中文摘要：

指出了目前参数成本估算对生产耗用资源市场价格的波动考虑不充分、所用历史成本数据缺乏真实性与可比性、对成本形成原因不能识别和反馈控制等缺陷．依据作业成本法“产品消耗作业，作业消耗资源”的思想，建立了从产品属性确认开始，经过作业动因、资源动因、再到作业成本、产品成本的参数成本估算模型．给出了运用主成分分析对产品主要特性指标降维的方法，运用径向基函数神经网络构建了作业估算关系（AER），并进行了案例分析．

英文摘要：

This paper points out the three limitations in current parametric cost estimation, which the price fluctuation of resources market is not adequately considered, the absence of the facticity and the comparability of historical cost data, and the reasons of cost forming can not be identified, fed back and controlled. According to the ideas from Activity-based costing that ＂products consume activities and activities consume resources＂, a cost estimation model that starts from obtaining product characteristics, to activity drivers, to resource drivers, to activity cost and product cost is established. A product characteristic parameter dimension debasing method based on Principal Component Analysis is discussed. The Radical Based Function （RBF） neural network is adopted to establish the Activity Estimation Relationships （AER） and an example of AER is given.