Data and Analysis Materials for "Interpretable Symbolic Regression for Psychological Questionnaire Data: Prediction, Descriptor Discovery, and Nonlinear Interaction Analysis"

his repository contains the data and statistical materials associated with the manuscript “Interpretable Symbolic Regression for Psychological Questionnaire Data: Prediction, Descriptor Discovery, and Nonlinear Interaction Analysis.” The study evaluates GoodRegressor, a hierarchical, depth-controlled symbolic-regression framework, using two independent psychological questionnaire datasets of middle-aged wives from Japan (n = 201) and South Korea (n = 220). Marital satisfaction was modeled using demographic and socioeconomic variables, family-viewpoint composites, marital conflict-resolution strategies, and systematically derived interaction descriptors. The Japanese and South Korean datasets were analyzed as independent methodological testbeds rather than as directly comparable cultural groups because the family-viewpoint measure did not show sufficient cross-national measurement invariance. The repository provides the raw data and analysis materials used for the reported statistical analyses, including descriptive statistics, confirmatory factor analyses and measurement-invariance tests, predictor and descriptor construction, five-fold held-out benchmark results for GoodRegressor and eight comparison methods, matched-sample sensitivity analyses, and statistical outputs supporting the full-data symbolic-model interpretation. Supplementary materials document the predictor definitions, model settings, benchmark metrics, symbol-stability results, and representative GoodRegressor expressions. These materials are provided to support transparency, reproducibility, and secondary methodological evaluation of the analyses reported in the manuscript. The discovered symbolic terms and interactions should be interpreted as exploratory, noncausal hypotheses rather than as evidence of causal effects or national differences in psychological mechanisms.