This project is the R-implement version of pk4adi. Please refer the doc here for more information.
The package's name rpk4adi is short for "R-implement PK for anesthetic depth indicators". The PK (Prediction probability) was first proposed by Dr. Warren D. Smith in the paper Measuring the Performance of Anesthetic Depth Indicators in 1996. Dr. Warren D. Smith and his team provide a tool to calculate PK written using the MS Excel macro language.
Our team provide a reimplementation of the PK tools developed using the R language with easy-to-use APIs in this package. The project is fully open source on github. The latest released version could be found here.
A GUI version of pk4adi called pk4adi_gui is also under development. This project is also open source on github.
Please feel free to contact us ([email protected]). Any kind of feedback is welcome. You could report any bugs or issues when using pk4adi on github project.
Please refer the changelog.md for details.
data.table >= 1.10
stats
To install rpk4adi, run the following in the command prompt.
install.packages('pk4adi')
- calculate_pk
calculate_pk <- function(x_in, y_in)
@title Compute the pk value to Measure the Performance of Anesthetic Depth Indicators.
@param x_in a vector, the indicator.
@param y_in a vector, the state.
@return a list containing all the matrices and variables during the calculation.
The value list$type is "PK", which indicated the list is return-value of the function calculate_pk().
The type of list$basic is also a list, which contains the most important results of the function.
The type of list$matrices is also a list, which contains all the matrices during the calculation.
The type of list$details is also a list, which contains all the intermediate variables during the calculation.- compare_pks()
compare_pks <- function(pk1, pk2)
@title Compare two answers of the pk values.
@description Both of the two input have to be the output of the function calculate_pk().
@param pk1 a list, the output of the function calculate_pk().
@param pk2 a list, the output of the function calculate_pk().
@return a list containing all the variables during the calculation.
The value list$type is "PKC", which indicated the list is return-value of the function compare_pk().
The type of list$group is also a list, which contains the normal distribution test results for the group variables.
The type of list$pair is also a list, which contains the t distribution test results for the pair variables.
The type of list$details is also a list, which contains all the intermediate variables during the calculation.The best way to use this package is to use R scripts.
x1 <- c(0, 0, 0, 0, 0, 0)
y1 <- c(1, 1, 1, 1, 1, 2)
ans1 <- calculate_pk(x1, y1)
## show the most important results.
print(ans1$basic)
x2 <- c(1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 4, 5, 5, 5, 5, 6, 6, 6, 6, 6)
y2 <- c(1, 1, 1, 1, 1, 2, 1, 1, 3, 3, 2, 2, 2, 2, 2, 1, 3, 3, 3, 3, 3, 3, 3, 3)
ans2 <- calculate_pk(x2, y2)
## show the full results.
print(ans2$basic)You will get the following output.
$PK
[1] 0.5
$SE0
[1] 0
$SE1
[1] 0
$jack_ok
[1] FALSE
$PKj
[1] NaN
$SEj
[1] NaN
$PK
[1] 0.8670213
$SE0
[1] 0.06503734
$SE1
[1] 0.06587109
$jack_ok
[1] TRUE
$PKj
[1] 0.8664848
$SEj
[1] 0.07011821
x1 <- c(1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 4, 5, 5, 5, 5, 6, 6, 6, 6, 6)
y1 <- c(1, 1, 1, 1, 1, 2, 1, 1, 3, 3, 2, 2, 2, 2, 2, 1, 3, 3, 3, 3, 3, 3, 3, 3)
pk1 <- calculate_pk(x_in = x1, y_in = y1)
x2 <- c(1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 4, 5, 5, 5, 5, 6, 6, 6, 6, 6)
y2 <- c(1, 1, 2, 1, 1, 2, 1, 2, 3, 3, 2, 2, 1, 2, 2, 2, 3, 3, 3, 3, 2, 3, 3, 2)
pk2 <- calculate_pk(x_in = x2, y_in = y2)
ans <- compare_pks(pk1, pk2)
print(ans$group)
print(ans$pair)You will get the following output.
$PKD
[1] 0.06757172
$SED
[1] 0.1010385
$ZD
[1] 0.6687717
$ZP
[1] 0.5036411
$ZJ
[1] "P > 0.05"
$DF
[1] 23
$PKDJ
[1] 0.02971846
$SEDJ
[1] 0.06558182
$TD
[1] 0.4531508
$TP
[1] 0.3273431
$TJ
[1] "P > 0.05"You could get the all the matrices and variables in the returned lists of the function calculate_pk() and compare_pks(). Then just get the value with the key of the lists!
Please feel free to contact us ([email protected]). Any kind of feedback is welcome and appreciated.