基于RF-RFECV和Stacking集成学习的脑卒中预测研究

Research on stroke prediction based on RF-RFECV and Stacking integrated learning

脑卒中具有发病率高、死亡率高和致残率高的特点,提早发现和治疗显得至关重要。在脑卒中预测方法中,机器学习相对于其他方法具有更好的表现。针对传统的单一机器学习模型在预测的精度或稳定性上都存在局限性的问题,提出了一种基于RF-RFECV和Stacking集成学习的脑卒中预测方法。通过实验证明,该方法可以有效地降低特征维度,获得最优特征子集,与其他的单一模型以及其他集成算法模型相比,Stacking模型的预测精度明显提升,可以更有效地预测脑卒中。

Stroke has the characteristics of high incidence rate,high mortality and high disability rate.Early detection and treatment are essential.Among stroke prediction methods,machine learning has better performance than other methods.Aiming at the limitation of traditional single machine learning model in prediction accuracy or stability,stroke prediction method based on RFRFECV and Stacking ensemble learning is proposed.Experiments show that this method can effectively reduce feature dimensions and obtain the optimal feature subset.Compared with other single models and other integrated algorithm models,the prediction accuracy of Stacking model is significantly improved.The research can more effectively predict stroke.