Noninvasive triglyceride estimation system based on PPG signal

The invention discloses a noninvasive triglyceride estimation system based on PPG signals, the system comprising: a data acquisition module for acquiring PPG signals, sample data comprising patient fingertip PPG signals acquired by a professional sensor, and recording triglyceride indexes corresponding to patients; a data preprocessing module used for carrying out single-cycle segmentation on the collected fingertip PPG signals and removing invalid samples measured due to interference of movement, light and the like; a neural network building module used for building a convolutional neural network for training and prediction; and a model training module used for substituting a part of the sample data into the convolutional neural network for training, updating network weight parameters, obtaining a trained neural network model, and taking the other part of reserved samples as a test for prediction analysis. According to the method, the average absolute error of prediction of samples is small, accurate measure