基于结构化神经网络挖掘的农产品产量预测方法

农产品市场的准确预测对指导农业生产、农产品流通和农产品消费有重要作用。本文提出了一种改进的结构化神经网络(ISNN),并基于ISNN构建了农产品产量预测模型;设计了一种优化性能更好的混合遗传算法(MGA),并采用MGA对ISNN预测模型进行训练。应用训练好的预测模型对某县最近10年的玉米总产品进行了预测分析。实验结果表明,该方法收敛速度快、学习能力强、预测精度较高、误差率较小。

The Forecasting Approach to Agriculture Products Based on an Improved Structure-Based Neural Network

The forecasting approach to agriculture products is very important to agricultural development.An improved structure-based neural network(ISNN) is proposed and applied to construct a forecasting model for agricultural products.An outstanding multiprogramming genetic algorithm(MGA) is designed and used to train the ISNN forecasting model.The proposed approach is evaluated by the agricultural products of the recent ten years.The experimental results suggest that the proposed approach has more favorable characteristics such as the convergence rate,learning ability,forecasting precision and estimation error.