A Multimodal Network Security Framework for Healthcare Based on Deep Learning

Algorithm 1

k-fold cross validation used to test the stability.

(1)	Begin:
(2)	Break the data into folds. Data capacity per fold is ;
(3)	Construct model joint-training based on CNN model and GRU model;
(4)	for each in do
(5)	validation data = data ;
(6)	training data = rest of data;
(7)	train CNN model on the training data, observe it on the validation data;
(8)	train GRU model on the training data, observe it on the validation data;
(9)	construct model pretraining based on CNN model and GRU model ;
(10)	train model pretraining and model joint-training;
(11)	Model pretraining prediction on the test data;
(12)	Model joint-training prediction on the test data;
(13)	end for
(14)	End