基于混合智能算法的支持向量机的粮食产量预测  

Grain yield prediction of support vector machine based on hybrid intelligent algorithm

作　　者：高心怡 韩飞[1] GAO Xinyi;HAN Fei(School of Computer Science & Communication Engineering, Jiangsu University, Zhenjiang, Jiangsu 212013, China)

机构地区：[1]江苏大学计算机与通信工程学院,江苏镇江212013

出　　处：《江苏大学学报（自然科学版）》2020年第3期301-306,共6页Journal of Jiangsu University：Natural Science Edition

基　　金：国家自然科学基金资助项目(61572241,61976108);国家重点研发计划项目(2017YFC0806600);江苏省“六大人才高峰计划”高层次人才计划项目(2015DZXX024);江苏省“333工程”第三层次人才计划项目((2016)III-0845)。

摘　　要：考虑到粮食产量的非线性,提出一种基于混合智能算法的支持向量机的预测模型,针对粒子群算法(PSO)易于陷入局部最优的特点,融合改进的粒子群算法(GAPSO)以及人工鱼群算法(AFSA)得到混合智能算法(GAPSO-AFSA),通过种群自身内部的变异交叉以及族外竞争机制,使得目标函数值向全局最优解快速收敛,提高了算法的全局搜索能力,最终得到支持向量机的最优参数组合.将支持向量机预测模型用于中国粮食产量预测中,通过试验验证了模型的正确性,并证实了该预测模型拥有较好的预测效果.Considering the non-linearity of grain yield,the prediction model of support vector machine was proposed based on hybrid intelligent algorithm.To solve the problem that particle swarm optimization(PSO)was easy to fall into a local optimum,the hybrid intelligent algorithm(GAPSO-AFSA)was proposed by combining the improved PSO and the artificial fish swarm algorithm(AFSA).The objective function value was quickly converged to the global optimal solution by the mutation cross within the swarm and the external competition mechanism.The global search ability of the algorithm was improved to obtain the optimal parameter combination of support vector machine.The support vector machine prediction model was used to predict Chinese grain yield,and the model correctness was verified by the experiments.The results show that the prediction model has good prediction result.

关 键 词：粮食产量 支持向量机 混合智能算法 粒子群算法 人工鱼群算法 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]