Latent profiling of five-dimensional psychological resilience across generations: a deep clustering and behavioural divergence analysis in pre-conflict Iran

Psychological resilience is increasingly conceptualized as a multidimensional construct encompassing identity, emotional, cognitive, behavioural, and social domains. Using data from 620 Iranian adults (aged 18-64 years; 52% female), collected through an online self-report survey, this study applied unsupervised machine-learning techniques combining a deep autoencoder for dimensionality reduction with a Gaussian Mixture Model (GMM) for latent clustering-to examine psychological resilience profiles in pre-conflict Iran. Thirty-seven standardized psychological subscales were aggregated into five theoretically grounded dimensions: Self-Identity and Meaning, Emotional Regulation, Cognitive Flexibility, Coping and Growth, and Social Support and Connectedness. Unsupervised analysis identified four latent archetypes-Fragile Striver, Reactive Idealist, Hidden Reactor, and Stable Withdrawer-that reflected nonlinear configurations of resilience capacities across generational and gender groups. However, because the research employed a cross-sectional and self-report design, findings illustrate associative rather than causal relationships, and representativeness is limited to online participants. These contextual and demographic influences suggest that resilience is embedded within Iran's evolving social environment. Despite these limitations, the study demonstrates the potential of AI-based latent profiling to clarify the multidimensional nature of resilience within culturally demanding contexts.