Clinically Constrained and Simulated Multimodal Tear Biomarker Dataset for Non-Invasive Ocular and Systemic Risk Analysis

This dataset is a clinically constrained synthetic multimodal tear-fluid dataset created to support machine learning–based early risk prediction of ocular and systemic pathologies. The dataset integrates tear biochemical biomarkers (osmolarity, electrolytes, total protein, lactoferrin, lysozyme), inflammatory cytokines (IL-6, TNF-α), physiological indicators (blink rate, redness score), lifestyle parameters (screen time, sleep duration), and engineered diagnostic indices including Inflammation Index, Tear Film Dysfunction Index, and Composite Risk Score. All values were generated within physiologically and pathologically reported ranges derived from peer-reviewed ophthalmology, tear proteomics, and systemic disease literature. The dataset is intended for research, benchmarking, explainable AI analysis, and feasibility studies in non-invasive tear-based diagnostics. No real patient data are included.