Banca de QUALIFICAÇÃO: Helena Santos Brandão

Uma banca de QUALIFICAÇÃO de MESTRADO foi cadastrada pelo programa.
STUDENT : Helena Santos Brandão
DATE: 21/07/2023
TIME: 14:00
LOCAL: https://teams.microsoft.com/l/meetup-join/19%3a2vplckHZkagBUOI6ZERuouq6QhtZ7c7jZyjpwRYiCBs1%40thread
TITLE:

Applying differential entropy to financial data analysis

KEY WORDS:

Differential entropy, financial risk, non-parametric entropy estimators, kernel estimators

PAGES: 32
BIG AREA: Ciências Exatas e da Terra
AREA: Probabilidade e Estatística
SUMMARY:

In the realm of financial risk, the conventional approach has typically linked risk to the variance of a variable, such as the return of a stock or portfolio. Alternative measures that aim to tackle downside risk or extreme outcomes have emerged to address the limitations of this approach. Given the need for a comprehensive risk management approach, incorporating a variety of risk metrics is crucial. Therefore, new methods that provide additional insights beyond mere variability are highly valued. One such complementary measure is the uncertainty measure, which enables us to capture and describe different aspects of risk, going beyond traditional notions of variability alone. In particular, our study focuses on providing a review of plug-in type estimators for entropy. Initially, we made empirical experiments to observe the properties of some well-known non-parametric entropy estimators (based on nearest neighbor distances, sample-spacings, and Kernel Estimators). After, we applied these estimators to real financial data, which allowed us to glean valuable insights from the entropy estimation process. It was also pointed out that estimators based on Sample-Spacing and Nearest Neighbor methods have limitations when the data contains repeated values (often caused by rounding limitations). There is also seems to be a relation between the population distribution and the shape of the kernel function employed. These findings highlight the need for adaptations in the Kernel estimator method, encouraging further refinement. Furthermore, we intend to expand our applications to encompass a broader range of financial data and explore diverse scenarios.

COMMITTEE MEMBERS:
Interno - 1040312 - ANTONIO EDUARDO GOMES
Interno - 2531979 - JAMES MATOS SAMPAIO
Presidente - 1171224 - RAUL YUKIHIRO MATSUSHITA
Interno - 2357165 - ROBERTO VILA GABRIEL