面向商业数据分析实验的数据合成系统设计  

Design of a data synthesizing system for business data analytics experiments

作　　者：李洋 李云矫[1] 韩赟 LI Yang;LI Yunjiao;HAN Yun(College of Management,Yunnan Normal University,Kunming 650032,China)

机构地区：[1]云南师范大学管理学院,云南昆明650032

出　　处：《无线互联科技》2025年第6期77-81,共5页Wireless Internet Science and Technology

基　　金：云南师范大学新文科研究与改革实践项目,项目编号:PX-30242094。

摘　　要：实验数据是实验教学和科研创新的关键要素。商业数据分析领域的科研实验和教学实验长期受限于数据的隐私性和敏感性而无法获取足量的真实数据。将机器学习技术应用于合成实验数据,基于Python编程语言生态,使用机器学习算法库SDV,文章设计了一种为商业数据分析实验提供合成数据的软件系统SD4BA,分析真实数据样本模式,训练数据生成器,产生任意数量的合成数据。通过对合成数据的主要特征进行评估,其与真实数据样本具有相似模式并实现了数据脱敏,可为商业数据分析实验提供较高保真度的足量数据。Experimental data is a key element in teaching,research and scientific innovation.Research and teaching experiments in the field of business analytics have been limited by the privacy and sensitivity of data for a long time,making it difficult to obtain sufficient amounts of real data.Applying machine learning techniques to synthesize experimental data,based on the Python programming language ecosystem and using the machine learning algorithm library SDV,a system SD4BA is designed to provide synthetic data for business analytics experiments.The system analyzes real data patterns,trains data generators,and generates any number of synthetic data.By evaluating the main features of the synthetic data,it has similar patterns to the real dataset and it is desensitized.SD4BA can provide sufficient data with high fidelity for business analytics experiments.

关 键 词：商业数据分析 机器学习 合成数据 数据脱敏 

分 类 号：G482[文化科学—教育学]