Liu Ran
School of Software, North China University of Water Resources and Electric Power, 450011, Zhengzhou, China
Bu Hui
Department of Information Engineering, North China University of Water Resources and Electric Power, 450011, Zhengzhou, China
ABSTRACT
In view of the existing problems in grain yield, this study introduces a nonlinear combination forecasting model based on adaptive neural network, support vector machine and relevance vector machine. In this combination model, we take the prediction result of the previous single model as the input data of the next single model. In addition, a new method of determining the weight coefficient is proposed based on rough set theory. During the forecast operation, the core modules of the three models are organized together organically which improves the precision and stability of the combination model. Test results show that this method overcomes random and mutations of traditional methods and the mean absolute error of the prediction result is lower than the traditional model.
PDF References Citation
How to cite this article
Liu Ran and Bu Hui, 2013. Study on Nonlinear Combination Forecasting Model for Grain Yield. Information Technology Journal, 12: 4666-4672.
DOI: 10.3923/itj.2013.4666.4672
URL: https://scialert.net/abstract/?doi=itj.2013.4666.4672
DOI: 10.3923/itj.2013.4666.4672
URL: https://scialert.net/abstract/?doi=itj.2013.4666.4672
REFERENCES
- Tipping, M.E., 2001. Sparse Bayesian learning and the relevance vector machine. J. Machine Learn. Res., 1: 211-244.
Direct Link - Wu, H.X., H.X. Dong and J.Q. Su, 2008. 3D reconstruction from section plane views based on Self-adaptive neural network. Proceedings of the 2nd Intelligent Information Technology Application, Volume 3, December 20-22, 2008, Shanghai, pp: 84-88.
CrossRef - Wu, H. and F. Wang, 2010. RVM-BASED ore grade forecasting model and its application. Proceedings of the 3rd IEEE International Conference on Computer Science and Information Technology, Volume 3, July 9-11, 2010, Chengdu, pp: 449-451.
CrossRef