Tests on synthetic data

• TWA amplitudes were 2, 4, 6, 8, 10, 16, 22, 28, 34, 40, 60 V

• Maximum error after scaling:

• Clean records – all methods better than 2 V

• White noise with standard deviation of 5, 10, 20, 30, 40 V

• • SM ‘standard’ better than 6 V, SM ‘differences’ better than 5 V where successful (TWA greater than 0.35*noise level)
  • MMA better than 5 V where SM is successful; fails to reject low TWA/noise ratio
  • All methods improve as noise decreases and TWA increases

• Baseline wander (NST DB) added – all methods better than 7 V