Management of Data Mining Model Lifecycle to Support Intelligent Business Services

Technology (IT) management is going through its third phase of evolution. In the first phase, it was about managing silos of devices (servers, storage) and applications. In the next phase, best practices frameworks such as IT Infrastructure Library (ITIL) prescribed a service-oriented approach to IT management. Now, we are seeing a shift of focus from the bottom-up IT-driven approaches to top-down business-driven strategies. This final phase also called Business Service Management (BSM) will enforce business accountability on people as well as IT by automating and managing business processes and leveraging IT Service Management (ITSM) practices. While the IT industry is moving forward with these modern concepts, they have left behind critical processes and soft IT assets unmanaged, especially at the intersection of business processes with Business Intelligence (BI). In this paper, we describe the design and implementation of a system that manages data mining model assets of an organization that can support business processes in making real-time decisions and forecasts. We address technical challenges related to model aging, scalability, model lifecycle, and metadata management. We also identify a number of grand challenges related to organizational ethnography (such as analysis of semantic gaps between business analysts and statisticians), visualization support, and longitudinal studies needed to identify evaluation metrics.