. 1602, American Institute of Physics Inc., pp. 94-99, 3rd International Conference on Mathematical Sciences, ICMS 2013, Kuala Lumpur, 17/12/13. https://doi.org/10.1063/1.4882472
Adenan NH, Md. Noorani MS. Improved nonlinear prediction method. In AIP Conference Proceedings. Vol. 1602. American Institute of Physics Inc. 2014. p. 94-99 https://doi.org/10.1063/1.4882472
Adenan, Nur Hamiza ; Md. Noorani, Mohd. Salmi. / Improved nonlinear prediction method. AIP Conference Proceedings. Vol. 1602 American Institute of Physics Inc., 2014. pp. 94-99
@inproceedings{816f2cf5dcc246a69e63f1cb39a7d201,
title = "Improved nonlinear prediction method",
abstract = "The analysis and prediction of time series data have been addressed by researchers. Many techniques have been developed to be applied in various areas, such as weather forecasting, financial markets and hydrological phenomena involving data that are contaminated by noise. Therefore, various techniques to improve the method have been introduced to analyze and predict time series data. In respect of the importance of analysis and the accuracy of the prediction result, a study was undertaken to test the effectiveness of the improved nonlinear prediction method for data that contain noise. The improved nonlinear prediction method involves the formation of composite serial data based on the successive differences of the time series. Then, the phase space reconstruction was performed on the composite data (one-dimensional) to reconstruct a number of space dimensions. Finally the local linear approximation method was employed to make a prediction based on the phase space. This improved method was tested with data series Logistics that contain 0{\%}, 5{\%}, 10{\%}, 20{\%} and 30{\%} of noise. The results show that by using the improved method, the predictions were found to be in close agreement with the observed ones. The correlation coefficient was close to one when the improved method was applied on data with up to 10{\%} noise. Thus, an improvement to analyze data with noise without involving any noise reduction method was introduced to predict the time series data.",
keywords = "Improved method, Noise, Nonlinear prediction method, Time series data",
author = "Adenan, {Nur Hamiza} and {Md. Noorani}, {Mohd. Salmi}",
year = "2014",
doi = "10.1063/1.4882472",
language = "English",
isbn = "9780735412361",
volume = "1602",
pages = "94--99",
booktitle = "AIP Conference Proceedings",
publisher = "American Institute of Physics Inc.",

}

TY - GEN

T1 - Improved nonlinear prediction method

AU - Adenan, Nur Hamiza

AU - Md. Noorani, Mohd. Salmi

PY - 2014

Y1 - 2014

N2 - The analysis and prediction of time series data have been addressed by researchers. Many techniques have been developed to be applied in various areas, such as weather forecasting, financial markets and hydrological phenomena involving data that are contaminated by noise. Therefore, various techniques to improve the method have been introduced to analyze and predict time series data. In respect of the importance of analysis and the accuracy of the prediction result, a study was undertaken to test the effectiveness of the improved nonlinear prediction method for data that contain noise. The improved nonlinear prediction method involves the formation of composite serial data based on the successive differences of the time series. Then, the phase space reconstruction was performed on the composite data (one-dimensional) to reconstruct a number of space dimensions. Finally the local linear approximation method was employed to make a prediction based on the phase space. This improved method was tested with data series Logistics that contain 0%, 5%, 10%, 20% and 30% of noise. The results show that by using the improved method, the predictions were found to be in close agreement with the observed ones. The correlation coefficient was close to one when the improved method was applied on data with up to 10% noise. Thus, an improvement to analyze data with noise without involving any noise reduction method was introduced to predict the time series data.

AB - The analysis and prediction of time series data have been addressed by researchers. Many techniques have been developed to be applied in various areas, such as weather forecasting, financial markets and hydrological phenomena involving data that are contaminated by noise. Therefore, various techniques to improve the method have been introduced to analyze and predict time series data. In respect of the importance of analysis and the accuracy of the prediction result, a study was undertaken to test the effectiveness of the improved nonlinear prediction method for data that contain noise. The improved nonlinear prediction method involves the formation of composite serial data based on the successive differences of the time series. Then, the phase space reconstruction was performed on the composite data (one-dimensional) to reconstruct a number of space dimensions. Finally the local linear approximation method was employed to make a prediction based on the phase space. This improved method was tested with data series Logistics that contain 0%, 5%, 10%, 20% and 30% of noise. The results show that by using the improved method, the predictions were found to be in close agreement with the observed ones. The correlation coefficient was close to one when the improved method was applied on data with up to 10% noise. Thus, an improvement to analyze data with noise without involving any noise reduction method was introduced to predict the time series data.

KW - Improved method

KW - Noise

KW - Nonlinear prediction method

KW - Time series data

UR - http://www.scopus.com/inward/record.url?scp=84904117055&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84904117055&partnerID=8YFLogxK

U2 - 10.1063/1.4882472

DO - 10.1063/1.4882472

M3 - Conference contribution

SN - 9780735412361

VL - 1602

SP - 94

EP - 99

BT - AIP Conference Proceedings

PB - American Institute of Physics Inc.

ER -