2. Predictive Analytics

Explainability, Monitoring and Drift

Explain the relevant system claim and detect when prediction or policy stops working

Explainability, Monitoring and Drift

Explain the right object

ObjectQuestion
datawhere did this feature come from and when was it available?
modelwhich inputs changed this score under the fitted function?
uncertaintyhow stable is the score across samples or plausible models?
decisionwhy did this score, threshold and constraint produce this action?
outcomedid the action improve the intended result?

A local feature attribution answers none of the other four automatically.

Three levels of explanation

  • global: broad model structure, important inputs, interactions and limits;
  • cohort: performance and behaviour for an operational or affected group;
  • local: evidence and rules behind one prediction/action.

Correlated inputs can make local attributions unstable. Perturbations may create impossible cases. An explanation should be tested for fidelity, stability and usefulness to its intended reader.

For high-stakes decisions, Rudin’s interpretable-model argument asks whether an intrinsically understandable model can meet performance needs before explaining an avoidable black box. Longo et al.’s XAI 2.0 agenda broadens the problem toward actionable, interdisciplinary and falsifiable explanations.

Drift taxonomy

DriftWhat changedHarborMart signal
data/covariateP(X)P(X)more express orders
priorP(Y)P(Y)lateness rises during disruption
conceptP(YX)P(Y\mid X)utilisation no longer maps to delay after automation
labeloutcome measurementnew delivery-event rule
policywho receives actionthreshold or capacity changed
feedbackaction changes future datare-promised orders disappear from late labels

Feature drift alone does not prove performance drift, and stable averages can hide concept failure. Monitor inputs, labels, calibration, decisions and outcomes.

Monitoring card

LayerMetricTriggerOwner response
datamissing route-distance rateabove 2% for two hoursfall back to rule; inspect feed
modelweekly calibration erroroutside validated bandrecalibrate or investigate shift
decisionoverride rate and reasonabrupt store-specific changereview policy usability
outcomelate cost per accepted orderworse than baseline guardrailrollback
fairnessfalse-negative gap by reviewed groupexceeds approved toleranceimpact review, not silent retuning

Every trigger needs a window, comparator, owner and action. “Monitor accuracy” is not an operating control.

Airbnb’s Chronon is a current organisational example of treating temporally correct backfills, online/offline consistency and feature monitoring as infrastructure concerns. It illustrates architecture, not a universal vendor prescription.

A compact model card

  • intended decision, population and non-use cases;
  • training/evaluation cutoff and feature availability;
  • baseline and performance with uncertainty;
  • calibration and subgroup results;
  • threshold, capacity and human override;
  • known failure modes and stress tests;
  • monitoring, rollback and accountable owners;
  • current version and change history.

Quick check

Input distributions are stable, but the same score band now has twice the observed late rate. What drift occurred?

Answer
The score–outcome relationship changed: concept or calibration drift. Investigate operations and labels; input-distribution monitoring alone would miss it.

Next: Prescriptive Analytics

Copyright © 2026