Convolutional neural network-based health record integration method and device, and medium

The invention discloses a health record integration method and device based on a convolutional neural network, and a medium, and the method comprises the steps: collecting the medical record and medical record of an individual, and generating a health data set; preprocessing the health data set, performing feature extraction on the health data set, and generating information expression suitable for model learning; a multi-channel convolutional neural network model is constructed, channels are used for processing information expressions of a single type, channel fusion is carried out on data features of different types based on the model, feature extraction is carried out on all feature dimensions after fusion so as to realize duplicate removal and classification, and further integration is carried out to form an individual health file. For the time sequence change characteristics of the health record information data and the historical medical record information, the missing health record information data can