基于改进支持向量机的作物叶水势软测量建模

Soft Sensor Modeling of Leaf Water Potential Based on Improved Support Vector Machine

摘要:

在标准最小二乘支持向量机(least square support vector machine,LSSVM)的基础上,利用改进的粒子群算法(improved particle swarm optimization,IPSO)来优化LSSVM模型参数,提出了基于IPSOLSSVM的软测量建模方法,建立了作物叶水势软测量模型.仿真结果表明,该方法比基本LSSVM和PSOLSSVM模型具有更高的精度,能够很好地预测作物叶水势信息.

关键词:作物叶水势; 软测量; 最小二乘支持向量机; 粒子群

Abstract:

Based on study on least square support vector machine(LSSVM),the paper presents an improved particle swarm optimization (IPSO) algorithm to select the parameters of LSSVM.The soft sensor modeling of the leaf water potential is established based on IPSOLSSVM.Simulation results indicate that the method based on IPSOLSSVM is of a higher accuracy than the basic LSSVM and LSSVM based on PSO,which can well predict leaf water potential.

Key words:leaf water potential; soft sensing; least square support vector machine(LSSVM); particle swarm optimization(PSO)