基于贝叶斯网络的业务系统数据质量优化方法

Data Quality Optimization Method of Business System Based on Bayesian Networks

在信息化系统数据产生的过程中,系统中的数据质量是一个非常重要的问题,是系统可靠性运行的保障。目前业内在修正业务系统的数据时大多数采用的是人工监控事后修复的方法。由于信息化系统涉及大量不同业务规则和不同的应用场景,数据千差万别,这使得人工难以精准定位影响数据质量问题,且难以标准化管理。该文为解决这个问题,以自动定位问题字段及提高定位问题字段的精确度为目标,研究并提出了基于贝叶斯网络的业务系统数据质量优化方法,该方法主要分为2个阶段:(1)预先建立关联模型;(2)对数据质量进行分类,在数据质量不达标时,定位问题字段并由管理员对问题字段进行修改。最后,进行了具体的实例研究,证明了该方法是可行有效的。

In the process of producing data of information systems,the data quality in the system is a very important issue.Currently,the majority of the industry uses manual monitoring of repair methods to correct business system data.As information systems involve a large number of different business rules and different application scenarios,the data vary greatly,which makes it difficult for workers to accurately locate the problems that affect the data quality and is difficult to standardize management.In order to solve this problem,this paper focuses on locating the irregular fields automatically and improving the accuracy of the locating irregular fields.The business systems data quality optimization method based on Bayesian networks is studied and proposed.The method is divided into two phases：1.Pre-established association model;2.Classify the data quality,The method can locate the problem field and modify the question field by the administrator when the data quality does not reach the standard.Finally,we conduct a specific case study to prove that this method is feasible and effective.