The Classification Window Calculator Manual
Installation (Windows Only)
To install the package download cwcalc.zip and
unzip it wherever you like. The created directories will contain the
file ClassificationWindowCalculator.exe, which is the
executable for running the calculator.
Using the Calculator
In order to run the calculator, select the calculation you would like,
fill in the required fields (in white) and hit "Calculate", an
hourglass will appear while the calculator runs its calculations (this
may take a while). Fields and buttons on the calculator have the
following meanings:
Disease Properties
m: The median incubation period of the disease we are
interested in classifying.
d: The dispersion of the incubation period. This is the
anti-log of the standard deviation of the incubation period on the log
scale. The calculator assumes that disease incubation periods follow a
log normal distribution.
R: The incidence rate ratio or relative risk of acquiring
the disease on interest in the hospital versus the acquiring it in the
community.
Window size and performance
w:The length of the classifying window, in days, that we wish to
evaluate, or that was calculated as optimal given the desired
criteria.
ppv: The desired or calculated positive predictive value of
the classifying window. A number between 0 and 1.
ppv: The desired or calculated negative predictive value of
the classifying window. A number between 0 and 1.
Calculation Options
ppv and npv given w Takes in the disease parameters and a
classifying window length, and gives the expected positive and
negative predictive values of this window for classifying the source
of infection for this disease.
npv and w given ppv Takes the disease parameters and a
desired positive predictive value and returns the maximum negative
predictive value achievable and the classifying window length
achieving this performance. (can be slow)
ppv and w given npv Takes the disease parameters and a
desired negative predictive value and returns the maximum positive
predictive value achievable and the classifying window length
achieving this performance. (can be slow)
all with npv=ppv Takes the disease parameters and finds
the classifying window length where the positive and negative
predictive value are equal.