Research on cleaning inaccurate data in production management module in ERP

With the rapid development of information technology, data quality has become a key factor in successfully operating and implementing ERP system. The problem of how to improve and enhance data quality in ERP has become an important research direction. However, because of the hugeness and complexity of ERP, this paper focuses on production management module and mainly aims at inaccurate data in it. Inaccurate data includes continuous abnormal data, discrete abnormal data and approximately duplicate records. Moreover, this paper designs different processes for detecting and cleaning different types of inaccurate data and then applies these processes to production management module in ERP system. At last, this paper illustrates how to use SOM clustering method and BP neural network to detect inaccurate data in production management module. It has certain directive significance for improving data quality in actual ERP system.