| Type: | Package |
| Title: | Person-Oriented Method and Perturbation on the Model |
| Version: | 0.2.1 |
| Maintainer: | Xiao Yang <vwendy@gmail.com> |
| Description: | An implementation of a hybrid method of person-oriented method and perturbation on the model. Pompom is the initials of the two methods. The hybrid method will provide a multivariate intraindividual variability metric (iRAM). The person-oriented method used in this package refers to uSEM (unified structural equation modeling, see Kim et al., 2007, Gates et al., 2010 and Gates et al., 2012 for details). Perturbation on the model was conducted according to impulse response analysis introduced in Lutkepohl (2007). Kim, J., Zhu, W., Chang, L., Bentler, P. M., & Ernst, T. (2007) <doi:10.1002/hbm.20259>. Gates, K. M., Molenaar, P. C. M., Hillary, F. G., Ram, N., & Rovine, M. J. (2010) <doi:10.1016/j.neuroimage.2009.12.117>. Gates, K. M., & Molenaar, P. C. M. (2012) <doi:10.1016/j.neuroimage.2012.06.026>. Lutkepohl, H. (2007, ISBN:3540262393). |
| License: | GPL-2 |
| Encoding: | UTF-8 |
| LazyData: | true |
| RoxygenNote: | 7.1.1 |
| Depends: | R (≥ 3.0.0) |
| Imports: | lavaan (≥ 0.5-23.1097), ggplot2 (≥ 2.2.1), reshape2 (≥ 1.4.2), qgraph, utils |
| Suggests: | knitr, rmarkdown, testthat |
| VignetteBuilder: | knitr |
| NeedsCompilation: | no |
| Packaged: | 2021-01-22 15:46:39 UTC; Xiao Yang |
| Author: | Xiao Yang [cre, aut], Nilam Ram [aut], Peter Molenaar [aut] |
| Repository: | CRAN |
| Date/Publication: | 2021-02-15 00:40:02 UTC |
Bootstrapped iRAM (including replications of iRAM and corresponding time profiles) for the bivariate time-series (simts2node)
Description
Bootstrapped iRAM (including replications of iRAM and corresponding time profiles) for the bivariate time-series (simts2node)
Usage
bootstrap_iRAM_2node
Format
An object of class list of length 5.
Details
Data bootstrapped from the estimated three-node network structure with 200 replications.
Examples
bootstrap_iRAM_2node$mean # mean of bootstrapped iRAM
bootstrap_iRAM_2node$upper # Upper bound of confidence interval of bootstrapped iRAM
bootstrap_iRAM_2node$lower # lower bound of confidence interval of bootstrapped iRAM
bootstrap_iRAM_2node$time.profile.data # time profiles generated from the bootstrapped beta matrices
bootstrap_iRAM_2node$recovery.time.reps # iRAMs generated from the bootstrapped beta matrices
Bootstrapped iRAM (including replications of iRAM and corresponding time profiles) for the 3-variate time-series (simts)
Description
Bootstrapped iRAM (including replications of iRAM and corresponding time profiles) for the 3-variate time-series (simts)
Usage
bootstrap_iRAM_3node
Format
An object of class list of length 5.
Details
Data bootstrapped from the estimated three-node network structure with 200 replications.
Examples
bootstrap_iRAM_3node$mean # mean of bootstrapped iRAM
bootstrap_iRAM_3node$upper # Upper bound of confidence interval of bootstrapped iRAM
bootstrap_iRAM_3node$lower # lower bound of confidence interval of bootstrapped iRAM
bootstrap_iRAM_3node$time.profile.data # time profiles generated from the bootstrapped beta matrices
bootstrap_iRAM_3node$recovery.time.reps # iRAMs generated from the bootstrapped beta matrices
Generate iRAM (impulse response anlaysis metric) from model fit.
Description
Generate iRAM (impulse response anlaysis metric) from model fit.
Usage
iRAM(
model.fit,
beta,
var.number,
lag.order = 1,
threshold = 0.01,
boot = FALSE,
replication = 200,
steps = 100
)
Arguments
model.fit |
model fit object generated by lavaan |
beta |
beta matrix for a point estimate |
var.number |
number of variables in the time series |
lag.order |
lag order of the model to be fit |
threshold |
threshold of calculation of recovery time (duration of perturbation), default value is 0.01 |
boot |
to bootstrap, default value is FALSE |
replication |
number of replication of bootstrap, default value is 200 |
steps |
number of steps of impulse response analysis, default value is 100 |
Value
iRAM matrix. Rows represent where the orthognal impulse was given, and columns represent the response. Dimension is var.number by var.number.
