individual conditional expectation
In plain wordsA chart that traces, for each example separately, what a model would predict as one input is swept across its range, exposing differences between examples that an overall average would hide.
A plot that draws, for each instance in a dataset, the curve of the model's prediction as one feature is varied over a grid while that instance's other features are held fixed. The partial dependence plotpartial dependence plot, n.: A chart showing how a model's average prediction changes as one input is varied while everything else stays as it is in the data — for example, predicted house price against floor area. is the pointwise average of these curves, so ICE plots show what the average conceals: heterogeneous effects — a feature that raises predictions for some instances and lowers them for others — signal interactions with other features. Centered ICE plots anchor every curve at a common starting value so that differences in shape stand out, and derivative ICE plots show where effects change. Like partial dependence, ICE curves can visit unrealistic feature combinations when features are correlated, and with many instances they become cluttered, so analysts plot samples or clusters of curves.
The ICE plot split into two bundles: age raised predicted risk for smokers but not for non-smokers.
Origin
Introduced by Alex Goldstein, Adam Kapelner, Justin Bleich and Emil Pitkin in 'Peeking Inside the Black Box: Visualizing Statistical Learning with Plots of Individual Conditional Expectation' (arXiv, 2013; Journal of Computational and Graphical Statistics, 2015).
Related
- partial dependence plotpartial dependence plot, n.: A chart showing how a model's average prediction changes as one input is varied while everything else stays as it is in the data — for example, predicted house price against floor area.
- accumulated local effectsaccumulated local effects, n.: A chart of how a model's predictions change with one input, built from small local comparisons so it stays reliable when inputs are related to each other, unlike simpler average-based charts.
- local explanationlocal explanation, n.: An explanation of why an AI model gave one particular answer for one particular case — this patient, this loan — rather than how it behaves in general.
- global explanationglobal explanation, n.: An explanation of how an AI model behaves overall — which factors matter most across all its decisions, and in what way — rather than why it made one particular decision.
- model-agnosticmodel-agnostic, adj.: Describes a method that works with any kind of AI model, because it only needs to feed the model inputs and observe its outputs, never to look inside.
Sources
- Goldstein, A., Kapelner, A., Bleich, J. & Pitkin, E. (2015). Peeking Inside the Black Box: Visualizing Statistical Learning with Plots of Individual Conditional Expectation. Journal of Computational and Graphical Statistics 24(1), 44–65.
Mentioned in
partial dependence plotpartial dependence plot, n.: A chart showing how a model's average prediction changes as one input is varied while everything else stays as it is in the data — for example, predicted house price against floor area., accumulated local effectsaccumulated local effects, n.: A chart of how a model's predictions change with one input, built from small local comparisons so it stays reliable when inputs are related to each other, unlike simpler average-based charts., model-agnosticmodel-agnostic, adj.: Describes a method that works with any kind of AI model, because it only needs to feed the model inputs and observe its outputs, never to look inside.