simAcc: Simulates accelerometer data based on a hidden Markov model

In acc: Exploring Accelerometer Data

Description

Simulates accelerometer data. The simulation function is based on a hidden Markov model, as described in the example for function acc. This function is provided for convenience to generate data from a pre-specified transition probabilities to mimic activity levels of low, moderate and high. To generate data from a specific transition probabilities and distributions, please refer to the example for function acc.

Usage

simAcc(timelength,paLevel,epoch,startDate,endDate,
       mu,sigma,seedset,tpm)

Arguments

timelength	Number of observations to be generated.
paLevel	Pre-specified levels of physical activity for convenience. User can specify all parameter as preferred, by stating the option as paLevel=NULL. Default is ‘moderate’. Options: 1) ‘low’, ‘moderate’, or ‘high’. Low specifies a hidden markov model with transition probabilities 0.95, 0.04, 0.01, 0.09, 0.9, 0.01, 0.1, 0.2, 0.7, respectively for P11, P12, P13, P21, P22, P23, P31, P32, P33, respectively. Moderate specifies a hidden markov model with transition probabilities 0.95, 0.04, 0.01, 0.09, 0.8, 0.11, 0.1, 0.1, 0.8 respectively for P11, P12, P13, P21, P22, P23, P31, P32, P33, respectively. High specifies a hidden markov model with transition probabilities 0.95, 0.04, 0.01, 0.09, 0.7, 0.21, 0.1, 0.1, 0.8, respectively for P11, P12, P13, P21, P22, P23, P31, P32, P33, respectively. For all levels, it is assumed that the activity intensities are realized from a mixture of two normal distributions (for sedentary activity and mvpa) and a constant at zero (for non-wear time), with means mu = c(0, 30, 2500) and variance sigma = c(0, 30, 1000).
epoch	Epoch size. User can specify desired epoch size in units of time larger than seconds. Defaults to 1 minute epoch.
startDate	Start date in ISOdate format. For example ISOdate(2017,1,1,hour=0,min=0,sec=0,tz="GMT").
endDate	End date in ISOdate format. For example ISOdate(2017,1,1,hour=0,min=0,sec=0,tz="GMT").
mu	Mean levels for each activity type.
sigma	Standard deviations for each activity type.
seedset	Sets seed for random data generation. Defaults to 1234.
tpm	Transition probability matrix that specify probability of change from one activity state to another.

Value

A simulated dataset is returned with two columns: [TimeStamp, counts]

Author(s)

Jaejoon Song <[e-mail]>

Examples

## Not run: 
##
## Example: Simulate a dataset for two days, for an individual with low activity level.
##
mvpaLowData <- simAcc(timelength=(60*24*2),paLevel='low')
summary <- acc(data=mvpaLowData, tri='FALSE', axis=NULL,
                     spuriousDef=20, nonwearDef=60, minWear=600, 
                     patype=c('Sedentary','MVPA'),pacut=c(c(0,99),c(1952,Inf)), 
                     boutsize=c(10,10), tolerance=c('FALSE','TRUE'))
summary

##
## Example: Simulate a dataset for two days, for an individual with moderate activity level.
##
mvpaLowData <- simAcc(timelength=(60*24*2),paLevel='moderate')
summary <- acc(data=mvpaLowData, tri='FALSE', axis=NULL,
                     spuriousDef=20, nonwearDef=60, minWear=600, 
                     patype=c('Sedentary','MVPA'),pacut=c(c(0,99),c(1952,Inf)), 
                     boutsize=c(10,10), tolerance=c('FALSE','TRUE'))
summary

##
## Example: Simulate a dataset for two days, for an individual with high activity level.
##
mvpaLowData <- simAcc(timelength=(60*24*2),paLevel='high')
summary <- acc(data=mvpaLowData, tri='FALSE', axis=NULL,
                     spuriousDef=20, nonwearDef=60, minWear=600, 
                     patype=c('Sedentary','MVPA'),pacut=c(c(0,99),c(1952,Inf)), 
                     boutsize=c(10,10), tolerance=c('FALSE','TRUE'))
summary


## End(Not run)

Example output

Loading required package: mhsmm
Loading required package: mvtnorm
        Date wearTime Sedentary.minutes Sedentary.num.bouts MVPA.minutes
1 2017-01-01      978               910                  17           10
2 2017-01-02     1284              1235                  13           22
  MVPA.num.bouts
1              1
2              2
        Date wearTime Sedentary.minutes Sedentary.num.bouts MVPA.minutes
1 2017-01-01      977               612                  26          112
2 2017-01-02     1284               973                  24          177
  MVPA.num.bouts
1              7
2             11
        Date wearTime Sedentary.minutes Sedentary.num.bouts MVPA.minutes
1 2017-01-01      977               514                  25          166
2 2017-01-02     1284               883                  28          273
  MVPA.num.bouts
1             11
2             14

acc documentation built on May 1, 2019, 9:26 p.m.