A new hybrid model for time-series prediction

Author

Pan, Feng ; Xia, Min ; Bai, En Jian

Author_Institution

Coll. of Inf. Sci. & Technol., Donghua Univ., Shanghai, China

fYear

2009

fDate

15-17 June 2009

Firstpage

281

Lastpage

286

Abstract

This paper proposed a new hybrid model in order to increase time series prediction accuracy. This hybrid model considers the routine time prediction technique like AR, ANN or any others as atomic building block. A linear hybrid technique is used to combine their forecast result into the final result. The hybrid algorithm was tested against three different kinds of time series data. Experiments results showed the effectiveness of the proposed hybrid model.

Keywords

autoregressive moving average processes; forecasting theory; time series; artificial neural nets; autoregression process; linear hybrid technique; time-series prediction; Accuracy; Artificial neural networks; Educational institutions; History; Information science; Neural networks; Paper technology; Predictive models; Smoothing methods; Testing; ANN; Autoregression; Hybrid; Prediction; Time series;

fLanguage

English

Publisher

ieee

Conference_Titel

Cognitive Informatics, 2009. ICCI '09. 8th IEEE International Conference on

Conference_Location

Kowloon, Hong Kong

Print_ISBN

978-1-4244-4642-1

Type

conf

DOI

10.1109/COGINF.2009.5250728

Filename

5250728

Link To Document

https://search.isc.ac/dl/search/defaultta.aspx?DTC=49&DC=3391625