Table 4

Population HMM: parameter estimates for discrete-time and heterogeneous-time setting.

Methods: Estimate (SE)
TimeTransitionParametersPMMDT-HMMCT-HMMPH-HMM
DiscreteActiveβ102.8973(0.8384)10.8254(1.4074)9.0154(1.2162)8.5893(1.2036)
Timetoβ11−0.4182(0.0850)−1.2431(0.1428)−1.0828(0.1232)−1.0395(0.1219)
Restβ120.0288(0.0121)0.0379(0.0184)0.0257(0.0126)0.0257(0.0126)
β130.0070(0.0124)0.0052(0.0189)0.0054(0.0130)0.0050(0.0129)
formula0.10990.25850.11710.1160
Restβ20−6.6357(0.3846)−16.2898(1.3782)−6.0404(0.4941)−5.9631(0.5029)
toβ210.5359(0.0405)1.5280(0.1401)0.4671(0.0512)0.4599(0.0521)
Activeβ22−0.0278(0.0190)−0.0441(0.0203)−0.0357(0.0183)−0.0355(0.0183)
β23−0.0431(0.0191)−0.0541(0.0208)−0.0503(0.0187)−0.0505(0.0188)
formula0.28020.32100.26310.2634
μ18.53218.08668.02518.0114
μ20.52380.42630.41950.4153
HeterogeneousActiveβ100.3714(0.7637)6.4420(1.2828)2.8110(1.1055)2.0846(1.0867)
Timetoβ11−0.2567(0.0786)−0.7771(0.1305)−0.5526(0.1130)−0.5209(0.1120)
Restβ120.0541(0.0214)0.0496(0.0231)0.0528(0.0233)0.0685(0.0285)
β130.0867(0.0221)0.0094(0.0241)0.0840(0.0240)0.1127(0.0296)
formula0.35540.39410.41020.6132
Restβ20−6.4771(0.3775)−20.1172(1.2551)−6.9784(0.4544)−6.8518(0.4673)
toβ210.4597(0.0411)2.0046(0.1298)0.5196(0.0481)0.4889(0.0500)
Activeβ22−0.0211(0.0235)−0.0620(0.0294)−0.0383(0.0217)−0.0327(0.0241)
β230.0033(0.0240)−0.0874(0.0309)−0.0199(0.0225)−0.0077(0.0252)
formula0.40510.62070.34220.4040
μ19.39399.05308.95228.9428
μ21.05340.95790.95280.9534
Methods: Estimate (SE)
TimeTransitionParametersPMMDT-HMMCT-HMMPH-HMM
DiscreteActiveβ102.8973(0.8384)10.8254(1.4074)9.0154(1.2162)8.5893(1.2036)
Timetoβ11−0.4182(0.0850)−1.2431(0.1428)−1.0828(0.1232)−1.0395(0.1219)
Restβ120.0288(0.0121)0.0379(0.0184)0.0257(0.0126)0.0257(0.0126)
β130.0070(0.0124)0.0052(0.0189)0.0054(0.0130)0.0050(0.0129)
formula0.10990.25850.11710.1160
Restβ20−6.6357(0.3846)−16.2898(1.3782)−6.0404(0.4941)−5.9631(0.5029)
toβ210.5359(0.0405)1.5280(0.1401)0.4671(0.0512)0.4599(0.0521)
Activeβ22−0.0278(0.0190)−0.0441(0.0203)−0.0357(0.0183)−0.0355(0.0183)
β23−0.0431(0.0191)−0.0541(0.0208)−0.0503(0.0187)−0.0505(0.0188)
formula0.28020.32100.26310.2634
μ18.53218.08668.02518.0114
μ20.52380.42630.41950.4153
HeterogeneousActiveβ100.3714(0.7637)6.4420(1.2828)2.8110(1.1055)2.0846(1.0867)
Timetoβ11−0.2567(0.0786)−0.7771(0.1305)−0.5526(0.1130)−0.5209(0.1120)
Restβ120.0541(0.0214)0.0496(0.0231)0.0528(0.0233)0.0685(0.0285)
β130.0867(0.0221)0.0094(0.0241)0.0840(0.0240)0.1127(0.0296)
formula0.35540.39410.41020.6132
Restβ20−6.4771(0.3775)−20.1172(1.2551)−6.9784(0.4544)−6.8518(0.4673)
toβ210.4597(0.0411)2.0046(0.1298)0.5196(0.0481)0.4889(0.0500)
Activeβ22−0.0211(0.0235)−0.0620(0.0294)−0.0383(0.0217)−0.0327(0.0241)
β230.0033(0.0240)−0.0874(0.0309)−0.0199(0.0225)−0.0077(0.0252)
formula0.40510.62070.34220.4040
μ19.39399.05308.95228.9428
μ21.05340.95790.95280.9534

Note: Parameter estimates using EM algorithm and asymptotic standard errors. Regression for state transitions are given as formula, where formula are individual specific random intercepts, Android devices and males serve as baseline.

Table 4

Population HMM: parameter estimates for discrete-time and heterogeneous-time setting.

