Adaptable Services for Novelty Mining

Novelty mining is the process of mining relevant information on a given topic. However, designing adaptable services for real-world novelty mining faces several challenges like real-time processing of incoming documents, computational efficiency, multi-user working environment, diverse system requirements, and integration of domain knowledge from different users. In this paper, the authors bridge the gap between generic data mining methodologies and domain-specific constraints by providing adaptable services for intelligent novelty mining that model user preferences by synthesizing the parameters of novelty scoring, threshold setting, performance monitoring, and contextual information access. The resulting novelty mining system has been tested in a variety of performance situations and user settings. By considering the special issues based on domain knowledge, the authors’ adaptable novelty mining services can be used to support a real-life enterprise. DOI: 10.4018/978-1-4666-1767-4.ch015