Animation of inferred information flow




This film shows the same data as before only instead of nodes turning red when they repost, they turn puruple or yellow. A node turning purple signifies that the 'Linear Threshold Model' accurately predicted that the node would post. Turning yellow means that the model predicted the user would not post, but they did anyway.

Many nodes due in fact turn purple rather than yellow, suggesting that the model does accurately predict infections for real data. This is actual data gathered, and the specific topic being watched is defined as posts that contain the word 'superbowl'.



Back to the main page