References
Lütkepohl, H. (2007). New introduction to multiple time-series analysis. Berlin: Springer.
Examples
boot.iRAM <- iRAM(model.fit = usemmodelfit,
beta = NULL,
var.number = 3,
lag.order = 1,
threshold = 0.01,
boot = TRUE,
replication = 200,
steps = 100
)
boot.iRAM$mean
Generate iRAM (impulse response anlaysis metric) in the equilibrium form.
Description
Generate iRAM (impulse response anlaysis metric) in the equilibrium form.
Usage
iRAM_equilibrium(beta.matrix, var.number, lag.order)
Arguments
beta.matrix |
beta matrix for a point estimate |
var.number |
number of variables in the time series |
lag.order |
lag order of the model to be fit |
Value
a list of equilibria. First numeric number in the variable name indicate where the impulse was given, and the second numeric number indicate the response, e.g., e12 indicates equilibrium of node 2 when node 1 is given an impulse.
Examples
iRAM_evalue <- iRAM_equilibrium(beta.matrix = true_beta_3node,
var.number = 3,
lag.order = 1
)
iRAM_evalue
Provide model summary.
Description
Provide model summary.
Usage
model_summary(model.fit, var.number, lag.order)
Arguments
model.fit |
model fit object generated by lavaan |
var.number |
number of variables in the time-series |
lag.order |
lag order of model |
Details
Model fit criteria: 3 out of 4 rule, meaning 3 out of 4 critea should be satisfied, including CFI and TLI should be greater than 0.95, RMSEA and SRMR should be less than 0.08.
Value
beta matrix estimates
matrix of standard error of beta
matrix of psi estimates
fit statistics CFI
fit statistics TLI
fit statistics RMSEA
fit statistics SRMR
Examples
mdl <- model_summary(model.fit = usemmodelfit,
var.number = 3,
lag.order = 1)
mdl$beta
mdl$beta.se
mdl$psi
mdl$cfi
mdl$tli
mdl$rmsea
mdl$srmr
Parse the beta from model fit object
Description
Parse the beta from model fit object
Usage
parse_beta(var.number, model.fit, lag.order, matrix = F)
Arguments
var.number |
number of variables in the time series |
model.fit |
model fit object generated by lavaan |
lag.order |
lag order of the model to be fit |
matrix |
output beta in matrix format or estimates format, default value is FALSE (as estimates) |
Value
beta
Examples
data(usemmodelfit)
beta.matrix <- parse_beta(var.number = 3,
model.fit = usemmodelfit,
lag.order = 1,
matrix = TRUE)
beta.matrix
Plot distribution of recovery time based on bootstrapped version of iRAM
Description
Plot distribution of recovery time based on bootstrapped version of iRAM
Usage
plot_iRAM_dist(recovery.time.reps)
Arguments
recovery.time.reps |
bootstrapped version of recovery time |
Examples
plot_iRAM_dist(bootstrap_iRAM_3node$recovery.time.reps)
Plot the time profiles in the integrated form
Description
Plot the time profiles in the integrated form
Usage
plot_integrated_time_profile(beta.matrix, var.number, lag.order = 1)
Arguments
beta.matrix |
matrix of temporal relations, cotaining both lag-1 and contemporaneous |
var.number |
number of variables in the time series |
lag.order |
lag order of the model to be fit |
Examples
plot_integrated_time_profile(beta.matrix = true_beta_3node,
var.number = 3,
lag.order = 1)
Plot the network graph
Description
Plot the network graph
Usage
plot_network_graph(beta, var.number)
Arguments
beta |
matrix of temporal relations, cotaining both lag-1 and contemporaneous |
var.number |
number of variables in the time series |
Examples
plot_network_graph(beta = true_beta_3node,
var.number = 3)
Plot time profiles given a time-series generated by impulse response analysis
Description
Plot time profiles given a time-series generated by impulse response analysis
Usage
plot_time_profile(time.series.data, var.number, threshold = 0.01, xupper = 20)
Arguments
time.series.data |
data of impulse response in long format |
var.number |
number of variables in the time-series |
threshold |
threshold of asymptote of equilibrium |
xupper |
upper limit of x-axis |
Examples
plot_time_profile(time.series.data = bootstrap_iRAM_2node$time.profile.data,
var.number = 2,
threshold= .01,
xupper = 20)
Simulated bivariate time-series data
Description
Simulated bivariate time-series data
Usage
simts_2node
Format
An object of class data.frame with 200 rows and 2 columns.
