Banca de QUALIFICAÇÃO: FELIPE ROCHA DE ARAUJO

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
DISCENTE: FELIPE ROCHA DE ARAUJO
DATA: 02/02/2023
HORA: 14:30
LOCAL: GERCOM
TÍTULO:

A Machine Learning Approach for Transportation Mode Classification using Information Theory Quantifiers

PALAVRAS-CHAVES:

Information Theory, human mobility e ML

PÁGINAS: 50
GRANDE ÁREA: Ciências Exatas e da Terra
ÁREA: Ciência da Computação
SUBÁREA: Sistemas de Computação
ESPECIALIDADE: Teleinformática
RESUMO:

The disorderly growth caused by migration and the increasing world population demand solutions to improve the infrastructure of cities. The academic and industrial communities are investing in solutions based on human mobility that can provide more sustainable development and reduce several problems, such as the economy, transport, traffic management, environment, and many other factors, to improve the life quality of humans. In this context, location-aware services provide valuable information for capturing human mobility patterns. Exploring the dynamics of mobility time series, such as the speed variation of the transportation modes, leads to better solutions by understanding the underlying data-generating process and identifying different patterns. In light of this, strategies based on Information Theory measures associated with ordinal pattern methods, for example, Complex-Entropy Causality Plane and Fisher-Shannon Causality Plane, have reached relevant advancements in distinguishing different time series dynamics. Consequently, they are promising tools to explain those complex behaviors to improve human mobility-based services. This thesis proposal presents a machine-learning approach to classify transportation modes through the Information Theory quantifiers. These quantifiers are computed over speed time series derived from the Geolife data set. Then, we characterize each transportation series by observing its dynamics over time and correlating their associated Information Theory quantifiers with colored noises mapped onto the causal planes. Evaluation results show the potential of our study, allowing us to identify motorized and non-motorized means of transportation regimes, estimate the transportation switching based on causal plane mappings, and efficiently classify the means of transportation through machine learning models.

MEMBROS DA BANCA:
Presidente - 1700540 - EDUARDO COELHO CERQUEIRA
Interno - 2244893 - DENIS LIMA DO ROSÁRIO
Externo à Instituição - ALLAN DOUGLAS BENTO DA COSTA
Externo à Instituição - IAGO LINS DE MEDEIROS
Externo à Instituição - THAIS LIRA TAVARES DOS SANTOS