outlier_stats: Outlier statistics and outlier gains

In guidopowell/scv: SCV

Description

Outlier detection from benchmark Poisson-Gamma distribution (a distribution drawn from a benchmark SCV). Outlier gains are the additions or reductions (depending on whether it is a desired outcome or not) to the system that would occur under a scenario where all identified upper outliers or lower outliers are brought into the 95% bounds of the benchmark distribution

Usage

outlier_stats(obsx, expx, scvtx = 3, alphax = 0.95)

Arguments

obsx	observed cases for a given population
expx	expected cases from standardization of given population
scvtx	benchmark SCV, default 3
alphax	alpha level, default .95

Value

obs

exp

upper_lim_qnbinom Upper limit O/E of benchmark Poisson-Gamma distribution

lower_lim_qnbinom Lower limit O/E of benchmark Poisson-Gamma distribution

is_outlier Identifes O/E as upper or lower outlier (else is NULL)

pval_ref Probability of O/E ratio in benchmark Poisson-Gamma distribution

outlier_diff Difference between outlier and benchmark upper or lower limit

upper_outlier_gain System gains under scenario of normalizing upper outliers

lower_outlier_gain System gains under scenario of normalizing lower outliers

Examples

data(nc.sids, package= "spData")

obs<-nc.sids$SID74

exp<-nc.sids$BIR74*sum(nc.sids$SID74)/sum(nc.sids$BIR74)

outlier_stats(obs,exp,scvtx=3, alphax=.95)

guidopowell/scv documentation built on Nov. 4, 2019, 1:24 p.m.