jeya-maria-jose / cuff_less_bp_prediction Goto Github PK

how to get 'cleaned_further.csv'?

I have ppg signal images and I was wondering that is there a way to extract the features from them rather than from Part_n.mat files?

Hi,Thank you for your work.
The code in ppg_features.m line 72.

if(O1P(1,i)<=(v(j)*pk(1,1)+ pk1(1,1)))
Should the correct code be
if(O1P(1,i)<=(v(j)*pk(1,1)+ pk1(1,2)))

I am looking forward to your reply, thank you.

Is that also extract the fixed features point from ecg or ppg? or put the whole signal of these to network depend on the cycle or other factors？

Is it possible to use your ML code and data set to estimate blood pressure only with ppg signal?

can you tell me where the data you provided through data link is originated from? is physionet? if yes, which dataset?

Clear.

1. I noticed a feature named 'alpha' in your file - feature_extraction_xx.m, does thie feature has any Physiological significance？ could you give any link about this feature definition?
2. S=real(ifft(fft(ppg))), can you tell me the function of this code?
   thank you very much! Waiting your reply.

Hi, I tried downloading the dataset from the google drive link you provided. However, I feel there's some mismatch between the length of the cleaned dataset.

Number of test CSV files: 251 | Number of train CSV files: 4011 | Number of Val CSV files: 751
Total number of CSV files: 5013

meanwhile

Number of rows in new_bp.csv: 5011
Number of rows in bp_test_new.csv: 250
Number of rows in bp_train_new.csv: 4009
Number of rows in bp_val_new.csv: 750
test+train+val: 5009

Do I miss something?