An event based framework for improving information quality that integrates baseline models, causal models and formal reference models

Abstract

We introduce a framework for improving information quality in complex distributed systems that integrates: 1) Analytic models that describe baseline values for attributes and combinations of attributes and components that detect statistically significant changes from baselines. These models determine whether a significant change has occurred, and if so, when. 2) Casual models that help determine why a statistically significant change has occurred and what its impact is. These models focus on the reasons for a change. 3) Formal business and technical reference models so that data and information quality problems are less likely to occur in the future. In this note, we focus on the first two types of models and describe how this framework applies to data quality problems associated with electronic payments transactions and highway traffic patterns.