参数优化支持向量机的农业大棚温室温度预测模型

利用支持向量机核函数linear,polynomial,radial basis function和sigmoid,通过粒子群算法对惩罚参数c和gamma寻优,建立农业大棚温室温度预测模型.试验结果表明:通过粒子群算法设定惩罚参数c为14.392,gamma为0.01时,得到的P_RBF预测模型对由24个测试时间所测数据组成的训练集拟合程度达90.849%,对加入随机影响因子的由5个测试时间所测数据组成的预测集拟合程度达90.545%,显示该预测模型具备相当的鲁棒性;P_RBF模型对温室内温度预测具备相当的可靠性,可以准确预测温室内温度变化趋势,解决温室控制系统中温度难以预测的问题.

Prediction Model on Agricultural Greenhouse Temperature Based on Support Vector Machine with Parameter Optimization

The paper constructs a prediction model on the temperature in the agricultural greenhouse based on kernel function linear,polynomial,radial basis function,sigmoid and the penalty parameters c and gamma values optimized by Particle Swarm Algorithm.The paper constructs the prediction model named as P_RBF model in which penalty parameters c is 14.392 and gamma value is 0.01 optimized by Particle Swarm Algorithm,on which the prediction accuracy of the train set consisted of 24 times,measured data is 90.849%.Based on P_RBF model,the prediction accuracy of the prediction set consisted of 5 times,measured data is 90.545%,which shows the model is robust.P_RBF model indicates that the prediction result is reliable on the temperature in the greenhouse and the temperature change trend may be accurately predicted,and solve the key factor on the greenhouse control being intelligent,which is difficult to predict the temperature.