Abstract:Link Discovery is a relatively new form of data mining. Link Discovery is the technique for mining large amounts of data to find hidden patterns, extract valuable knowledge and discover hidden links. A survey on link discovery is presented in the paper, including the definition, the research range, characteristics and challenges of Link Discovery. And several typical Link Discovery methods are detailed, including unsupervised Link Discovery methods (novel Link Discovery), relational data mining with inductive logic programming for Link Discovery, multi-hypothesis abductive reasoning for Link Discovery, Link Discovery based on Correlation Analysis and the KOJAK group finder. We emphasize on the performance evaluation metrics for Link Discovery systems and measuring confidence intervals in Link Discovery. Further more the aplication of Link Discovery is presented: EELD. Lastly some issues about Link Discovery are discussed and the future directions of Link Discovery are pointed out.

Link Discovery:A New Data Mining Method