"AyurKOSH Dataset_ A Machine-Readable Ayurvedic Knowledge Resource for Knowledge Graph and Computational Intelligence"

"Ayurveda, India\u2019s traditional medical system, contains highly interconnected knowledge regarding diseases, symptoms, herbs, and formulations (compounds). However, texts such as Charaka Samhita, Bhaishajya Ratnawali remain largely unstructured and inaccessible for computational analysis. This work presents AyurKOSH, a machine-readable, high-quality Ayurvedic dataset designed as a Knowledge Graph (KG) to support Artificial Intelligence-driven research. AyurKOSH contains structured information on diseases, associated symptoms, drug formulations, herbal components, and detailed pharmacological properties including Rasa, Guna, Virya, Vipaka, and Karma. Additionally, botanical metadata and herb-substitution relationships are included for synergy prediction and drug repurposing. The dataset enables applications in biomedical NLP, automated reasoning systems, clinical decision assistance, and pedagogy for integrative medicine. AyurKOSH can used for academic and non-commercial research.To summarize, this dataset presents a machine-readable Ayurvedic Knowledge Graph constructed from classical Ayurvedic texts and curated domain sources. It contains structured entities such as Vy\u0101dhi (diseases), Lak\u1e63a\u1e47a (symptoms), Aushadhi (herbs), and compounds, along with their semantic relationships. The dataset is designed to support NLP, knowledge graph construction, graph neural networks, and computational Ayurveda research."