piiR

Predictive Information Index (PII): Quantifying predictive utility of scores

The piiR package provides tools for computing the Predictive Information Index (PII), which evaluates how much outcome-relevant information is retained in various types of scores (e.g., sum scores, CFA scores, subscale scores) in predictive models.

📦 Installation

```r # Install from GitHub remotes::install_github(“TheotherDrWells/piiR”)

Or from CRAN (once accepted)

install.packages(“piiR”)

🚀 Example

library(piiR)

set.seed(123) full <- rnorm(100) score <- full + rnorm(100, sd = 0.5)

Compute RMSE-based PII

pii(full, score, type = “rm”)

Try R²-based PII

pii(full, score, type = “r2”)

📘 Learn More Vignette: vignette(“piiR_intro”)

Docs: CRAN page (once available)