摘要
随着中国对外开放进程的加快,高科技产业迎来了更加严峻的挑战,只注重新产品开发或技术开发已不足以满足企业发展的需要。企业应该关注其项目成本管理能力,尤其需要选择合理的方法预测项目成本。本文通过探究作业成本法和机器学习各自的特点,尝试将两者结合起来,确定一种符合企业项目成本特点的预测方法。通过本文案例公司的数据测试,作业成本法和BP神经网络结合构建的成本分析模型误差控制在4%以内,表明此法在成本预测方面有一定的适用性,可用于高科技企业项目成本预测。
By exploring the characteristics of activity-based costing and machine learning,this paper attempts to build a prediction method that is consistent with the cost characteristics of the high-tech enterprise.The testing results of the case company show that the cost analysis model constructed by the combination of activitybased costing and BP neural network has an error rate within 4%,which indicates that this method has certain applicability in cost prediction and can be used for project cost forecasting in high-tech enterprises.
作者
商婷婷
陈亚盛
Shang Tingting;Chen Yasheng
出处
《管理会计研究》
2019年第5期76-86,88,共12页
MANAGEMENT ACCOUNTING STUDIES
