BIGDATA: Collaborative Research: IA: F: Too Interconnected to Fail? Network Analytics on Complex Economic Data Streams for Monitoring Financial Stability

BIGDATA：协作研究：IA：F：互联性太强以至于不会失败？

• 批准号: 1632730
• 负责人: George Michailidis
• 金额: $ 47万
• 依托单位: University of Florida
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2016
• 资助国家: 美国
• 起止时间: 2016-09-01 至 2020-12-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1632730&HistoricalAwards=false
• 关键词: BIGDATA; Collaborative; Research; IA; Too

The recent financial crisis has accentuated the need for effective monitoring, oversight and regulation of financial markets and institutions. Complex market structures involving intricate interconnected relationships among financial institutions can help propagate and amplify shocks and hence also foster systemic risk. This project develops an integrative framework, based on accounting principles, that leverages a wide array of diverse quantitative financial datastreams, complemented by metadata and market announcements for the purpose of identifying and predicting market participants that could endanger the overall financial system.The proposed research builds upon modern statistics and computer science works, as well as recent financial and economic ideas aimed at assessing threats to financial stability and uncovering the complexity of financial systems in different market conditions. It will result in both new methods for complex Big Data and empirical results that can advance the state-of-the-art in financial research, as well as tools that support and enhance financial policymaking and decision-making. Key tasks of the project include: (1) Develop a rigorous accounting framework to integrate multiple financial and econometric data streams from many platforms and technologies. (2) Develop and customize a range of new network models and analysis tools for use with multiple financial data streams. An important idea will be to extend network and econometric tools in order to compare the structural evolution of different types of networks in response to external events and policy changes.

最近的金融危机突出表明，需要对金融市场和机构进行有效的监测、监督和监管。复杂的市场结构涉及金融机构之间错综复杂的相互关联关系，可能有助于传播和放大冲击，从而也会助长系统性风险。该项目开发了一个基于会计原则的综合框架，该框架利用各种不同的量化金融数据流，辅之以元数据和市场公告，以识别和预测可能危及整个金融体系的市场参与者。拟议的研究建立在现代统计和计算机科学工作的基础上，以及最近的金融和经济思想，旨在评估对金融稳定的威胁，并揭示金融体系在不同市场条件下的复杂性。它将为复杂的大数据和实证结果带来新的方法，这些方法可以推动金融研究的最新发展，以及支持和加强金融政策制定和决策的工具。该项目的主要任务包括：（1）开发一个严格的会计框架，以整合来自许多平台和技术的多个金融和计量经济数据流。(2)开发和定制一系列新的网络模型和分析工具，用于多个金融数据流。一个重要的想法是扩大网络和计量经济学工具，以便比较不同类型的网络在应对外部事件和政策变化时的结构演变。

项目成果

Estimation of graphical models through structured norm minimization

通过结构化范数最小化估计图模型

• 发表时间: 2018
• 期刊: Journal of machine learning research
• 影响因子: 6
• 作者: Tarzanagh, Davoud Attae;Michailidis, George
• 通讯作者: Michailidis, George

Fast Randomized Algorithms for t-Product Based Tensor Operations and Decompositions with Applications to Imaging Data

基于 t 产品的张量运算和分解的快速随机算法及其在成像数据中的应用

• DOI: 10.1137/17m1159932
• 发表时间: 2018
• 期刊: SIAM Journal on Imaging Sciences
• 影响因子: 2.1
• 作者: Tarzanagh, Davoud Ataee;Michailidis, George
• 通讯作者: Michailidis, George

Finite Time Identification in Unstable Linear Systems

• DOI: 10.1016/j.automatica.2018.07.008
• 发表时间: 2017-10
• 期刊: ArXiv
• 作者: Mohamad Kazem Shirani Faradonbeh;Ambuj Tewari;G. Michailidis

Regularized Estimation of High-dimensional Factor-Augmented Vector Autoregressive (FAVAR) Models

高维因子增强向量自回归 (FAVAR) 模型的正则化估计

• 发表时间: 2020
• 期刊: Journal of machine learning research
• 影响因子: 6
• 作者: Jiahe Lin, George Michailidis

Sequential change-point detection in high-dimensional Gaussian graphical models

• 发表时间: 2018-06
• 期刊: ArXiv
• 作者: Hossein Keshavarz;G. Michailidis;Y. Atchadé