Acerca de
Experiencia y educación
Licencias y certificaciones
-
Generative AI with Large Language Models
DeepLearning.AI, Amazon Web Services
Publicaciones
-
A Inteligência Artificial impõe o maior desafio energético ao século XXI
Um só Planeta
Treinamento de um único modelo de linguagem pode consumir o equivalente ao consumo de centenas de residências em um ano
Otros autoresVer publicación -
Enabling data-driven results
Medium
Thoughts about integrating decision systems into business processes.
-
Dynamic Time Scan Forecasting for multi-step wind speed prediction
Renewable Energy
Multi-step forecasting of wind speed time series, especially for day-ahead and longer time horizons, is still a challenging problem in the wind energy sector. In this paper, a novel analog-based methodology to perform multi-step forecasting in univariate time series, named dynamic time scan forecasting (DTSF), is presented. DTSF is a fast time series forecasting methodology for large data sets. Thus, the proposed method is optimal for forecasting renewable energy features such as wind speed, in…
Multi-step forecasting of wind speed time series, especially for day-ahead and longer time horizons, is still a challenging problem in the wind energy sector. In this paper, a novel analog-based methodology to perform multi-step forecasting in univariate time series, named dynamic time scan forecasting (DTSF), is presented. DTSF is a fast time series forecasting methodology for large data sets. Thus, the proposed method is optimal for forecasting renewable energy features such as wind speed, in which standard statistical and soft computing methods present limitations. A scan procedure is applied to identify similar patterns, named best matches, throughout the time series. As opposed to euclidean distance, more flexible similarity functions, using polynomial regression models, are dynamically estimated and Goodness-of-fit statistics are used to find the best matches. The observed values following the best matches and the fitted similarity functions are used to predict k-steps ahead, as well as forecasting intervals. An ensemble version of the method, named eDTSF, combines different predictions using different set of parameters thus, further improving forecasting performance. Remarkably, eDTSF achieved competitive results for multi-step forecasting of wind speed time series, even in situations of very high variability, as compared to eleven selected concurrent forecasting methods.
Otros autoresVer publicación -
O futuro das pessoas e da gestão
HSM Management
Quatro pilares, em atuação interdependente, permitem antecipar os próximos anos das empresas: sustentabilidade, capital, trabalho e inovação.
Otros autoresVer publicación -
Artificial Intelligence: a concise conceptual introduction
Towards Data Science
A concise description of the main concepts around Artificial Intelligence and Machine Learning.
-
Bayesian detection of clusters in efficiency score maps: An application to Brazilian energy regulation
Applied Mathematical Modelling
An original application of an approach first used in epidemiology investigation was developed and implemented in energy regulation benchmarking. Using Brazilian electricity energy distribution utilities, the proposed methodology applies spatial Bayesian analysis to estimate the number of clusters and the utilities in each cluster. By dividing the utilities into smaller, but geographically closer groups, it can be argued that local determinants of production or environmental components are…
An original application of an approach first used in epidemiology investigation was developed and implemented in energy regulation benchmarking. Using Brazilian electricity energy distribution utilities, the proposed methodology applies spatial Bayesian analysis to estimate the number of clusters and the utilities in each cluster. By dividing the utilities into smaller, but geographically closer groups, it can be argued that local determinants of production or environmental components are ed in the benchmarking model. Thus, the proposed method requires the spatial location of the utilities and their cost efficiencies, estimated previously by the regulator. Results show two detected clusters with high and low efficiencies located in the east and west of Brazil. After applying the regulator model to the detected groups, significant changes in cost efficiencies were estimated for a few utilities. This is important information that can be used by the regulator to estimate future cost incentives.
Otros autoresVer publicación -
Implementação do método Bayesiano de regionalização para análise espacial de índices de eficiência das empresas brasileiras distribuidoras de energia elétrica
L Simpósio Brasileiro de Pesquisa Operacional
-
Modeling of a Torch Ignition System Using One-Dimensional Model of Computational Simulation
23rd SAE Brasil International Congress and Display
An torch ignition system with homogeneous charge is numerically analyzed using a one-dimensional computational model. The new ignition system is implemented in a four-cylinder engine, spark ignition, 1600 cm3, 16 valves. Parameters such as mass burn fraction profile and pressure vs crank angle are compared with experimental data obtained with the torch ignition system operating homogeneous charge with stoichiometric mixture. The computational model uses information such as the pre-chamber…
An torch ignition system with homogeneous charge is numerically analyzed using a one-dimensional computational model. The new ignition system is implemented in a four-cylinder engine, spark ignition, 1600 cm3, 16 valves. Parameters such as mass burn fraction profile and pressure vs crank angle are compared with experimental data obtained with the torch ignition system operating homogeneous charge with stoichiometric mixture. The computational model uses information such as the pre-chamber pressure as a function of crack angle, intake and exhaust pressure, volumetric efficiency, maps of injection and ignition, valve discharge and valve intake coefficient, lifting valve, laminar flame speed, among others parameters. A simulation model of the torch ignition system, after being validated with experimental data, should provide information on increasing the burning rate, possibility of burning lean mixtures, more complete burn of the mixture, less sensitivity to composition and state of the mixture among others points that could be improved in control and ignition system design by torch ignition. To simulate the GT-POWER program and the results will used to obtain important information of the air flow and combustion in the cylinder in order to improve the original design of the torch ignition system.
Proyectos
-
Awesome Healthmetrics
A curated list of awesome resources at the intersection of healthcare and AI.
-
Machine Learning Timeline
Important milestones on the journey towards learning machines.
-
Machine Learning Feynman Experience
A collection of concepts I tried to implement using only Python, NumPy and SciPy on Google Colaboratory.
-
Machine Learning Knowledge Graph
Organizing concepts related to machine learning and artificial intelligence.
-
gbdcd
An R package implementing the Bayesian Detection of Clusters and Discontinuities.
-
Machine Learning Journey
-
A curated list of free learning resources on topics related to Machine Learning.
Idiomas
-
Portuguese
Competencia bilingüe o nativa
-
English
Competencia bilingüe o nativa
-
Italian
Competencia básica limitada