基于支持向量机的水稻净光合速率预测模型  

Prediction model on net photosynthetic rate of rice based on support vector machine

作　　者：刘彤 武海巍[2] 曹忠 

机构地区：[1]吉农水稻高新科技发展有限责任公司,吉林公主岭136100 [2]北华大学电气信息工程学院,吉林吉林132021

出　　处：《中国农机化学报》2016年第9期151-153,共3页Journal of Chinese Agricultural Mechanization

基　　金：国家留学基金(201408220077)

摘　　要：利用线性核函数、多项式核函数、径向基核函数和S型核函数,通过粒子群算法优化支持向量机中的惩罚参数c和gamma值,使用主成分分析对影响水稻净光合速率的影响因子进行特征提取,建立水稻净光合速率的支持向量机预测模型。试验结果表明,通过粒子群算法设定惩罚参数c为10,gamma值为0.904,通过训练集的训练,得到P_RBF预测模型,其对水稻净光合速率的预测效果最佳,预测精度达83%以上。This paper constructs a prediction model on net photosynthetic rate of rice based on linear kernel function,polynomial kernel function,radial basis function kernel function,sigmoid kernel function,principal component analysis which extracts the factors affecting the net photosynthetic rate of rice and the penalty parameters c and gammavalues optimized by particle swarm algorithm.The test results are as follows.The prediction model named as P_RBF model in which penalty parameters cis 10 and gammavalues is 0.904 optimized by particle swarm algorithm is the best prediction model and the prediction accuracy is up 83% on net photosynthetic rate of rice.

关 键 词：粒子群算法 支持向量机 水稻净光合速率 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]