Individual Doctor Recommendation in Large Networks by Constrained Optimization

Related Work

Traditional information retrieval models, such as Boolean Model (Ricardo, 1999), Multi-keyword method (Chen, 2012) and Vector Space Model (Salton, 1975), all compute similarity degree between query keywords and destination doctors. A closely related research topic is expertise search, such as expertise search based on candidate vote by Macdonald and Ounis (Macdonald, 2006), expertise mining from social network by Tang et al.(Tang, 2009), and transfer learning from expertise search to Bole search by Yang et al. (Heidelberger, 2011). The probabilistic topic models are often used in social ties/relationships analysis and mining (Zheng, 2012). Bloom Filter based distributed random replication scheme is proposed for content retrieval in unstructured P2P networks (Chen, 2012). Several recent studies focus on how to model the imbalanced and noisy data in order to improve relationships mining performance (Yuan, 2012; Zhou, 2009), how to search in P2P Networks (Chen, 2009; Chen, 2010) and how to obtain better classification results (Zhang, 2009).