eribul / rinca Goto Github PK

Generate functions to generate a fake but realistic data set. Then use these functions to genereta (on or more) fake datasets to be used for examples.

Take a real data set from INCA
substitute personnummer with random dates ~U(1900, currentYear - 18). This distribution can be changed and mayby updated later with something more realistic. The four last digits could just be randomly assigned U(1000, 9999) or similar. But it should be possible to update later with a more realistic numers (rules from Skatteverket or Wikipedia).
Diagnos date and/or similair could be keped but reordered randomly (to not conect dates with patients).
Later dates can be assigned from diadat + X where XExp(lambda) with some more or less realistic distribution.
KON_VALUE could also be taken from original data (to keep the propoortion between the genders) but be shuffled.
Same with codes for hospital, clinics and PAT/CYT code.
LKF-codes must be kept to not confuse hospitals etc between regions.
Patients, doctors and monitor names should initialy be just removed. We might later generate fake names but that has low priority!

• I själva skripten
• I testet
• I exempel

avs 3.4.3 i Advanced R kanske ger alernativ till logsplines?

# Select 6 bootstrap replicates
df[sample(nrow(df), 6, rep = T), ]
#>     x y z
#> 2   1 5 b
#> 6   3 1 f
#> 5   3 2 e
#> 1   1 6 a
#> 6.1 3 1 f
#> 3   2 4 c

Package to generate random names:

http://cran.r-project.org/web/packages/randomNames/randomNames.pdf

http://www.ihsn.org/home/software/disclosure-control-toolbox

Problem med Travis annars

koda om code-funktionen map method sispatch:

UseMethod() takes two arguments: the name of the generic function, and the argument to use for method dispatch. If you omit the second argument it will dispatch on the first argument to the function. There’s no need to pass any of the arguments of the generic to UseMethod() and you shouldn’t do so. UseMethod() uses black magic to find them out for itself.

Mayby later, inspiration for survival analysis data

http://www.jstatsoft.org/v59/i02