【摘要】 企业产品销售额的分布具有动态的特性,循环神经网络能够直接地反映系统动态、非线性的特性,因此,考虑用循环神经网络来预测企业产品的销售额,同时为了避免训练陷入区域极小值,用遗传算法来改进网络的权值算法。实验表明基于遗传算法的循环神经网络可以预测出销售额的变化趋势,能取得不错的预测效果。更多还原

【Abstract】 Recurrent neural networks can direct reflection the dynamic property of system,the same saleroom has dynamic property.The monthly sale of market can be predicted with recurrent neural networks.Therefore Genetic Algorithm is adopted to optimize weights of network in order to avoid being trapped into a poor local minima.This software is finished with java.更多还原