Techniques for intelligent novelty mining
Tsai, Flora S.
Date of Issue2010
School of Electrical and Electronic Engineering
Intelligent novelty mining addresses the domain-specific problem of mining novel information from text data with specific regard to the user context. Traditional novelty mining focuses on optimizing the retrieval of novel information with a predefined evaluation measure. This tends not be useful in realistic situations where a fixed unit of measure is unable to adapt to the performance requirements of different users. Intelligent novelty mining, on the other hand, aims to balance the technical significance and business concerns to create techniques that are useful in real-world scenarios. This thesis addressed the problem of intelligent novelty mining, which is a big challenge in the data mining and business community. The proposed techniques aim to leverage the use of novelty metrics, novelty decision, and novelty feedback to improve the results of mining new information from text data.
DRNTU::Engineering::Electrical and electronic engineering
Nanyang Technological University