| Type: | Package |
| Title: | Fits the Bayesian Piecewise Linear Log-Hazard Model |
| Version: | 1.5 |
| Date: | 2022-10-19 |
| Author: | Andrew G Chapple |
| Maintainer: | Andrew G Chapple <achapp@lsuhsc.edu> |
| Description: | Contains posterior samplers for the Bayesian piecewise linear log-hazard and piecewise exponential hazard models, including Cox models. Posterior mean restricted survival times are also computed for non-Cox an Cox models with only treatment indicators. The ApproxMean() function can be used to estimate restricted posterior mean survival times given a vector of patient covariates in the Cox model. Functions included to return the posterior mean hazard and survival functions for the piecewise exponential and piecewise linear log-hazard models. Chapple, AG, Peak, T, Hemal, A (2020). Under Revision. |
| License: | GPL-2 |
| Encoding: | UTF-8 |
| Imports: | Rcpp (≥ 0.12.18) |
| LinkingTo: | Rcpp, RcppArmadillo |
| RoxygenNote: | 7.1.1 |
| NeedsCompilation: | yes |
| Packaged: | 2022-10-19 13:41:46 UTC; achapp |
| Repository: | CRAN |
| Date/Publication: | 2022-10-20 14:08:01 UTC |
Returns the approximate restricted posterior mean survival for the PLLH model.
Description
Uses a grid and parameter values to approximate the restricted posterior mean survival for the PLLH model using the integral of the survival function.
Usage
ApproxMean(Y, s, lam, J)
Arguments
Y |
Sequence from 0.01 to the maximum observed event time used to compute the approximate restricted mean survival time. Smaller spaced sequences results in better approximation but longer computation time. |
s |
Vector of split points. The first and last entries must be 0 and max(Y). |
lam |
Vector of log-hazard values at each split point location. Must be same length as s. |
J |
Number of split points. |
Value
Returns the approximate restricted posterior mean survival time for the PLLH model.
Examples
##Generate Data
Y1=rweibull(100,4,1)
##Create sequence from (0,max(Y1)) for approximation
Y=seq(.01,max(Y1),.01)
##Parameters used to approximate the mean
s=c(0,1,max(Y1))
lam=c(-2,0,-2)
J=1
ApproxMean( Y, s, lam, J)
Samples from the PEH model without covariates.
Description
Samples from the Piecewise Exponential Hazard (PEH) model and returns a list containing posterior parameters and posterior restricted mean survival.
Usage
BayesPiecewiseHazard(Y, I1, Poi, B)
Arguments
Y |
Vector of event or censoring times. |
I1 |
Vector of event indicators. |
Poi |
Prior mean number of split points. |
B |
Number of iterations for MCMC. |
Value
Returns a list containing posterior samples of (1) the split point locations, (2) the log-hazards at each split point, (3) the number of split points, (4) the variance parameter for the log-hazard values, (5) the posterior mean restricted survivial time.
Examples
##Generate Data
Y=rweibull(20,4,1)
I=rbinom(20,1,.5)
##Hyperparameter for number of split points
Poi=5
##Number of iterations for MCMC
B=200
BayesPiecewiseHazard( Y, I, Poi, B)
Samples from the PEH Cox model with a patient covariate vector.
Description
Samples from the Piecewise Exponential Hazard (PEH) Cox model with a patient covariate vector and returns a list containing posterior parameters and posterior restricted mean survival.
Usage
BayesPiecewiseHazardCOV(Y, I1, COV, Poi, B)
Arguments
Y |
Vector of event or censoring times. |
I1 |
Vector of event indicators. |
COV |
Matrix of size nxp containing p patient covariates. |
Poi |
Prior mean number of split points. |
B |
Number of iterations for MCMC. |
Value
Returns a list containing posterior samples of (1) the split point locations, (2) the log-hazards at each split point, (3) the number of split points, (4) the variance parameter for the log-hazard values, (5) the coefficients in the Cox model.
Examples
##Generate Data
Y=rweibull(20,4,1)
I=rbinom(20,1,.5)
COV = matrix(rnorm(40,0,1),ncol=2)
##Hyperparameter for number of split points
Poi=5
##Number of iterations for MCMC
B=200
BayesPiecewiseHazardCOV( Y, I,COV, Poi, B)
Samples from the PEH Cox model with a patient covariate vector.
