Chronic disease condition change event prediction device based on a recurrent neural network

Abstract

The invention discloses a chronic disease condition change event prediction device based on a recurrent neural network, and the device comprises a memory, a processor, and a computer program, a preprocessing module and a chronic disease condition change event prediction model are stored in the memory, and the prediction model comprises a preprocessing module, a condition feature extraction module,and a classification module. When the processor executes a computer program, the following steps are realized: receiving long-term longitudinal data generated by multiple hospitalization of a patient, performing data preprocessing on the number by the preprocessing module, and reconstructing the data of each hospitalization into a feature vector as a to-be-tested data set; Taking the to-be-detected data set as input, extracting disease characteristics by a disease characteristic extraction module, and inputting the disease characteristics into a classification module; And enabling the classification module to output the prediction probability of various events indicating that the illness state changes. The prediction device can predict the event that the chronic disease patient has markeddisease condition change in the target time window, thereby assisting the doctor to formulate reasonable diagnosis and treatment measures and reducing the medical expenditure.

Description

technical field [0001] The invention belongs to the field of data processing, and in particular relates to a chronic disease condition change event prediction device based on a cyclic neural network. Background technique [0002] Chronic diseases are the main cause of death of Chinese residents. According to statistics, in recent years, about 86% of the dead population in my country died of various chronic diseases. The main characteristics of chronic diseases are that they last for a long time, the disease is difficult to reverse, and the causes of disease are relatively complicated. Predicting the possible future condition change events of chronic diseases is very important for evaluating the prognosis of chronic diseases, and it also helps to carry out more precise treatment of chronic diseases. [0003] Most of the existing prediction methods for major events of chronic diseases are the results of clinical medical research, and the technical means used are basically sim...

AI Technical Summary

Problems solved by technology

However, although the original design of the recurrent neural network is good, it is not suitable for receiving long-term patient data, because it is generally used to receive sequence data with the same time interval, an

Images