anfis-based modeling of pulse parameters for objective determination of warmness, coldness, wetness, dryness, and heart strength in persian medicine

This study developed a computational framework to objectify pulse diagnosis in persian medicine by quantitatively predicting the primary temperamental qualities—warmness, coldness, wetness, dryness, and heart strength.methods: an adaptive neuro-fuzzy inference system (anfis) model was designed. using data from 68 participants, eleven pulse characteristics (e.g., length, strength, speed) were used as inputs, with the five temperament qualities serving as expert-defined outputs. the model was trained to learn the complex relationships between these inputs and outputs. results: the anfis model demonstrated exceptional predictive accuracy, achieving an average r² of 0.9322 and an average rmse of 0.2239 across all outputs. performance was notably high for dryness (r² = 0.9998) and heart strength (r² = 0.9891).conclusion: the high accuracy of the model validates its capability to emulate the sophisticated diagnostic reasoning of a persian medicine practitioner. this establishes anfis as a powerful tool for quantifying subjective pulse-based diagnosis.