DocumentCode :
3444601
Title :
Algorithm of fertilization knowledge model based on fuzzy-neural network
Author :
Yang, Yushu ; Ran, Cao ; Wenyang, Liu
Author_Institution :
Eng. Coll., Northeast Agric. Univ., Haerbin, China
Volume :
3
fYear :
2010
fDate :
29-31 Oct. 2010
Firstpage :
40
Lastpage :
43
Abstract :
When traditional mathematical statistic method is used to build fertilization model, the structure design and factor choosing largely rely on prior knowledge of experts in the field. Consequently, the result is somewhat casual and subjective. It is very difficult to settle this universal nonlinear and uncertain problems in fertilization. A new network for fuzzy-neural system which is easy to distill the fuzzy rules is proposed. The network structure is adjusted by FBP(Fuzzy Back Propagation) learning algorithm to acquire network parameters and variable weights of the initial fertilization model. For an optimal fertilization model comes into being, IIP(Improved iterative pruning) algorithm which is applied can lessen the network structure and reduce the complexity of compute to speed up the respond rate of output. Soybean experiments in different plants and years show that this approach can build very accurate model without any prior knowledge, which contributes as theoretical in countryside.
Keywords :
backpropagation; crops; fertilisers; fuzzy neural nets; iterative methods; statistics; fertilization knowledge model; fuzzy back propagation learning algorithm; fuzzy rules; fuzzy-neural network; improved iterative pruning algorithm; mathematical statistic method; network structure; soybean experiments; Artificial neural networks; Cognition; Computational modeling; FMLP(Fuzzy Neural Network); IP(iterative pruning) algorithm; Network structure; fertilizer knowledge model;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Intelligent Computing and Intelligent Systems (ICIS), 2010 IEEE International Conference on
Conference_Location :
Xiamen
Print_ISBN :
978-1-4244-6582-8
Type :
conf
DOI :
10.1109/ICICISYS.2010.5658551
Filename :
5658551
Link To Document :
بازگشت