优化BP神经网络算法在油茶产量预测中的应用

针对标准的BP神经网络模型对我国南方的重要木本油料树种油茶产量进行预测过程中存在的缺陷,采用相对误差逐步优化回溯算法在迭代过程的参数,使原始的BP神经网算法在运算精度和计算速度上均得到显著的提高,即一种改进的回溯算法优化BP神经网络预测模型。经过试验效验与仿真证明,得到的改进算法不仅能提高油茶产量预测的收敛速度,而且在油茶产量预测精度上也有很大的提高。优化后的BP神经网络算法为多要素因子之间相互影响事件的结果预测,提供了新的设计思路和更好的解决方法。

Optimized BP Neural Network Algorithm in Predicting Camellia oleifera Yield

For the defects in predicting Camellia oleifera yield, we used standard BP neural network model and relative error to optimize the parameters of back-track algorithm in iterative procedure gradually, and improved the calculation precision and calculation speed of original BP neural network algorithm significantly.Through test calibration and simulation proof, the improved algorithm improved not only the convergence rate for predicting C.oleifera yield, but also the prediction pre-cision greatly.