pridit - Composite Scoring via Principal Component Analysis of Ridit Scores

Implements 'PRIDIT' (Principal Component Analysis applied to 'RIDITs'), an unsupervised, nonparametric method for aggregating ordinal, categorical, and continuous indicators into a single interpretable composite score. Originally proposed by Brockett et al. (2002) <doi:10.1111/1539-6975.00027> for insurance fraud detection and extended to hospital quality measurement by Lieberthal (2008) <doi:10.1111/j.1475-6773.2007.00821.x> and Lieberthal and Comer (2013) <doi:10.1111/rmir.12009>. The package provides: (1) low-level functions ridit(), PRIDITweight(), and PRIDITscore(); (2) a unified pridit() entry point returning a classed object with print, summary, 'autoplot', and 'coef' methods; (3) pridit_boot() for bootstrap confidence intervals on scores and weights; (4) a step_pridit() recipe step for out-of-sample scoring within the 'tidymodels' framework; and (5) pridit_longitudinal() for panel data, computing cross-period stability of scores and weights.