Diabetes insipidus prediction method based on incremental neural network model and prediction system

Abstract

The invention discloses a diabetes insipidus prediction method based on an incremental neural network model. The diabetes insipidus prediction method comprises the following steps that a diabetes insipidus daily data database is established; a neural network model is trained; daily life data is acquired and sent to a server and is saved in a daily data record sheet of a user; current day data is extracted from the daily data record sheet of the user to form n-dimensional vectors, normalization processing is performed, then the vectors are input to a diabetes insipidus pathology neural network model for diabetes insipidus probability prediction; an intelligent household diabetes insipidus nursing device judges whether a diabetes insipidus probability value is greater than 0.5 or not; the users judges that the user suffers from diabetes insipidus and goes to a hospital for examination by himself/herself, an examination result is transmitted back to the server through the intelligent household diabetes insipidus nursing device, and the server judges whether the examination result is correct or not; when the examination result is wrong, an incremental algorithm is executed, and the neural network model is dynamically corrected. The diabetes insipidus prediction method is accurate in prediction, and the neural network model is customized for each user.

Description

technical field [0001] The invention belongs to the field of medical technology, in particular to a prediction method and prediction system for diabetes insipidus based on an incremental neural network model. Background technique [0002] At present, each domestic health management system has set up diabetes insipidus prediction evaluation, and the prediction method used is data matching. The principle is to input personal life data into the system, and the system matches the fixed data and then obtains the probability of disease. However, due to the complexity and unpredictability of the human body and disease, in terms of the manifestations of biological signals and information, and the law of change (self-change and changes after medical intervention), the detection and signal expression of biological signals and information, the data and information obtained There are very complex nonlinear relationships in analysis, decision-making and many other aspects. Therefore, t...

AI Technical Summary

Problems solved by technology

However, due to the complexity and unpredictability of the human body and diseases, the detection and signal expression of biological signals and information in the form of expression and change law (self-change and change after medical intervention), the acquired data and information There are very complex nonlinear relationships in analysis, decision-making and many other aspects

Therefore, the use of traditional data matching can only be blind data screening, unable to judge the logical relationship between data and variables, and the obtained value range deviation is large, resulting in very poor specificity of system prediction, so the current domestic health management The system cannot effectively predict an individual's diabetes insipidus

[0003] Previously, most of the predictions of diabetes insipidus used the BP neural network model, but when new detection data is generated, the neural network model must be trained again, and the calculation efficiency is extremely low

And when the scale of system users increases, the server will not be able to complete the training tasks in time

Images