Banca de QUALIFICAÇÃO: Cláudia Raquel da Rocha Eirado

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
STUDENT : Cláudia Raquel da Rocha Eirado
DATE: 29/11/2023
TIME: 16:30
LOCAL: a definir
TITLE:

Essays on economics using networked data sets and machine learning

KEY WORDS:

#centrality, #complexnetworks, #indirectagion, #machinelearning, #artificial intelligence, #systemic risk

PAGES: 60
BIG AREA: Ciências Sociais Aplicadas
AREA: Economia
SUMMARY:

This work comprises three papers in Economics utilizing networked datasets and machine learning. In the first paper, we conduct a review of machine learning applications for solving complex network problems. We cover concepts in machine learning, including supervised learning, unsupervised learning, and reinforcement learning, along with methods such as clustering, embedding, and PCA. Additionally, we explore network construction and centrality concepts, encompassing node and link prediction. The paper also discusses natural language approaches, incorporating theories from Natural Language Processing (NLP) , subarea of Artificial Intelligence - AI. The second paper delves into the concept of systemic risk in the financial domain, investigating its potential to trigger simultaneous losses across a specific sector, affecting multiple institutions and, ultimately, the entire system. The study emphasizes the role of human emotions and sentiments, beyond traditional financial data, in influencing market dynamics and the rationality of economic agents. The analysis extends to indirect contagion, examining how localized events can lead to severe instability or collapse across industries and economies. Drawing from the aftermath of the 2008 global financial crisis, the research investigates interconnectedness among institutions and channels of contagion that contribute to market instability. The exploration includes measures of connectedness in finance and insurance industries, associating them with systemic risk using text data on news. The study also highlights the growing importance of textual digital data in economic research, especially within the context of media influence on investor behavior. Techniques such as web scraping and artificial intelligence (AI) and machine learning (ML) are employed to analyze sentiments expressed in news articles and their impact on stock market movements. Natural Language Processing (NLP) is discussed as a crucial tool in understanding and interpreting textual data. A key aspect of the paper involves the application of a model utilizing a news similarity network to predict stationary probabilities as a proxy for centrality in the network and relations between firms, establishing a relationship between them and identifying paths of indirect contagion. By examining interactions and contagion propagation between firms based on news articles, the study aims to uncover insights into the interconnectedness and ripple effects within the financial system. The paper concludes with a discussion on the potential applications of AI and ML in understanding and predicting systemic risk in the financial landscape. The third paper is an empirical exercise on Oriented Modeling of Complex Networked Systems and Their Dynamics.

COMMITTEE MEMBERS:
Interno - 1196877 - BERNARDO PINHEIRO MACHADO MUELLER
Presidente - 1642911 - DANIEL OLIVEIRA CAJUEIRO
Externo ao Programa - 1997721 - HERBERT KIMURA - null