A tiny logistic model learns to classify fictional character profiles as Jets or Sharks using simple social cues. It is a toy classifier inspired by connectionist ideas, not a full interactive-activation model.
What to do: Click Train model to let the network learn from labelled training cases. Once trained, toggle the cue checkboxes on the left and watch how the Jet probability on the right changes. Pay attention to the learned weights in the table.
Think about: Which features received the largest positive or negative weights? A large positive weight means the feature pushes the prediction toward "Jet"; a large negative weight pushes it toward "Shark." This is how even a simple network captures statistical patterns in data — the same principle behind larger neural networks in modern AI.
| Feature | Weight |
|---|