Identifying Leaders in an Online Cancer Survivor Community

Motivations:

Online communities are important sources of social support for cancer survivors and their informal caregivers. Leaders in the online communities will have significant impacts on other members. Thus identifying the leaders in the online communities may contribute a lot to better supports for cancer survivors and the caregivers.

Research Questions:

How to identify leaders in a popular online forum for cancer survivors and caregivers? What features may the leaders have?

Methodoloty:

Classification techniques are used. User features are extracted from many different perspectives. Based on these features, the structure of the network among users is exploited, and new neighborhood-based and cluster-based features are generated. With all these features, classifications are conducted.

Main Results:

Classification results revealed that the features extracted are discriminative for leader identification. Using these features, a hybrid approach based on an ensemble classifier is developed and it performs better than many traditional metrics. This research has important implications for understanding and managing similar online communities.