Evaluation of the model for individual behavior rate estimate: Social network data
The paper described the Bayesian belief network model for individual behavior rate estimate based on data about the last episodes of that behavior. For model evaluation we used data from social network VKontakte about episodes of publishing posts. The sample size is 1123 users with 160555 posts in total for the half-year period. Rate values were discretized and form eight classes. There was no statistical difference between estimated rate distribution in the group and the real rate distribution; individual rate estimates showed 91% classification accuracy. Hence, the model allows behavior rate estimating on the base of data about limited number of episodes that is often an important issue, for example, in epidemiological studies.