Diastolic heart failure is a chronic condition which touch
450 000 people in UK each year. A model fitted to a patient gives strong
insight of the state of the heart of the patient. It could help to target
treatment at an individual level. Nevertheless fitting a cardiac model is
challenging because accurate cardiac models are slow. We propose a new
framework which consists to transform the cardiac model into a statistical
model called emulator. The emulator is used to select all the plausible sets of
parameters which would fit well the data.