ClassBD Achieves Exceptional Anti-Noise Performance on HIT Dataset with F1 Score Above 96% | HackerNoon

The HIT dataset was sourced from the MIIT Key Laboratory, featuring faulty bearings measured under a standard methodology, providing high-fidelity data for analysis.

Using an acceleration sensor with a NI USB-6002 device, approximately 561,152 data points were recorded per class, collecting significant data across various defect levels.

Read at Hackernoon