Machine Learning (ML) affords researchers tools to advance beyond research methods commonly employed in psychology, business, and public policy studies of federal nutrition programs and participant food decision-making. It is a sub domain of AI that is applied for feature extraction – a crucial step in decision making. These features are used in context-specific automated decisions resulting in predictive AI models. Whereas many prior studies rely on retrospective, static, “one-shot” decision-making in controlled laboratory environments, ML allows researchers to refine predictions about participation and food behaviors using large-scale datasets. We propose a case study using ML to predict an aspect of participation in a large, publicly funded nutrition education program (The Expanded Food and Nutrition Education Program). Participation has important downstream implications for diet quality, food security, and other important nutrition related decisions. We then suggest a process for validating the ML insights using qualitative research and survey data.

中文翻译：

---

用于了解政府营养计划参与情况的机器学习 (ML) 方法


机器学习 (ML) 为研究人员提供了超越心理学、商业和公共政策研究中常用的联邦营养计划和参与者食品决策研究方法的工具。它是人工智能的一个子领域，用于特征提取——决策中的关键步骤。这些功能用于特定环境的自动化决策，从而产生预测性人工智能模型。尽管许多先前的研究依赖于受控实验室环境中的回顾性、静态、“一次性”决策，但机器学习允许研究人员使用大规模数据集来完善对参与和食物行为的预测。我们提出了一个案例研究，使用机器学习来预测参与大型公共资助营养教育计划（扩展食品和营养教育计划）的某个方面。参与对饮食质量、粮食安全和其他重要的营养相关决策具有重要的下游影响。然后，我们建议使用定性研究和调查数据来验证机器学习见解的流程。