基于粒子群优化的神经网络训练算法在产品种类预测中的应用  

Category forecast application of neural network algorithm trained by particle swarm optimization

作　　者：高亮[1] 杨林[1] 周驰[1] 胡映兵[1] 

机构地区：[1]华中科技大学机械科学与工程学院工业工程系,湖北武汉430074

出　　处：《计算机集成制造系统》2006年第3期465-469,共5页Computer Integrated Manufacturing Systems

基　　金：国家自然科学基金资助项目(50305008);国家863/CIMS主题资助项目(2002AA413720)。~~

摘　　要：市场环境的变化导致产品更新换代加快,产品种类预测成为新的难题。传统的线性预测方法只能对产品需求的数量或价格等数值进行预测,而无法对产品的发展趋势和未来种类做出正确预测。通过对产品种类预测、数据挖掘和粒子群优化算法的研究,建立种类预测模型,利用基于粒子群优化的神经网络训练算法进行产品种类预测,并以手机为例进行预测,结果证明该方法是有效的。Market changes have shortened product life cycles, therefore category forecast has turning out to be a new difficult problem. But traditional linear forecasting methodology could only make predictions on numeric value such as requested quality and price, and it could not accurately predict product development trend and future categories. Category forecast model was constructed based on the study of product category prediction, data mining and Particle Swarm Optimization （PSO）. As the result, product category forecast could be conducted on neural network algorithm trained by PSO. Finally an example of category forecast on mobile telephone was used to verify the effectiveness of the proposed methodology.

关 键 词：数据挖掘 神经网络 种类预测 粒子群优化算法 

分 类 号：F274[经济管理—企业管理]