Description
Samples from the Piecewise Linear Log-Hazard (PLLH) Cox model and returns a list containing posterior parameters and posterior restricted mean survival.
Usage
BayesPiecewiseHazardTrt(Y, I1, Trt, Poi, B)
Arguments
Y |
Vector of event or censoring times. |
I1 |
Vector of event indicators. |
Trt |
Vector containing patient treatment/control assignment. |
Poi |
Prior mean number of split points. |
B |
Number of iterations for MCMC. |
Value
Returns a list containing posterior samples of (1) the split point locations, (2) the log-hazards at each split point, (3) the number of split points, (4) the variance parameter for the log-hazard values, (5) the treatment coefficient, (6) the mean restricted survivial time of the control therapy, (7) the mean restricted survival time of the treatment therapy.
Examples
##Generate Data
Y=rweibull(20,4,1)
I=rbinom(20,1,.5)
Trt=rbinom(20,1,.5)
##Hyperparameter for number of split points
Poi=5
##Number of iterations for MCMC
B=200
BayesPiecewiseHazardTrt( Y, I,Trt, Poi, B)
Samples from the PLLH model without covariates.
Description
Samples from the Piecewise Linear Log-Hazard (PLLH) model and returns a list containing posterior parameters and posterior restricted mean survival.
Usage
BayesPiecewiseLinearLogHazard(Y, I1, Poi, B)
Arguments
Y |
Vector of event or censoring times. |
I1 |
Vector of event indicators. |
Poi |
Prior mean number of split points. |
B |
Number of iterations for MCMC. |
Value
Returns a list containing posterior samples of (1) the split point locations, (2) the log-hazards at each split point, (3) the number of split points, (4) the variance parameter for the log-hazard values, (5) the posterior mean restricted survivial time.
Examples
##Generate Data
Y=rweibull(20,4,1)
I=rbinom(20,1,.5)
##Hyperparameter for number of split points
Poi=5
##Number of iterations for MCMC
B=200
BayesPiecewiseLinearLogHazard( Y, I, Poi, B)
Samples from the PLLH Cox model with a patient covariate vector.
Description
Samples from the Piecewise Linear Log-Hazard (PLLH) Cox model with a patient covariate vector and returns a list containing posterior parameters and posterior restricted mean survival.
Usage
BayesPiecewiseLinearLogHazardCOV(Y, I1, COV, Poi, B)
Arguments
Y |
Vector of event or censoring times. |
I1 |
Vector of event indicators. |
COV |
Matrix of size nxp containing p patient covariates. |
Poi |
Prior mean number of split points. |
B |
Number of iterations for MCMC. |
Value
Returns a list containing posterior samples of (1) the split point locations, (2) the log-hazards at each split point, (3) the number of split points, (4) the variance parameter for the log-hazard values, (5) the coefficients in the Cox model.
Examples
##Generate Data
Y=rweibull(20,4,1)
I=rbinom(20,1,.5)
COV = matrix(rnorm(40,0,1),ncol=2)
##Hyperparameter for number of split points
Poi=5
##Number of iterations for MCMC
B=200
BayesPiecewiseLinearLogHazardCOV( Y, I,COV, Poi, B)
Samples from the PEH Cox model with a treatment indicator.
Description
Samples from the Piecewise Exponential Hazard (PEH) Cox model with a treatment indicator and returns a list containing posterior parameters and posterior restricted mean survival.
Usage
BayesPiecewiseLinearLogHazardTrt(Y, I1, Trt, Poi, B)
Arguments
Y |
Vector of event or censoring times. |
I1 |
Vector of event indicators. |
Trt |
Vector containing patient treatment/control assignment. |
Poi |
Prior mean number of split points. |
B |
Number of iterations for MCMC. |
Value
Returns a list containing posterior samples of (1) the split point locations, (2) the log-hazards at each split point, (3) the number of split points, (4) the variance parameter for the log-hazard values, (5) the treatment coefficient, (6) the mean restricted survivial time of the control therapy, (7) the mean restricted survival time of the treatment therapy.
