Title :
Grain Security Risk Level Prediction Using ANFIS
Author :
Kadir, Muhd Khairulzaman Abdul ; Hines, Evor L. ; Arof, Saharul ; Illiescu, D. ; Leeson, Mark ; Dowler, Elizabeth ; Collier, Rosemary ; Napier, Richard ; Kefaya, Qaddoum ; Ghafari, R.
Author_Institution :
Sch. of Eng., Univ. of Warwick, Coventry, UK
Abstract :
Food security is a major worldwide issue nowadays. One of the supporting indicators of the food security level is the trend of the global agriculture output per capita. In this study, grain data from China between 1997 and 2007 is used as a means to indicate the level of grain security. The inputs for this study are based on 3 categories, productive indexes, consumptive indexes, disaster indexes, in total there are 11 input indexes to the system with 2 membership functions (MFs) for each input. The system output is the level of the grain security, where the target data is based on a previous study of China grain security level. We use an Adaptive Neuro-Fuzzy Inference System (ANFIS) to predict the grain security level. In this case data pre-processing with the Principal Component Analysis (PCA) technique was used to reduce inputs to 6 to avoid too many rule parameters which would affect the optimization performance of the model. A Multi-Layer-Perceptron-Neural-Network (MLP-NN) model is used to compare with the performance of ANFIS. The result of this study shows that the resulting regression value in the case of ANFIS is around 0.99 which is better than that for the NN, which is around 0.60. Hence the ANFIS model is shown to offer better predictor of grain security level. It may also be an attractive method to explore further as a means for food security early warning monitoring systems.
Keywords :
agricultural products; agriculture; food safety; fuzzy reasoning; multilayer perceptrons; principal component analysis; regression analysis; ANFIS; adaptive neuro-fuzzy inference system; consumptive index; disaster index; food security; food security early warning monitoring systems; global agriculture output per capita; grain security risk level prediction; membership functions; multilayer-perceptron-neural-network model; principal component analysis technique; productive index; regression value; Data models; Educational institutions; Indexes; Mathematical model; Predictive models; Principal component analysis; Security; ANFIS; Neural Network; food security; grain security; risk level;
Conference_Titel :
Computational Intelligence, Modelling and Simulation (CIMSiM), 2011 Third International Conference on
Conference_Location :
Langkawi
Print_ISBN :
978-1-4577-1797-0
DOI :
10.1109/CIMSim.2011.27