Nutritional Component Analysis and Hypertension Prediction Based on XGBoost Model

2023, Issue 2

Nutritional Component Analysis and Hypertension Prediction Based on XGBoost Model

Abstract: Hypertension is a common chronic disease, and early detection and intervention can reduce the risk of complications. Although the onset and development of hypertension are related to various factors, diet is recognized as one of the main influences on hypertension. Machine learning models can effectively predict diseases and provide auxiliary treatments. The author proposes a hypertension prediction scheme based on XGBoost by analyzing nutritional components. This scheme consists of five parts: data transformation, feature selection, data cleaning and normalization, model building, classification, and evaluation. Experimental results show that XGBoost achieved an F1 score of 0.859 and an accuracy rate exceeding 85% in hypertension prediction, outperforming Random Forest, Support Vector Machine, and Artificial Neural Network. Additionally, by analyzing the influence of different nutritional components on hypertension prediction, the top 10 nutritional features affecting hypertension were identified, most of which align with medical conclusions, validating the model’s effectiveness.

Long press the QR code to follow