Algorithm 2. Data Preprocessing and Feature Extraction

1: Data Preprocessing 2: for each d in participant data do 3:     s ← MAF(PPG) {moving average filter} 4:     d′ ← SGF(s) {Savitzky–Golay filter} 5:     X ← a(n) 6: end for 7: Feature Extraction 8: for each d′ do 9:     AR ← CalcAR(d′) 10:    Fppg ← {AR.F 1, AR.F 2, AR.F 3, AR.F 4, AR.F 5} 11:    KTE ← CalcKTE(d′) 12:    FKT E ← {KTE.µ, KTE.skew, KTE.IQR, KTE.σ} 13:    HRV ← CalcHRV(d′) 14:    FHRV ← {HRV.α, HRV.β, HRV.IQR, HRV.skew} 15:    SE ← CalcSE(d′) 16:    FSE ← {SE.α, SE.β, SE.IQR, SE.skew} 17:    EP ← CalcLogEP(d′) 18:    FLogEP ← {EP.AR.F 1, EP.AR.F 2, EP.AR.F 3, EP.AR.F 4, EP.AR.F 5, EP.β,       EP.IQR} 19:    PTT ← CalcPTT(d′) 20:    FP T T ← {PTT.α, PTT.β, PTT.IQR, PTT.skew} 21:    PPI ← CalcPPI(d′) 22:    FP P I ← {PPI.α, PPI.β, PPI.IQR, PPI.skew} 23:    FV ← Combine(Fppg, FKT E, FHRV, FSE, FLogEP, FP T T, FP P I ) 24: end for 25: Data Splitting 26: TR, TE ← Split(FV, 0.8) {80:20 Split}