Methods: Estimate (SE)
TimeTransitionParametersPMMDT-HMMCT-HMMPH-HMM
DiscreteActiveβ102.8973(0.8384)10.8254(1.4074)9.0154(1.2162)8.5893(1.2036)
Timetoβ11−0.4182(0.0850)−1.2431(0.1428)−1.0828(0.1232)−1.0395(0.1219)
Restβ120.0288(0.0121)0.0379(0.0184)0.0257(0.0126)0.0257(0.0126)
β130.0070(0.0124)0.0052(0.0189)0.0054(0.0130)0.0050(0.0129)
formula0.10990.25850.11710.1160
Restβ20−6.6357(0.3846)−16.2898(1.3782)−6.0404(0.4941)−5.9631(0.5029)
toβ210.5359(0.0405)1.5280(0.1401)0.4671(0.0512)0.4599(0.0521)
Activeβ22−0.0278(0.0190)−0.0441(0.0203)−0.0357(0.0183)−0.0355(0.0183)
β23−0.0431(0.0191)−0.0541(0.0208)−0.0503(0.0187)−0.0505(0.0188)
formula0.28020.32100.26310.2634
μ18.53218.08668.02518.0114
μ20.52380.42630.41950.4153
HeterogeneousActiveβ100.3714(0.7637)6.4420(1.2828)2.8110(1.1055)2.0846(1.0867)
Timetoβ11−0.2567(0.0786)−0.7771(0.1305)−0.5526(0.1130)−0.5209(0.1120)
Restβ120.0541(0.0214)0.0496(0.0231)0.0528(0.0233)0.0685(0.0285)
β130.0867(0.0221)0.0094(0.0241)0.0840(0.0240)0.1127(0.0296)
formula0.35540.39410.41020.6132
Restβ20−6.4771(0.3775)−20.1172(1.2551)−6.9784(0.4544)−6.8518(0.4673)
toβ210.4597(0.0411)2.0046(0.1298)0.5196(0.0481)0.4889(0.0500)
Activeβ22−0.0211(0.0235)−0.0620(0.0294)−0.0383(0.0217)−0.0327(0.0241)
β230.0033(0.0240)−0.0874(0.0309)−0.0199(0.0225)−0.0077(0.0252)
formula0.40510.62070.34220.4040
μ19.39399.05308.95228.9428
μ21.05340.95790.95280.9534
Methods: Estimate (SE)
TimeTransitionParametersPMMDT-HMMCT-HMMPH-HMM
DiscreteActiveβ102.8973(0.8384)10.8254(1.4074)9.0154(1.2162)8.5893(1.2036)
Timetoβ11−0.4182(0.0850)−1.2431(0.1428)−1.0828(0.1232)−1.0395(0.1219)
Restβ120.0288(0.0121)0.0379(0.0184)0.0257(0.0126)0.0257(0.0126)
β130.0070(0.0124)0.0052(0.0189)0.0054(0.0130)0.0050(0.0129)
formula0.10990.25850.11710.1160
Restβ20−6.6357(0.3846)−16.2898(1.3782)−6.0404(0.4941)−5.9631(0.5029)
toβ210.5359(0.0405)1.5280(0.1401)0.4671(0.0512)0.4599(0.0521)
Activeβ22−0.0278(0.0190)−0.0441(0.0203)−0.0357(0.0183)−0.0355(0.0183)
β23−0.0431(0.0191)−0.0541(0.0208)−0.0503(0.0187)−0.0505(0.0188)
formula0.28020.32100.26310.2634
μ18.53218.08668.02518.0114
μ20.52380.42630.41950.4153
HeterogeneousActiveβ100.3714(0.7637)6.4420(1.2828)2.8110(1.1055)2.0846(1.0867)
Timetoβ11−0.2567(0.0786)−0.7771(0.1305)−0.5526(0.1130)−0.5209(0.1120)
Restβ120.0541(0.0214)0.0496(0.0231)0.0528(0.0233)0.0685(0.0285)
β130.0867(0.0221)0.0094(0.0241)0.0840(0.0240)0.1127(0.0296)
formula0.35540.39410.41020.6132
Restβ20−6.4771(0.3775)−20.1172(1.2551)−6.9784(0.4544)−6.8518(0.4673)
toβ210.4597(0.0411)2.0046(0.1298)0.5196(0.0481)0.4889(0.0500)
Activeβ22−0.0211(0.0235)−0.0620(0.0294)−0.0383(0.0217)−0.0327(0.0241)
β230.0033(0.0240)−0.0874(0.0309)−0.0199(0.0225)−0.0077(0.0252)
formula0.40510.62070.34220.4040
μ19.39399.05308.95228.9428
μ21.05340.95790.95280.9534

Note: Parameter estimates using EM algorithm and asymptotic standard errors. Regression for state transitions are given as formula, where formula are individual specific random intercepts, Android devices and males serve as baseline.

Close
This Feature Is Available To Subscribers Only

Sign In or Create an Account

Close

This PDF is available to Subscribers Only

View Article Abstract & Purchase Options

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

Close