Symptom-Based Health Assessment Dataset for Healthcare NLP

This dataset contains 10,897 patient symptom records designed for research in healthcare natural language processing, triage modelling, and clinical decision-support experimentation. Each record includes demographic attributes, symptom lists, vital signs, and a natural-language description resembling patient-reported symptom narratives.The dataset was generated using structured symptom indicators derived from a publicly available medical symptom ontology. All records, values, demographics, and textual descriptions were artificially created using controlled randomization and templated narrative generation. No real patient information was used, and no identifiable data is present.This dataset supports tasks such as:Text classificationSymptom extractionEarly triage predictionHealthcare NLP benchmarkingEducational demonstrationsThe dataset is provided under a CC0 Public Domain license to enable unrestricted reuse, redistribution, modification, and integration into downstream machine learning workflows.