Prediction device for cardiovascular adverse events of percutaneous coronary intervention based on machine learning

Abstract

The invention discloses a prediction device for cardiovascular adverse events of percutaneous coronary intervention based on machine learning. The prediction device is characterized in that a cardiovascular event prediction model is saved in a memory, the cardiovascular event prediction model comprises a trained XGboost model, a LightGBM model, an SVM model, an NN model and corresponding weights of each model. The operating process of the prediction device comprises the steps that to-be-detected clinical characteristic data is received, missing value filling is conducted on the to-be-detectedclinical characteristic data; correlation detection is conducted on the clinical characteristic data on which missing value filling is conducted, and the clinical characteristic value with a crossed relationship is removed; computation is conducted on the clinical characteristic data on which correlation detection is conducted using the trained XGboost model, the LightGBM model, the SVM model andthe NN model to obtain four prediction probabilities, and weighted summation is conducted on the four prediction probabilities to obtain the prediction probability predicted by using the cardiovascular adverse event prediction model.

Description

technical field [0001] The invention belongs to the field of medical artificial intelligence, and in particular relates to a machine learning-based device for predicting cardiovascular adverse events in percutaneous coronary intervention. Background technique [0002] With the great development of machine learning in the field of statistics, machine learning-based methods are also widely used in medical data. [0003] In the field of coronary heart disease, percutaneous coronary intervention (PCI) is currently one of the most effective treatments. In addition to the therapeutic effect of opening occluded blood vessels and improving cardiac function, PCI is often accompanied by the possibility of long-term complications such as in-stent restenosis and recurrent myocardial infarction. Prediction of major adverse cardiovascular events (MACE) after PCI has also become an important direction of clinical research. Among them, MACE mainly includes myocardial infarction, stent thr...

AI Technical Summary

Problems solved by technology

[0007] In the course of decades of development, various learning algorithms including linear models and tree models have appeared in machine learning. Different learning algorithms have different adaptability to different predictions. However, there are not many clinical studies on cardiovascular adverse events. Learning algorithm application experience

Images