DiD: MIning Relationships Among variables in large datasets from CompLEx systems (MIRACLE)

社会科学家面临的一些最紧迫的问题--从更好地理解人类与环境之间的相互作用，到预测人口老龄化将如何影响美国和全球经济--需要理解大量数据。因此，社会科学家越来越多地使用新的计算建模方法来探索人类互动的动态和后果。这些新方法，包括基于代理的模型，提供了探索研究问题的方法，不能使用传统的统计方法进行调查。 但是，仍然缺乏合适的方法来挖掘、分析和综合大规模复杂模型输出数据，以回答社会科学研究问题。传统的分析方法是针对线性、连续和正态分布的数据而设计的，而来自复杂社会生态系统模型的数据是非线性、不连续和幂律分布的。在这个项目中，研究人员试图通过开发、应用和传播一个集成环境来解决这些挑战，该环境用于分析和可视化复杂系统模型生成的数据。一个重要的更广泛的影响是，该研究将导致工具，将允许利益相关者，政策制定者和公众探索，互动，并提供反馈，否则难以理解的模型。该项目建立在项目团队正在进行的研究，并使用NSF支持的CoMSES网络计算建模库作为平台，使这套协作，开源工具广泛使用。这个社区环境将允许任何用户发布模型输出沿着相关元数据，可视化和分析输出数据，评论和分享分析，并利用其他项目的数据进行比较和元分析。这种网络基础设施将提供发现关系的半自动化手段，这些关系可以导致关于社会系统如何工作的新理论，测试模拟的真实性，并提出新的研究方向。

Some of the most pressing questions for social scientists -- ranging from efforts to better understand the interactions between humans and the environment to predicting how an aging population will impact the US and global economy -- require making sense of large amounts of data. As a result, social scientists are increasingly using new computational modeling methods to explore the dynamics and consequences of human interactions. These new methods, including agent-based models, provide ways to explore research questions that cannot be investigated using traditional statistical approaches. But appropriate methods to mine, analyze, and synthesis large-scale complex model output data in order to answer social science research questions are still lacking. Traditional analysis methods are designed for data that are linear, continuous, and normally distributed, while data from models of complex socio-ecological systems are non-linear, discontinuous, and power-law distributed. In this project, the researchers seek to address these challenges by developing, applying, and disseminating an integrated environment for analysis and visualization of data generated by complex systems models. An important broader impact is that the research will lead to tools that will allow stakeholders, policy makers, and the general public to explore, interact with, and provide feedback on otherwise difficult-to-understand models.The project builds on ongoing research by the project team, and uses the NSF-supported CoMSES Net Computational Modeling Library as a platform to make this suite of collaborative, open-source tools broadly available. This community environment will allow any users to post model output along with associated metadata, visualize and analyze output data, comment on and share analyses, and conduct comparative and meta-analysis, drawing on data from other projects. This cyber-infrastructure will provide semi-automated means of discovering relationships that can lead to new theories about how social systems work, test the realism of simulations against knowledge from empirical systems, and propose new research directions to explore.