P7 · AEGIS · static

🛠️ Predictive Maintenance — Early-Warning

Project 7 · serves the AI4I2020 machine-failure classifier as a risk-banded early-warning tool (not a binary alarm). Enter the six machine readings — two engineered features (Temperature Difference, Power) are computed for you.

Machine readings — pre-filled with typical normal-operation values

Type

Air temperature (K)

Process temperature (K)

Rotational speed (rpm)

Torque (Nm)

Tool wear (min)

Production model —

Full model benchmark (8 models tested)

Bands are anchored on the model's own best-F1 threshold (not an arbitrary 0.50), consistent with Project 7's finding that the default cutoff understates every model's real quality on this imbalanced target.

Static edition — runs entirely in your browser, no server. Predictions are bit-exact reproductions of the original scikit-learn Random Forest pipeline (verified to machine epsilon). A feasibility/decision-support demo on the open AI4I2020 dataset — not a certified maintenance system.