Link prediction in undirected weighted disease network

Through the social network analysis we can get new data from various data and is intended to generate future predictions in the network structure. Based on the nodes and relationships between nodes, the future structure of the network and new relationships or give up relationships can be estimated. The definition, features and studies of link prediction in social networks are examined in this thesis. Before starting to this topic data mining, social networks, social network analysis are explained. In this application, "Undirected Weighted Disease Network" is created with using patent information who completed blood test in Fırat University Hospital. In Disease Network, each node is disease and each relation is relationship between disease. We developed an new link prediction approach with using social network analysis methods and link prediction methods. Through our approach, detecting what kind of disesase risks of patients have coming the hospital with specific complaints. The availability of these methods and applicability of link prediction to these networks is shown. Key Words: Disease Network,Link Prediction, Social Network Analysis, Proximity Metrics Algorithm, Data Mining