Exploring Local Explanations of Nonlinear Models Using Animated Linear Projections

The increased predictive power of nonlinear models comes at the cost of interpretability of its terms. This trade-off has led to the emergence of eXplainable AI (XAI). XAI attempts to shed light on how models use predictors to arrive at a prediction with local explanations, a point estimate of the linear feature importance in the vicinity of one instance. These can be considered linear projections and can be further explored to understand better the interactions between features used to make predictions across the predictive model surface. Here we describe interactive linear interpolation used for exploration at any instance and illustrate with examples with categorical (penguin species, chocolate types) and quantitative (soccer/football salaries, house prices) output. The methods are implemented in the R package cheem, available on CRAN.Comment: 24 pages, 9 figures, 0 table