Case-Based Reasoning for Summarizing Simulation Results

Simulations can produce large quantities of data. To reason about the results of simulations, machine-learning methods can be helpful. We explored a case-based reasoning approach to summarizing the results of a probabilistic simulation of naval combat involving missiles. We used a tree structure to index the data and showed that it gave good accuracy in estimating the results of this simulation with new parameters. We are now extending these ideas to a more complex military simulation.