A method of learning a pulse wave characteristic inference model by a processor includes analyzing a pulse wave index with respect to a plurality of pulse wave data for training, and clustering the pulse wave data for training into a plurality of pulse wave data clusters based on the pulse wave index Step, determining a membership function of the pulse wave data cluster based on a degree of membership indicating a degree to which the learning pulse wave data belongs to the pulse wave data cluster, and inferring a pulse wave characteristic based on the membership function It may include the step of generating a model that
展开▼