A Remote Health Diagnosis Method Based on Full Voting XGBoost Algorithm

At present, classification models are widely used in medical disease monitoring. In remote health diagnosis technology, classification algorithm based on machine learning also plays a key role. Among them, XGBoost model is a widely used classification model. However, due to the imbalance of disease detection data, it will lead to a series of problems such as insufficient model learning and poor classification effect. Researchers have proposed some data enhancement methods to solve this kind of problem, but they often do not perform well in polyphenol tasks. In order to solve this problem, this paper proposes a full vote algorithm based on XGBoost model. For small sample data, a special dataset is generated and multiple models are trained. When classifying, all models vote in full to determine the classification result. Finally, we verify that our method performs better in the classification task of small sample disease detection data in simulation experiments.