On the use of machine-assisted knowledge discovery to analyze and reengineer measurement frameworks

We call the set of metrics, data collection mechanisms, and measurement models used by organizations in running their businesses a Measurement Framework. This paper [1] describes how a knowledge discovery technique called Attribute Focusing (AF) can be combined with a measurement planning approach called the Goal/Question/Metric Paradigm (GQM) to analyze and reengineer the Measurement Framework of an organization. The GQM Paradigm is widely used by the software engineering community to handle Measurement Frameworks in a top-down, goal-oriented fashion. The AF technique is a machine-assisted knowledge discovery technique which has been widely used to help domain experts search for knowledge in a database of measurement (attribute-valued) data. Using our experience analyzing Software Customer Satisfaction survey data at IBM, we illustrate how the AF Technique can be combined with GQM to improve a Measurement Framework. We argue that this may be a good approach to reengineering and improving existing Measurement Frameworks.