Avanish Choudhary completed Predictive Maintenance: Industrial ML for Fault Detection

Built a compact bearing-fault detector with a shallow decision tree trained only on training bearings, then evaluated it on a held-out set and justified an alarm threshold using catch-rate and false-alarm trade-offs. The memo also identified the outer-ring fault from the dominant BPFO signature and translated the result into plain maintenance guidance.

• Leakage-Aware Model Training
• Held-Out Evaluation
• Cost-Based Thresholding