Hi there Chandan, First of all thank you for your help!
We have chosen not to use adaptive calibration because we found that further calibrations would disturb the overall accuracy of the algorithm. But once again NuPIC got me thinking… If NuPIC is continuously adapting the model, there’s no need for further calibrations in our case because the blood pressure is known to be a very pattern-filled data varying in time, much like the PPG signal and the ECG signal. So more and more I’m getting to think that HTM would even improve our accuracy to levels we would never experience with our current algo! Thank you once again and I guess I’ll be going to do a lot more talking in this community from now on! :D Gonçalo Sá
