Start Date

11-12-2016 12:00 AM

Description

Online communities suffer from the 1-9-90 principle, which states that 1% of the community's user base generates original content, an additional 9% is limited to interacting with existing content, while the remaining 90% of the participants is passively lurking. In this work we present a data-driven stochastic framework that estimates (1) the activation potential (i.e., the users that are currently lurkers but present a high likelihood of becoming heavy contributors) of an online community and (2) when and which users are more likely to become heavy contributors. Our proposed framework captures the transitional evolution of a user by a Hidden Markov Model, and estimates each user's propensity to become a heavy contributor by employing parametric survival models. We build and evaluate our models on a unique large dataset of a specialized online community about diabetes.

Share

COinS
 
Dec 11th, 12:00 AM

Realizing the Activation Potential of Online Communities

Online communities suffer from the 1-9-90 principle, which states that 1% of the community's user base generates original content, an additional 9% is limited to interacting with existing content, while the remaining 90% of the participants is passively lurking. In this work we present a data-driven stochastic framework that estimates (1) the activation potential (i.e., the users that are currently lurkers but present a high likelihood of becoming heavy contributors) of an online community and (2) when and which users are more likely to become heavy contributors. Our proposed framework captures the transitional evolution of a user by a Hidden Markov Model, and estimates each user's propensity to become a heavy contributor by employing parametric survival models. We build and evaluate our models on a unique large dataset of a specialized online community about diabetes.