Business rule discovery through data mining methods

Engineering asset management processes rely heavily on input of data and also produce a large amount of data. Many asset management organisations need to manage their data for a long period of time (e.g. Water supply pipelines data will be kept for more than 100 years in utility companies). Due to the inability to access original data requirements and system design documentation, it is difficult for these organisations to redesign their asset management systems which often results in ongoing data quality problems. Our nation-wide data quality survey of 2500 Australian engineering management organisations and a pilot study have revealed that many of the data quality problems emanate from inconsistent applications of business rules that govern the behaviour of data (e.g. data management, data flow, system interactions and so on) within asset management information systems. Thus, this research will investigate the problems of business rule-based data and information integration from disparate sources in various forms found in asset management systems (e.g. Databases, Excel spreadsheets, etc.). This research aims at developing innovative methods to automatically discover undocumented business rules from disparate data sources and to use business rules for automated data integration in order to deliver quality asset data sets.