An Algorithm for Reducing the Complexity of Interaction Networks

We present an algorithm for reducing the size and complexity of the Interaction Network, a data structure used for storing solution paths explored by students in open-ended multi-step problem solving environments. Our method reduces the number of edges and nodes of an Interaction Network by an average of 90% while still accounting for 40% of actions performed by students, and preserves the most frequent half of solution paths. We compare our method to two other approaches and demonstrate why it is more effective at reducing the size of large Interaction Networks.