组合式神经网络企业产品销售量预测方法

Modular method of Artifical Neural Network Forecast on sales volume of enterprise product

采用回归分析法,去逼近预测模型中具有时间增长趋势的离散数据,用其逼近误差进行人工神经网络学习和训练,输出预测误差,然后对回归模型进行动态补偿,以实现对企业产品销售量进行组合预测。其预测结果比单一的回归分析法和线性规划法计算精度要高,表明组合式人工神经网络对预测企业产品销售量具有良好的适应性。

This paper introduced the modular method of artifical neural network and regress analysis forecast on sales volume of enterprise product. The method is better than single method of regress analysis and linear programming by an example demonstrated ,the results show that the modular artifical neural network forecast is very efficient.