psi: Population Stability Index

In siyuany/rpsi: Population Stability Index

Description

psi calculates the popolation stability index.

Usage

psi(original, current, cut.points = NULL)

Arguments

original	The original set of a measurement, should be a factor or numeric
current	The current set of a measurement, should be a factor or numeric
cut.points	It won't work if original and current are factors, and it cannot be NULL if original and current are numerical. This function use this argument to binning original and current with left-closed right-open interval.

Details

psi measures the stablity of the population. Usually we can believe the population stays the same as the past if psi is less than 0.1, and a significant shift can be recognised if psi is greater than 0.25. The outcome of this function is a numeric, with details stored as attributes. You can use summary function to see all of the detailed information. Fot the situation where some of the levels has no element in either original population or current population and the psi does not exist for such levels, the empty levels will not be taken into account and a warning will inform you of this. Again, by using summary you could know everything inside.

Value

a psi object

Author(s)

Siyuan Yao

References

https://www.quora.com/What-is-population-stability-index

Examples

data("iris")
train <- sample(nrow(iris), nrow(iris) * .7)
train.species <- iris$Species[train]
test.species <- iris$Species[-train]
p <- psi(train.species, test.species)
p
summary(p)

siyuany/rpsi documentation built on May 29, 2019, 11 p.m.