基于机器学习的商品报价模型  

Pricing Model Based on Machine Learning

作　　者：高晨芳 陈斌全 陈亚盛 Gao Chenfang;Chen Binquan;Chen Yasheng

机构地区：[1]厦门大学管理学院 [2]厦门正航软件科技有限公司

出　　处：《管理会计研究》2022年第4期16-26,共11页MANAGEMENT ACCOUNTING STUDIES

摘　　要：定价决策是企业实现利润最大化的直接途径,也是管理会计领域广泛关注的话题。本文以一家电子商贸公司为样本,运用机器学习的方法,通过利用企业积累的销售数据和从企业外部获取的大数据,构建客户订单价格估算模型。本文构建的价格估算模型共选取了包括宏观环境、客户属性、产品属性、需求属性和供应属性共5个维度的17个价格影响因素。模型使用12109笔历史销售记录作为训练集。经过比较,发现以随机森林算法为基础的估算模型效果最好。本文的发现,进一步验证了将人工智能方法应用在管理会计特别是定价决策中的可行性。Pricing decision is not only a direct way for enterprises to maximize profits but also a topic of widespread concern in the field of management accounting.Taking an e-commerce company as a sample,this paper uses machine learning methods to build a customer order price estimation model by using the sales data accumulated by the enterprise and the big data obtained from outside the enterprise.The price estimation model constructed in this paper selects 17 price influencing factors,including macro environment,customer attributes,product attributes,demand attributes and supply attributes.The model uses 12109 historical sales records as the training set.Through comparison,it is found that the estimation model based on random forest algorithm is the best.The findings of this paper further verify the feasibility of applying artificial intelligence methods to management accounting,especially pricing decision-making.

关 键 词：人工智能 机器学习 差别定价 算法可视化 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程] F274[自动化与计算机技术—控制科学与工程]