Exploiting evidence from unstructured data to enhance master data management

Master data management (MDM) integrates data from multiple<br/>structured data sources and builds a consolidated 360-<br/>degree view of business entities such as customers and products.<br/>Today’s MDM systems are not prepared to integrate<br/>information from unstructured data sources, such as news<br/>reports, emails, call-center transcripts, and chat logs. However,<br/>those unstructured data sources may contain valuable<br/>information about the same entities known to MDM from<br/>the structured data sources. Integrating information from<br/>unstructured data into MDM is challenging as textual references<br/>to existing MDM entities are often incomplete and<br/>imprecise and the additional entity information extracted<br/>from text should not impact the trustworthiness of MDM<br/>data.<br/>In this paper, we present an architecture for making MDM<br/>text-aware and showcase its implementation as IBM InfoSphere<br/>MDM Extension for Unstructured Text Correlation,<br/>an add-on to IBM InfoSphere Master Data Management<br/>Standard Edition. We highlight how MDM benefits from<br/>additional evidence found in documents when doing entity<br/>resolution and relationship discovery. We experimentally<br/>demonstrate the feasibility of integrating information from<br/>unstructured data sources into MDM.