Human Calibration

Compares Customer Twin prediction against real customer evaluation.

Illustrative Calibration Simulation

-0.06

Spearman ρ

Rank alignment across configurations

0.36

MAE

Average absolute score difference

55%

Decision Agreement

BUY/CONSIDER/REJECT agreement

0.91

Stability

1 − normalized MAE

AI–Human Alignment

Illustrative calibration simulation

Calibration Metrics

MetricValueInterpretation
Spearman ρ-0.06Rank alignment across configurations
MAE0.36Average absolute score difference (1-5 scale)
Decision Agreement55%BUY/CONSIDER/REJECT agreement
Stability0.911 − normalized MAE across illustrative responses

These values demonstrate how the system would report calibration after real human validation data are collected. Computed from 14 configuration(s) with ≥10 respondents each.