Visualizing high-dimensional predicitive model quality

Using inductive learning techniques to construct classification mod- els from large, high-dimensional data sets is a useful way to make predictions in complex domains. However, these models can be dif- ficult for users to understand. We have developed a set of visualiza- tion methods that help users to understand and analyze the behavior of learned models, including techniques for high-dimensional data space projection, display of probabilistic predictions, variable/class correlation, and instance mapping. We show the results of apply- ing these techniques to models constructed from a benchmark data set of census data, and draw conclusions about the utility of these methods for model understanding.