Table 7 Log rank results applying \((\eta =\frac{CA-BA}{CA})\) for different SBSI categories. Results are shown for model with all subjects, female-only, and male-only separately. Q1, Q2, etc. denote 1st quartile, 2nd quartile, etc.

From: Deep Learning using Convolutional LSTM estimates Biological Age from Physical Activity

  ConvLSTM* SBSI Q1 SBSI Q2 SBSI Q3 SBSI Q4
Chi-sq p-value Chi-sq p-value Chi-sq p-value Chi-sq p-value
ALL ConvLSTM* (λ = 1) 3.03 3.87E-01 7.75 5.14E-02 2.89 4.09E-01 15.83 1.23E-03
ConvLSTM* (λ = 0) 18.66 3.21E-04 5.78 1.23E-01 29.01 2.23E-06 61.52 2.78E-13
ConvLSTM* (λ = 0.9) 13.25 4.12E-03 8.57 3.55E-02 13.37 3.90E-03 38.01 2.81E-08
Female ConvLSTM* (λ = 1) 3.58 3.11E-01 4.86 1.82E-01 10.28 1.64E-02 12.88 4.90E-03
ConvLSTM* (λ = 0) 10.34 1.59E-02 11.35 9.95E-03 24.25 2.21E-05 65.76 3.45E-14
ConvLSTM* (λ = 0.9) 7.69 5.28E-02 6.25 1.00E-01 11.52 9.23E-03 26.58 7.23E-06
Male ConvLSTM* (λ = 1) 2.55 4.67E-01 7.69 5.29E-02 4.38 2.23E-01 3.71 2.94E-01
ConvLSTM* (λ = 0) 5.07 1.67E-01 1.58 6.63E-01 3.02 3.89E-01 8.00 4.60E-02
ConvLSTM* (λ = 0.9) 13.86 3.10E-03 3.56 3.14E-01 12.4 6.14E-03 27.84 3.92E-06