基于XGBoost模型的营养成分分析高血压预测方案

高血压是一种常见的慢性病，若能早发现、早采取措施可降低其引发并发症的风险。尽管高血压的产生与发展与诸多因素有关，但饮食被公认为影响高血压的主要因素之一。机器学习模型可以对疾病进行有效预测，并提供辅助治疗。笔者提出一种基于XGBoost的通过分析营养成分预测高血压的方案，该方案由数据转换、特征选择、数据清理与标准化、模型搭建、分类与评估5部分组成。实验结果表明，XGBoost在高血压预测中获得了0.859的F1分数且准确率超过85%,高于随机森林、支持向量机与人工神经网络。此外通过分析不同营养成分对高血压预测的影响因素，获得了影响高血压的前10个营养特征，大部分与医学结论相同，验证了模型的有效性。

Hypertension predicting scheme by analyzing nutritional ingredients based on xgboost model

Hypertension is a common chronic disease. Early detection and early measures can reduce the risk of complications. Although the occurrence and development of hypertension are related to many factors, diet is recognized as one of the main factors affecting hypertension. Machine learning models can effectively predict the disease and provide adjuvant therapy. Accordingly, this paper proposes a scheme based on XGBoost to predict hypertension by analyzing nutritional ingredients. The scheme consists of five parts:data conversion, feature selection, data cleaning and standardization, model building, classification and evaluation. The experimental results show that XGBoost obtains an F1_score of 0.859 in the prediction of and the accuracy rate exceeds 85%, which are higher than random forests, support vector machine and artificial neural network. By analyzing the influence factors of different nutritional ingredients on the prediction of hypertension, we obtain the top 10 nutritional characteristics that affect hypertension, most of which are the same as medical conclusions, verifying the effectiveness of the model.