【摘要】 家禽孵化是一个复杂的生物学过程,其过程参数的系统辨识对于实现后续精确控制有着重要意义。由于孵化的过程参数具有数据量获取困难、样本小的特点,而SVM是经过严格的数学推理而得,在解决小样本、非线性、过学习与欠学习、局部极小点问题中表现出许多特有的优势。为此,结合SVM的特点,将其应用于孵化过程。实验结果表明,采用SVM辨识方法对孵化过程控制系统的温度和湿度具有较好的预测效果。更多还原

【Abstract】 Poultry hatching is a complex biological process;it is of great significance for the achievement of follow-up precise control to identify the process parameters.Based on the hatching process parameters which is difficult to get and has small sample,it is necessary to find an advanced modeling method.SVM is applied to the hatching.SVM is derived through rigorous mathematical reasoning which has many unique advantages in resolving the small sample,non-linear sex,over-learning and less learning,local minimum problems.Experimental results show that the SVM identification method has good prediction on the hatching temperature and humidity.更多还原