Abstract: To solve problems of subjective experience, delay and inaccuracy caused by equalization in traditional standard cost determination, an activity-based standard cost setting method based on neighborhood rough set, random distribution and Support Vector Machine(SVM)was proposed. The attribute factors influencing activity cost was determined by using rough set theory to reduce. According to attribute reduction sets, the distribution of stochastic attributes was sampled and analyzed, and the confidence interval was obtained. The regression model of attributes and coset was built by SVM. By using confidence interval of stochastic attributes, the interval range of activity standard cost was set, and the activity standard cost was determined finally. The new standard cost regression model was built to predict the standard cost of small batch and new product. The implementation steps of this method were explained by its application in iron and steel enterprise, and the validity of this method was verified.

Key words: standard cost setting, rough set, random distribution, support vector machine, activity

摘要： 为解决传统标准成本制定过程中存在的主观经验性、不及时性和平均化带来的误差等问题，提出一种基于邻域粗糙集、随机分布理论和支持向量机的作业标准成本制定方法。首先通过邻域粗糙集理论进行属性约简，确定影响作业成本的属性因素，依据约简属性集从实例库中抽样并分析随机属性的分布规律，确定置信区间；然后用支持向量机建立随机属性和作业成本的回归模型，利用随机属性的置信区间确定作业标准成本的区间范围并最终确定作业标准成本值；最后建立作业标准成本的回归模型，预测小批量或新产品的标准成本。通过钢铁企业的应用实例，给出了该方法的具体实现过程，并验证了方法的有效性。

关键词: 标准成本制定, 粗糙集, 随机分布, 支持向量机, 作业