论文标题
检测患者对机器学习的帮助
Detecting the patient's need for help with machine learning
论文作者
论文摘要
开发支持健康分析的机器学习模型需要对自我评估表达式陈述的统计特性的更多了解。我们分析了有关冠状病毒Covid-19的流行病的自我评估表达式陈述，以识别受访者组之间的统计学上显着差异，并检测患者对机器学习的帮助。我们的量化研究收集了有关二十个与健康相关的表达陈述的“需要”评级，该陈述在11点李克特量表上涉及冠状病毒流行，以及关于该人的健康和福祉，性别和年龄的九个答案。在5月30日至2020年8月3日之间的在线受访者是从芬兰患者和残疾人组织，其他与健康有关的组织和专业人士以及教育机构（n = 673）招募的。我们分析了肯德尔等级相关的评级差异和依赖关系，以及余弦的相似性度量和Wilcoxon Rank-sum，Kruskal-Wallis和组之间的单向方差分析（ANOVA）的测试，并通过基本的卷积神经网络algorithm进行了机器学习实验。我们发现，各种与健康相关的表达式对之间的统计学意义相关性和高余弦相似性值，与“需要帮助”评级和背景问题对。我们还根据背景问题的答案值（例如，怀疑具有冠状病毒感染的评分，以及取决于估计的健康状况，生活质量和性别的质量，对分组的几种与健康相关的表达陈述的评级差异。我们使用卷积神经网络算法进行的实验显示了机器学习的适用性，以支持检测患者表达的帮助。
Developing machine learning models to support health analytics requires increased understanding about statistical properties of self-rated expression statements. We analyzed self-rated expression statements concerning the coronavirus COVID-19 epidemic to identify statistically significant differences between groups of respondents and to detect the patient's need for help with machine learning. Our quantitative study gathered the "need for help" ratings for twenty health-related expression statements concerning the coronavirus epidemic on a 11-point Likert scale, and nine answers about the person's health and wellbeing, sex and age. Online respondents between 30 May and 3 August 2020 were recruited from Finnish patient and disabled people's organizations, other health-related organizations and professionals, and educational institutions (n=673). We analyzed rating differences and dependencies with Kendall rank-correlation and cosine similarity measures and tests of Wilcoxon rank-sum, Kruskal-Wallis and one-way analysis of variance (ANOVA) between groups, and carried out machine learning experiments with a basic implementation of a convolutional neural network algorithm. We found statistically significant correlations and high cosine similarity values between various health-related expression statement pairs concerning the "need for help" ratings and a background question pair. We also identified statistically significant rating differences for several health-related expression statements in respect to groupings based on the answer values of background questions, such as the ratings of suspecting to have the coronavirus infection and having it depending on the estimated health condition, quality of life and sex. Our experiments with a convolutional neural network algorithm showed the applicability of machine learning to support detecting the need for help in the patient's expressions.
