Banca de DEFESA: JOSÉ EDVALDO DE OLIVEIRA NUNES

Uma banca de DEFESA de DOUTORADO foi cadastrada pelo programa.
STUDENT : JOSÉ EDVALDO DE OLIVEIRA NUNES
DATE: 18/02/2025
TIME: 09:00
LOCAL: https://meet.google.com/rzj-nkuw-yaf
TITLE:

Analysis of complex price networks of Brazilian agricultural commodities using the horizontal visibility graph method

KEY WORDS:

agricultural market; complex networks; time series; visibility graphs.

PAGES: 64
BIG AREA: Ciências Agrárias
AREA: Agronomia
SUMMARY:

Agricultural commodities are directly associated with agribusiness and play a fundamental role in the Brazilian economy. These products can be considered the most important, as their prices have a significant effect on the population's quality of life. Understanding the behavior of agricultural commodity price series is a challenge, due to the complexity of the market and the abrupt changes resulting from global crises. The methods currently applied to analyze nonlinear time series still present a high computational cost, requiring simpler and faster methods to extract information from the process that generated them, with the aim of understanding, modeling and forecasting. Motivated by this, we propose to apply the Horizontal Visibility Graph (HVG) method to analyze the temporal variation of the prices of Brazilian agricultural commodities. The data were provided by the Center for Advanced Studies in Applied Economics (CEPEA) of the Luiz de Queiroz College of Agriculture (ESALQ), University of São Paulo (USP) and divided into two groups: (i) ethanol and sugar, in relation to the COVID-19 pandemic, with records between January 2017 and May 2023; and (ii) sugar, cotton, cattle, soybeans, and coffee, in relation to the 2007/2008 financial crisis and the COVID-19 pandemic, covering the periods from May 2003 to September 2011 and from January 2017 to May 2023. The HVG, based on the theory of complex networks, maps a time series onto a graph, assigning each data point in the series to a node. The software used was VisGraAnalysis, developed in C programming language. The main idea is to study to what extent the techniques and focus of graph theory are useful as a way to characterize time series. The following network indices were calculated: clustering coefficient, transitivity, average shortest path length, degree distribution, and Kullback-Leibler divergence to identify the irreversibility of the series. The main results indicated that the HVG positively captures the structural properties of the networks, and price variations reflect changes in market dynamics in the face of global crises. The relationship between ethanol and sugar was analyzed, identifying the sugar market (in BRL) as more efficient and the ethanol market as less efficient during the COVID-19 pandemic. For the second group, lower efficiency was observed in the period before the 2007/2008 crisis, except for soybeans. Coffee showed greater sensitivity to the 2007/2008 crisis, while sugar and cattle showed increased efficiency during the pandemic (the value of λ decreased, indicating a less correlated series).

COMMITTEE MEMBERS:
Interno - ANTONIO SAMUEL ALVES DA SILVA
Presidente - BORKO STOSIC
Externo à Instituição - JOSÉ DOMINGOS ALBUQUERQUE AGUIAR - IFPE
Externa à Instituição - LIDIANE DA SILVA ARAUJO - Unipampa
Interno - LUCIAN BOGDAN BEJAN