Using Text Mining to Create Actionable Knowledge: Application to Network Failure Incident Reports

Leading in the innovation economy [1] requires an ability to create, comprehend, and utilize dynamic and distributed knowledge assets. This must be accomplished in a reliable, timely, and cost effective manner. Text mining technology has great potential to enhance knowledge management systems because it provides an objective analysis (reliable) of existing knowledge assets (cost effective) in a rapid manner (timely). We show, through a case study involving the analysis of computer network failure incident reports that deep analytics can be used to create actionable knowledge concerning the technical system, and entity extraction can be used to highlight the underlying social architecture of the system. Centering resonance analysis is employed to create a data model of each incident report, and hierarchical clustering, factor analysis, time series analysis, and social network analysis are used to generate insights into the management, execution, and control of the computer network system and its underlying social system.