Water resources security assessment based on support vector machine
Wei Ma · Ziran zaihai xuebao · 2011
Based on statistical learning theory,support vector machine(SVM) can transform the learning process into a convex quadratic planning problem to get a global optimization by using the rule of structure risk minimization,which is appropriate to solving small sample,nonlinear classification and regression. Based on the concept of water resources security,representative indicators were selected for the water resources security assessment indicator system.Water resources assessment model based on support vector machine was established.Water resources security standards were divided into five grades,named good,safe,critical,not safe and dangerous.Sample sets were formed by stochastic method according to water resources security standards and their grade values.180 samples were used for training to construct 5 two-classification support vector classifiers.Twenty samples were used for testing and all of which can be classified correctly.Applying the model to 11 cities in Shanxi Province,the results show that the algorithm is reasonable and feasible.