search

shawnlxh / Blood_Pressure_Prediction

31 stars 17 forks source link

The documentation is not clear about train_data #1

Open jagarzone6 opened 5 years ago

jagarzone6 commented 5 years ago

I would like a better explanation of what train_data really has and means for the model. Right now I only see random values between 0 and 1 that does not see related to " information data, such as height, weight, sex, age, and the observation data, including systolic blood pressure, diastolic blood pressure, heart rate, the time of the measure, whether taking in drug, mood." So a explanation of it would help.

Same for test data I dont know exactly what I am testing. What is input and output in human terms ?

If I generate a If I graph model .pb, how would I use it in a android app ? I am able to run the code in tensorflow v 0.12. Had to modified the batch_size to make it run with the train_data.

shawnlxh commented 5 years ago

For preprocessing part, I made gender and time (month) into one-hot vector and all other numerical values are converted within 0 and 1 by max-min regularizer. For blood pressure measurement data, it is a time series and I put it into RNN to train the model.

Sorry for my late response. Hope it can help you.