Details
Data simulated from a given three-node network structure with 200 measurements. Network structure is shown in the dataset true.beta. Process noise has mean of 0 and SD .1.
Examples
data(simts_2node)
Simulated 3-variate time-series data
Description
Simulated 3-variate time-series data
Usage
simts_3node
Format
An object of class data.frame with 100 rows and 3 columns.
Details
Data simulated from a given three-node network structure with 200 measurements. Network structure is shown in the dataset true.beta. Process noise has mean of 0 and SD .1.
Examples
data(simts_3node)
The true beta matrix (4 by 4) used in simulation.
Description
The true beta matrix (4 by 4) used in simulation.
Usage
true_beta_2node
Format
An object of class matrix (inherits from array) with 4 rows and 4 columns.
Details
true_beta_2node <- matrix(c(0,0,0,0, 0,0,0,0, 0.2,-.4,0,-0.25, 0,0.3,-0.2,0), nrow = 4, ncol = 4, byrow = TRUE)
Examples
true_beta_2node
The true beta matrix (6 by 6) used in simulation.
Description
The true beta matrix (6 by 6) used in simulation.
Usage
true_beta_3node
Format
An object of class matrix (inherits from array) with 6 rows and 6 columns.
Details
true_beta_3node <- matrix(c(0,0,0,0,0,0, 0,0,0,0,0,0, 0,0,0,0,0,0, 0.2,0,0.25,0,0,0.6, 0,0.3,0,-0.2,0,-0.6, 0,-0.2,0.3,0,0,0), nrow = 6, ncol = 6, byrow = TRUE)
Examples
true_beta_3node
Fit a multivariate time series with uSEM (unified Structural Equation Model).
Description
Fit a multivariate time series with uSEM (unified Structural Equation Model).
Usage
uSEM(var.number,
data,
lag.order = 1,
verbose = FALSE,
trim = FALSE)
Arguments
var.number |
number of variables in the time series |
data |
time series data, must be in long format |
lag.order |
lag order of the model to be fit, default value is 1. Note: Higher order (greater than 1) might not run. |
verbose |
print intermediate model fit (iterations), default value is FALSE |
trim |
to trim the insignificant betas (just one step, not iterative), default value is FALSE |
Details
The purpose of uSEM is to quantify the temporal relations (both contemporaneous and lag-1) between variables. Model specification and estimation can be found in the references.
Value
model fit object generated by lavaan
References
Kim, J., Zhu, W., Chang, L., Bentler, P. M., & Ernst, T. (2007). Unified Structural Equation Modeling Approach for the Analysis of Multisubject, Multivariate Functional MRI Data. Human Brain Mapping, 93, 85–93. doi:10.1002/hbm.20259
Gates, K. M., & Molenaar, P. C. M. (2012). Group search algorithm recovers effective connectivity maps for individuals in homogeneous and heterogeneous samples. NeuroImage 63(1), 310-319. doi: 10.1016/j.neuroimage.2012.06.026
Gates, K. M., Molenaar, P. C. M., Hillary, F. G., Ram, N., & Rovine, M. J. (2010). Automatic search for fMRI connectivity mapping: An alternative to Granger causality testing using formal equivalences among SEM path modeling, VAR, and unified SEM. NeuroImage, 50(3), 1118–1125. doi: 10.1016/j.neuroimage.2009.12.117
Examples
model.fit <- uSEM(var.number = 3,
data = simts_3node,
lag.order = 1,
verbose = FALSE,
trim = FALSE)
model.fit
Model fitbased on similated time-series by uSEM.
Description
Model fitbased on similated time-series by uSEM.
Usage
usemmodelfit
Format
An object of class lavaan of length 1.
Examples
data(usemmodelfit)