随机森林
计算机科学
健康
机器学习
人工智能
支持向量机
电子健康
自我管理
稳健性(进化)
哮喘
模式(遗传算法)
贝叶斯网络
医疗保健
医学
护理部
化学
内科学
经济
基因
生物化学
经济增长
心理干预
作者
Otilia Kocsis,Aris S. Lalos,Gerasimos Arvanitis,Κωνσταντίνος Μουστάκας
标识
DOI:10.1016/j.entcs.2019.04.007
摘要
Ambient intelligence and machine learning techniques are widely proposed by various eHealth and mHealth applications for home-care and self-management of various chronic health conditions. Their adoption for self-management of asthma, a multifactorial chronic disease, requires evaluation and validation in a real-life setups along with optimization at patient level to personalize predictions with respect to asthma control status and exacerbation risk. The current work proposes a novel short-term prediction approach for asthma control status, considering training of multiple classification models for each monitored parameters along with necessary pre-processing methods to enhance robustness and efficiency. The machine learning algorithms considered in this study are the Support Vector Machines, the Random Forests, AdaBoost and Bayesian Network. The Random Forests and Support Vector Machines classifiers demonstrated overall superior performance for the case studies (models) considered.
科研通智能强力驱动
Strongly Powered by AbleSci AI