A long-term risk prediction system for hemodialysis complications based on convolutional survival network

Abstract

The invention discloses a long-term risk prediction system for hemodialysis complications based on a convolutional survival network. The system includes a data acquisition module, a data preprocessing module, a learning prediction module and a result display module; Dimensional hemodialysis timing characteristics; convolutional neural network combined with Cox proportional hazards assumption, a convolutional survival network was proposed; on the basis of using convolutional survival network, Breslow was used to estimate the baseline risk function to calculate the long-term risk changes of patients. The present invention can make full use of common censored data in medical research; apply the main structure of convolutional neural network, facilitate visual analysis, and produce interpretable and enlightening results; and can predict long-term risk changes of patients.

Description

technical field [0001] The invention belongs to the technical field of medical treatment and machine learning, and in particular relates to a long-term risk prediction system for hemodialysis complications based on a convolutional survival network. Background technique [0002] The incidence of end-stage renal disease is on the rise worldwide, causing a huge disease burden. Most patients require hemodialysis (hemodialysis) to stay alive. Vascular access infection, hypertension, coronary heart disease and other concurrent diseases that may occur during long-term hemodialysis seriously affect the survival of patients, and have a huge impact and burden on patients, patients' families and society. Therefore, long-term risk prediction and early preventive treatment of hemodialysis complications are crucial to improving the quality of life of patients with end-stage renal disease. A large amount of time-series data has been accumulated during long-term hemodialysis, which brings...

AI Technical Summary

Problems solved by technology

However, the features recognized by the cyclic neural network are complex and abstract, and it is difficult to visualize and understand and provide heuristic results.

Convolutional neural networks can only predict the risk value up to a single time point, and cannot perform long-term and continuous risk prediction for the risk of complications, and it is difficult to provide accurate and effective decision support for clinicians

In addition, traditional recurrent neural networks and convolutional neural networks can only analyze simple data structures, but cannot handle truncated data that is common in clinical analysis

Images