Blood pressure prediction method and electronic device using the same

1. A blood pressure prediction method, adapted to an electronic apparatus comprising a processor and a storage circuit, comprising:

collecting a training data set;

establishing a first blood pressure prediction model according to the training data set;

receiving hemodialysis parameter data of a target patient, wherein the hemodialysis parameter data comprises a first hemodialysis parameter at a previous time point and a second hemodialysis parameter at a current time point;

calculating a hemodialysis parameter variation amount between the first hemodialysis parameter and the second hemodialysis parameter;

providing the hemodialysis parameter variation amount to the first blood pressure prediction model to generate a prediction blood pressure variation associated with a next time point; and

in response to the prediction blood pressure variation associated with the next time point being greater than an alert threshold, performing an operation according to the prediction blood pressure variation of the target patient, wherein the operation comprising:

generating a suggestion hemodialysis parameter value according to the first blood pressure prediction model and the second hemodialysis parameter at the current time point through minimizing an output of the first blood pressure prediction model; and

prompting the suggestion hemodialysis parameter value or set a dialysis apparatus according to the suggestion hemodialysis parameter value.

2. The blood pressure prediction method according to claim 1 , further comprising:

receiving physiological data and medical record data of the target patient and climate data; and

providing the physiological data, the medical record data and the climate data to the first blood pressure prediction model to generate the prediction blood pressure variation associated with the next time point.

3. The blood pressure prediction method according to claim 1 , wherein the first hemodialysis parameter and the second hemodialysis parameter comprise ultrafiltration rates, dialysate concentrations, dialysate temperatures or blood flow rates.

4. The blood pressure prediction method according to claim 1 , further comprising:

in response to the prediction blood pressure variation associated with the next time point being greater than the alert threshold, issuing an alarm notification.

5. The blood pressure prediction method according to claim 4 , further comprising:

receiving a test hemodialysis parameter;

providing another hemodialysis parameter variation amount between the second hemodialysis parameter and the test hemodialysis parameter to the first blood pressure prediction model to generate a simulation blood pressure variation; and

prompting the simulation blood pressure variation.

6. The blood pressure prediction method according to claim 1 , wherein the step of establishing the first blood pressure prediction model according to the training data set comprises:

acquiring a plurality of important feature variables from the training data set;

generating a classification factor of each patient according to the important feature variables of each patient, and classifying the patients into a plurality of patient clusters according to the classification factor of each patient; and

for each of the patient clusters, respectively training a plurality of sub training data sets respectively corresponding to the patient clusters in the training data set according to a machine learning algorithm to generate a plurality of second blood pressure prediction models respectively corresponding to the patient clusters, wherein the second blood pressure prediction models comprise the first blood pressure prediction model.

7. The blood pressure prediction method according to claim 6 , further comprising:

determining the target patient as belonging to one of the patient clusters; and

selecting the first blood pressure prediction model corresponding to the one of the patient clusters from the second blood pressure prediction models.

8. The blood pressure prediction method according to claim 6 , wherein the step of generating the classification factor of each patient according to the important feature variables of each patient and classifying the patients into the patient cluster according to the classification factor of each patient comprises:

calculating mutual information of each patient according to the important feature variables of each patient; and

clustering the patients into the patient clusters by comparing the mutual information of each patient with at least one cluster threshold.

9. The blood pressure prediction method according to claim 6 , wherein the machine learning algorithm is a supervised machine learning algorithm.

10. An electronic device, comprising:

a storage circuit, storing a plurality of modules; and

a processor, coupled to the storage circuit and configured to access the modules to:

collect a training data set;

establish a first blood pressure prediction model according to the training data set;

receive hemodialysis parameter data of a target patient, wherein the hemodialysis parameter data comprises a first hemodialysis parameter at a previous time point and a second hemodialysis parameter at a current time point;

calculate a hemodialysis parameter variation amount between the first hemodialysis parameter and the second hemodialysis parameter;

provide the hemodialysis parameter variation amount to the first blood pressure prediction model to generate a prediction blood pressure variation associated with a next time point; and

in response to the prediction blood pressure variation associated with the next time point being greater than an alert threshold, perform an operation according to the prediction blood pressure variation of the target patient, wherein the operation comprising:

generating a suggestion hemodialysis parameter value according to the first blood pressure prediction model and the second hemodialysis parameter at the current time point through minimizing an output of the first blood pressure prediction model; and

prompting the suggestion hemodialysis parameter value or set a dialysis apparatus according to the suggestion hemodialysis parameter value.

11. The electronic device according to claim 10 , wherein the processor is configured to:

receive physiological data and medical record data of the target patient and climate data; and

provide the physiological data, the medical record data and the climate data to the first blood pressure prediction model to generate the prediction blood pressure variation associated with the next time point.

12. The electronic device according to claim 10 , wherein the first hemodialysis parameter and the second hemodialysis parameter comprise ultrafiltration rates, dialysate concentrations, dialysate temperatures or blood flow rates.

13. The electronic device according to claim 10 , wherein the processor is configured to:

in response to the prediction blood pressure variation associated with the next time point being greater than the alert threshold, issue an alarm notification.

14. The electronic device according to claim 13 , wherein the processor is configured to:

receive a test hemodialysis parameter;

provide another hemodialysis parameter variation amount between the second hemodialysis parameter and the test hemodialysis parameter to the first blood pressure prediction model to generate a simulation blood pressure variation; and

prompt the simulation blood pressure variation.

15. The electronic device according to claim 10 , wherein the processor is configured to:

acquire a plurality of important feature variables from the training data set;

generate a classification factor of each patient according to the important feature variables of each patient, and classify the patients into a plurality of patient clusters according to the classification factor of each patient; and

for each of the patient clusters, respectively train a plurality of sub training data sets respectively corresponding to the patient clusters in the training data set according to a machine learning algorithm to generate a plurality of second blood pressure prediction models respectively corresponding to the patient clusters, wherein the second blood pressure prediction models comprise the first blood pressure prediction model.

16. The electronic device according to claim 15 , wherein the processor is configured to:

determine the target patient as belonging to one of the patient clusters; and

select the first blood pressure prediction model corresponding to the one of the patient clusters from the second blood pressure prediction models.

17. The electronic device according to claim 15 , wherein the processor is configured to:

calculate mutual information of each patient according to the important feature variables of each patient; and

cluster the patients into the patient clusters by comparing the mutual information of each patient with at least one cluster threshold.

18. The electronic device according to claim 15 , wherein the machine learning algorithm is a supervised machine learning algorithm.