A decision-making mechanism for assessing risk factor significance in cardiovascular diseases
A decision-making mechanism for assessing risk factor significance in cardiovascular diseases
复制标题
DOI：
10.1016/j.dss.2018.09.004
复制
发表时间：
2018-11-01
影响因子：
7.5
通讯作者：
Hsu, Wei-Yen
中科院分区：
文献类型：
--
作者：
Hsu, Wei-Yen
心血管疾病（CVD）是严重的疾病，其在全球范围内的发病率不断增加，刺激了国家医疗保健支出的增加。尽管有许多诊断和治疗建议，心血管疾病仍然值得研究，因为它们的风险因素多种多样，其中一些是正相关的，负相关的或不相关的。为了帮助医生和研究人员识别CVD危险因素的重要性，在这项研究中，我们提出了一种新的排名和属性（或特征）选择算法。我们应用了七种流行的机器学习技术来生成属性排名数据集，以便为每个分类器确定理想的因子/属性数量。比较结果表明，经过排序和属性选择后，部分因子的性能明显优于整体因子和几种最先进的算法。由于这些知识可以帮助正确选择CVD患者的因素，从而帮助医生在诊断和治疗中做出更好的决策，因此我们的研究结果可以降低治疗成本，从而降低医疗保健的经济负担。
Cardiovascular diseases (CVDs) are severe diseases whose growing incidence worldwide has spurred increased national healthcare spending. Despite numerous diagnostic and treatment suggestions, CVDs continue to merit investigation due to their diverse risk factors, some of which are positively, negatively, or not correlated. To assist doctors and researchers in identifying the significance of CVD risk factors, in this study we propose a novel ranking and attribute (or feature) selection algorithm. We applied seven popular machine learning technologies to generate attribute-ranked datasets in order to identify the ideal number of factors/attributes for each classifier. Above all, the results of the comparisons indicate that the performance of parts of factors after ranking and attribute selection was significantly better than the performance of whole factors and that of several state-of-the-art algorithms. Since such knowledge can aid the proper selection of factors of CVD patients and thereby assist doctors in making better decisions in diagnostics and treatment, our results can reduce treatment costs and thus lower the economic burden of healthcare.