Analyzing vaccination priority judgments for 132 occupations using word vector models.
Analyzing vaccination priority judgments for 132 occupations using word vector models.
复制标题
使用词向量模型分析 132 种职业的疫苗接种优先级判断。
DOI：
10.1145/3498851.3498933
复制
发表时间：
2021
期刊：
影响因子：
--
通讯作者：
H.
中科院分区：
文献类型：
--
作者：
Ueshima;A.;& Takikawa;H.
大多数人类社会都根据职业进行高度的劳动分工。然而，确定应该分配疫苗等稀缺资源的职业领域是一个有争议的话题，特别是考虑到COVID-19的情况。虽然理解和预测人们对资源分配优先级的判断是至关重要的，但对占用概念进行量化是一项艰巨的任务。在这项研究中，我们通过将人们对职业的知识表示量化为向量空间中的单词向量，研究了人们对不同职业的疫苗接种优先级的判断可以在多大程度上建模。结果表明，将职业量化为词向量的模型具有较高的样本外预测精度，使我们能够探索参与者判断背后的心理维度。这些结果表明，使用词向量来模拟人类对日常概念的判断，可以预测性能和理解判断机制。
Most human societies conduct a high degree of division of labor based on occupation. However, determining the occupational field that should be allocated a scarce resource such as vaccine is a topic of debate, especially considering the COVID-19 situation. Though it is crucial that we understand and anticipate people's judgments on resource allocation prioritization, quantifying the concept of occupation is a difficult task. In this study, we investigated how well people's judgments on vaccination prioritization for different occupations could be modeled by quantifying their knowledge representation of occupations as word vectors in a vector space. The results showed that the model that quantified occupations as word vectors indicated high out-of-sample prediction accuracy, enabling us to explore the psychological dimension underlying the participants’ judgments. These results indicated that using word vectors for modeling human judgments about everyday concepts allowed prediction of performance and understanding of judgment mechanisms.