computeOverlap: Compute overlap score between two jobs, and generate plot

In saharaja/imputeORS: Impute Missing Values in the BLS Occupational Requirements Survey

Description

Thanks to imputation, we arrive at a full distribution of population percentages for all jobs, across all requirements. Using this information, we can devise a way to measure the requirement "overlap" of any two occupations. We do this by taking the product of their estimates to yield a value on the interval [0,1]. This produces a way to measure similarity between different occupations.

Usage

computeOverlap(blended.results, jobA, jobB, print.plot = TRUE)

Arguments

blended.results	Blended predictions from imputation models, calculated at convergence iterations and blending proportions computed by computeBlendingRatio() (output of blendImputations())
jobA	First job to compare
jobB	Second job to compare
print.plot	Should plot file (.png) be generated; default is TRUE (create plot file)

Details

If ω_1ir is the population mean of the ith level of the rth requirement for Job 1 (average of simulation predictions for missing values, and actual value for known estimates), and ω_2ir is the same for Job 2, then we say that the overlap of rth requirement (O_r) for these two jobs is:

O_r = ∑(ω_1ir * ω_2ir)

Value

A list of length three, containing a data frame with all the data pertaining to the specified occupations, a data frame of the overlap values for these occupations by requirement, and a plot object displaying these overlap values; optionally produces a plot (.png file) of the overlap

See Also

blendImputations()

saharaja/imputeORS documentation built on Feb. 4, 2022, 12:27 a.m.