UC Santa Cruz

Nearly all enterprises are routinely collecting data and information and attempting to transform it into knowledge that can be applied to core business activities, e.g. product development, customer support, and marketing. In this work we develop and apply a representation-based methodology for building Knowledge Engineering systems to support the rapid and effective extraction of knowledge from unstructured data. The proposed methodology draws upon methods and techniques from Knowledge Engineering, Product Design, and Software Engineering in order to maximize the overall value of the system with respect to the needs of the users and organization. The domain of Product Design provides formal tools for identifying user needs, exploring different function realizations, and managing trade-offs between quality and cost. Application of the methodology at a large computer networking company produced a "Service Request Portal" tool that was well received by end-users and resulted in a 30% productivity improvement compared to previous tools.