OBJECTIVE: To compute quantitative estimates of the tridosha--the qualitative characterization that constitutes the core of diagnosis and treatment in Ayurveda--to provide a basis for biostatistical analysis of this ancient Indian science, which is a promising field of alternative medicine. SUBJECTS: The data sources were 280 persons from among the residents and visitors/training students at the Brahmvarchas Research Centre and Shantikuj, Hardwar, India. DESIGN/METHODOLOGY: A quantitative measure of the tridosha level (for vata, pitta, and kapha) is obtained by applying an algorithmic heuristic approach to the exhaustive list of qualitative features/factors that are commonly used by Ayurvedic doctors. A knowledge-based concept of worth coefficients and fuzzy multiattribute decision functions are used here for regression modeling. VALIDATION AND APPLICATIONS: Statistical validation on a large sample shows the accuracy of this study's estimates with statistical confidence level above 90%. The estimates are also suited for diagnostic and prognostic applications and systematic drug-response analysis of Ayurvedic (herbal and rasayanam) medicines. An application with regard to the former is elucidated, extensions of which might also be of use in investigating the role of nadis in Ayurvedic healing vis-a-vis acupuncture and acupressure techniques. The importance and scope of this novel approach are discussed. CONCLUSIONS: This pioneering study shows that the concept of tridosha has a sound empirical basis that could be used for the scientific establishment of Ayurveda in a new light.