Acute hypotension hybrid early warning method based on LSTM network

Abstract

The invention discloses an acute hypotension hybrid early warning method based on an LSTM network. The method comprises the following steps: firstly, a collected physiological data sequence is imported; hypotension event detection is carried out; if a patient is in a hypotension state at present, a result is output for hypotension early warning, if the patient is not in the hypotension state at present, predicting of an acute hypotension event is carried out, if the acute hypotension is predicted to occur in a next time period, the result is output for hypotension prediction early warning, then a blood pressure protection value is detected, and if the blood pressure exceeds the protection value, returning to two results is carried out for early warning. The prediction and early warning canbe carried out on acute hypotension, so that time is bought for monitoring personnel, the safety of the patient during rehabilitation training is guaranteed, and cardiac exercise rehabilitation becomes more scientific, efficient and safer.

Description

technical field [0001] The invention relates to the field of medical technology, in particular to a mixed early warning method for acute hypotension based on an LSTM network. Background technique [0002] Blood pressure is an important indicator of a person's health status, and it is also one of the main characteristics reflecting cardiovascular disease. Excessive blood pressure is prone to emergencies such as myocardial infarction, heart failure, and cerebral hemorrhage. In order to prevent adverse changes in blood pressure from causing serious consequences, it is very important to predict blood pressure in advance to monitor blood pressure changes. [0003] With the advancement of the concept of cardiovascular disease treatment, more and more medical workers began to advocate cardiac rehabilitation. Cardiac exercise rehabilitation, as one of the five major treatment methods of cardiac rehabilitation, has also attracted more and more attention. Cardiac exercise rehabilit...

AI Technical Summary

Problems solved by technology

Among them, the blood pressure prediction is carried out through the echo state network in the machine learning algorithm. The echo state network is proposed to solve the problem of gradient disappearance and gradient explosion in the traditional recurrent neural network (RNN). It adds a reserve pool to the RNN. Simple time-series data is memorized, but the network can only perform short-term memory and cannot handle complex dynamic problems well; the BP neural network and radial basis neural network in the neural network algorithm are respectively used to predict the diastolic state of the user by using the user's personal information to remind users whose diastolic blood pressure is higher than the normal range, and establish a relationship between user information and diastolic blood pressure. blood pressure warning

Images