Finding relationships of symptom-disease and disease-drug in online health forum sites

Social networks are one of the structures used to model relationships between the objects, individuals, organizations and communities in our environment. Many complex systems that we face in different areas and in different ways can be visualized with social networks. Social network is a useful tool for understanding the structure, development, and relationships of complex systems. Social network analysis provides valuable information about complex systems. This information can be used for many purposes. One of the most important research topics in social network analysis is link prediction. Social networks change and evolve over time. Link prediction is an effective way to investigate dynamics of networks and has become an important analysis topic in health as well as in many areas. In this thesis, three main studies are presented. The first study is based on finding of associations between diseases and symptoms obtained with analysis of the questions on the online forum sites. Link prediction, which has been applied in many areas over the past few years, has been carried out on a bipartite network constructed with data obtained from analysis of online forum sites, which is an important reference source in health field for people with rapid development of the Internet. The second study aims to determine relation of the diseases that correspond to user questions with the drugs in the online forum sites. For this purpose, a bipartite network consisting of diseases and drugs is modeled, and relations between diseases and drugs suitable to these diseases are found with link prediction. A recent study has focused on the reliability of online forum sites that people often use. With this study, how accurate the answers that people get from health professionals on these forums have been tested. Finally, this thesis study was supported by Project No:MF.17.04 the Master Thesis Research Project carried out within the scope of FÜBAP, and the studies were concluded with various international publications.