Earticle

초록

영어

Time series prediction is a challenging research area with broad application prospects in machine learning. Accurate prediction on a time series’ value can provide important information for the decision-makers. In the literature, many works were reported to extend different architecture of artificial neural networks to work with time series prediction. However, most of the work only considered the target time series itself, while neglecting the impact of the relevant time series. In this paper we proposed a novel method MANNP that makes use of multiple artificial neural networks to conduct the time series prediction. The proposed method creates time series model and forecast time series. To verify the effectiveness of the proposed method, we apply MANNP to a shipping price index time series prediction. The experimental results show that this method can improve accuracy of prediction when compared with traditional methods.

목차

Abstract
 1. Introduction
 2. Related Work
  2.1. Linear Model
  2.2. Nonlinear Model
 3. Multiple Artificial Neural Network Prediction (MANNP) Model
  3.1. Artificial Neural Network
  3.2. Description of MANNP Model
  3.3. Multi-step Forecasting
 4. Experiment
  4.1. Experiment Settings
  4.2. Experiment Result
 5. Conclusion
 Acknowledgements
 References

키워드

저자

• Fei Li [ College of Information Engineering, Shanghai Maritime University, 201306 ]
• Jin Liu [ College of Information Engineering, Shanghai Maritime University, 201306 ]
• Lei Kong [ The third branch of Shanghai Administration Institute, 200233, ] Corresponding author

참고문헌

발행기관

간행물

표지보기 관심저널 등록

• 간행물명

  International Journal of Hybrid Information Technology

• 간기

  격월간

• pISSN

  1738-9968

• 수록기간

  2008~2016

• 십진분류

  KDC 505 DDC 605