Examples
##Generate Data
Y=rweibull(20,4,1)
I=rbinom(20,1,.5)
Trt=rbinom(20,1,.5)
##Hyperparameter for number of split points
Poi=5
##Number of iterations for MCMC
B=200
BayesPiecewiseLinearLogHazardTrt( Y, I,Trt, Poi, B)
Computes the posterior distribution of hazard value for a vector x for the Piecewise Linear Log Hazard model (PLLH)
Description
Computes the posterior distribution of hazard value for a vector x for the Piecewise Linear Log Hazard model (PLLH)
Usage
GetALLHazLogSlope(x, G1)
Arguments
x |
Vector of times to compute the posterior mean hazard function |
G1 |
List of posterior samples from the BayesPiecewiseLinearLogHazard function. |
Value
Matrix containing the posterior distribution of hazard values h(x)
Computes the posterior hazard values for a vector x for the Piecewise Exponential Hazard model (PEH)
Description
Computes the posterior hazard values for a vector x for the Piecewise Exponential Hazard model (PEH)
Usage
GetALLHazPiece(x, G1)
Arguments
x |
Vector of times to compute the hazard. |
G1 |
List of posterior samples from the BayesPiecewiseHazard function. |
Value
Matrix containing the posterior distribution of hazard values h(x)
Computes the posterior distribution of survival probabilities for a vector x for the Piecewise Exponential Hazard model (PEH)
Description
Computes the posterior distribution of survival probabilities for a vector x for the Piecewise Exponential Hazard model (PEH)
Usage
GetALLSurvPEH(x, G1)
Arguments
x |
Vector of times to compute the posterior mean survival probability. |
G1 |
List of posterior samples from the BayesPiecewiseLinearHazard function. |
Value
Matrix containing the posterior distribution of survival probabilities S(x)
Computes posterior distribution of survival probabilities for a vector x for the Piecewise Linear Log Hazard model (PLLH)
Description
Computes posterior distribution of survival probabilities for a vector x for the Piecewise Linear Log Hazard model (PLLH)
Usage
GetALLSurvPLLH(x, G1)
Arguments
x |
Vector of times to compute the posterior mean survival probability. |
G1 |
List of posterior samples from the BayesPiecewiseLinearLogHazard function. |
Value
Matrix containing the posterior distribution survival probabilities S(x)
Computes the posterior mean hazard value for a vector x for the Piecewise Linear Log Hazard model (PLLH)
Description
Computes the posterior mean hazard value for a vector x for the Piecewise Linear Log Hazard model (PLLH)
Usage
PostMeanHazLogSlope(x, G1)
Arguments
x |
Vector of times to compute the posterior mean hazard function |
G1 |
List of posterior samples from the BayesPiecewiseLinearLogHazard function. |
Value
Vector containing the posterior mean hazard values h(x)
Computes the posterior mean hazard values for a vector x for the Piecewise Exponential Hazard model (PEH)
Description
Computes the posterior mean hazard values for a vector x for the Piecewise Exponential Hazard model (PEH)
Usage
PostMeanHazPiece(x, G1)
Arguments
x |
Vector of times to compute the posterior mean hazard. |
G1 |
List of posterior samples from the BayesPiecewiseHazard function. |
Value
Vector containing the posterior mean hazard values h(x)
Computes the posterior mean survival probabilities for a vector x for the Piecewise Exponential Hazard model (PEH)
Description
Computes the posterior mean survival probabilities for a vector x for the Piecewise Exponential Hazard model (PEH)
Usage
PostMeanSurvPEH(x, G1)
Arguments
x |
Vector of times to compute the posterior mean survival probability. |
G1 |
List of posterior samples from the BayesPiecewiseLinearHazard function. |
Value
Vector containing the posterior mean survival probabilities S(x)
Computes the posterior mean survival probabilities for a vector x for the Piecewise Linear Log Hazard model (PLLH)
Description
Computes the posterior mean survival probabilities for a vector x for the Piecewise Linear Log Hazard model (PLLH)
Usage
PostMeanSurvPLLH(x, G1)
Arguments
x |
Vector of times to compute the posterior mean survival probability. |
G1 |
List of posterior samples from the BayesPiecewiseLinearLogHazard function. |
Value
Vector containing the posterior mean survival probabilities S(x)