TPDT

Time resolved paired differences test

TPDT is an R package suited for the analysis of time-resolved observations. More specifically, we provide a collection of state of the art tools which all have the task of exploring the difference of two groups of functional observations over a specified time interval. Moreover, we provide an additional tool - the TPDT (Time-resolved Paired Differences Test) - which in some situations has a better performance in terms of AUC and power compared to the here implemented competitors (publication to appear soon).

Installation

1. Install devtools from CRAN with install.packages("devtools").

2. Until this package is published on CRAN, you can install it from github

with devtools::install_github("toreerdmann/TPDT")

After that you load it as usual with library(TPDT)

Demo

library(TPDT)

# Simulate paired data with underlying function f with
# a shift of 5 between the groups
f <- function(x) 2 * x * sin(x) + 10
simdata <- make_data(shift = 5, n = 2, sd1 = .5, sd2 = .5, 
                     ntimepoints = 10,type = "shift", f = f)
# run test
result <- TPDT(simdata, B = 200) 

# plot data and pvalue from the test
matplot(y = matrix(simdata$data, nrow = 10), 
        x = matrix(simdata$time, nrow = 10), 
        main = paste("pvalue =", result$p), 
        ylab = "y", xlab = "time", type = "b", lwd = 3)
plot(result$funcdata$func1, add = TRUE, lwd = 2)
plot(result$funcdata$func2, add = TRUE, lwd = 2, 
     col = c(3, 4))