APSO_LightGBM模型在高血压风险预测中的应用

APSO_LightGBM Model and Its Application in Hypertension Risk Prediction

APSO_LightGBM高血压风险预测模型将LightGBM算法与自适应粒子群优化算法相结合,弥补了LightGBM自身收敛速度慢的缺陷。实验表明,基于该模型的预测精度高于线性回归、决策树、SVM和LightGBM模型,参数寻优时间也远小于网格搜索与随机搜索,说明该模型可以更准确、高效地预测高血压风险,实现对疾病的早期筛查与干预。

A new APSO_LightGBM model is proposed to predict hypertension risk.This model combines the LightGBM algorithm with the adaptive particle swarm optimization algorithm to make up for the slow convergence of LightGBM itself.Experiments show that the accuracy of APSO_LightGBM model is higher than linear regression,decision tree,SVM and LightGBM,and the optimization time is much shorter than grid search and random search,which fully shows that the model can be more accurate and efficient to predict hypertension,and making early screening and intervention for diseases a reality.