针对冷却塔生产过程中物料无法精确到各零部件的特点,提出从产品的角度,而非零部件的角度计算物料用量的成本计算方法。针对冷却塔零部件众多的特点,采用粗糙集属性归约约简属性,之后采用神经网络建立成本预估模型,极大减少了神经网络的输入维数,简化了神经网络的学习强度,为企业进行成本核算和预估提供了方便。 Aiming at the characteristics that the materials in the cooling tower can not be accurate to the parts in the production process of the cooling tower, the cost calculation method of calculating the material consumption from the product point of view rather than the parts is put forward. Aiming at many characteristics of cooling tower components, the rough set attribute reduction reduction attribute is adopted, and then the neural network is used to establish the cost estimation model, which greatly reduces the input dimension of neural network and simplifies the learning intensity of neural network. Cost accounting and forecasting provides a convenient.