High Performance Intelligent Decision Systems


About our research group

Our research group works on design and implementation of decision support systems based on intelligent algorithms to work on high performance computer architectures. The goal of such algorithms is to solve semi-structured data mining and optimization problems. The researchers involved in this project, which are experts in complementary fields, intend to work together to combine their research to advance and contribute to the field of decision support systems. Such systems are extremely important to organizations as they provide valuable information that helps managers take strategic decisions.

Decision support systems must be able to deal with semi-structured problems and to quickly extract reliable information from large sources of data. To solve semi-structured problems, non-traditional algorithms must be employed, such as clustering, classification and optimization algorithms based on bio-inspired computing.

Due to the large set of parameters, queries and data associated with semi-structured problems, algorithms applied to such problems often require large computational power to obtain results. Therefore, the application of high performance computing techniques may lead to faster responses in such situations, so that these algorithms can be successfully applied to aid decision-making.

Our group considers innovation as the basis of the whole research project, due not only to the integration of distinct areas of Computer Science, but also to the focus in the development of products based on innovative concepts.


Published Works

Scientific and Technological Production

Check our scientific and technological production (content in Portuguese)

Our Team


A. L. S. Gradvohl

Post-Doc in Distributed Systems

A. E. A. Silva

Ph.D. in Electrical Engineering

G. P. Coelho

Ph.D. in Electrical Engineering

Graduate Students

Supervised by Prof. A. E. A. Silva

Alciomar Hollanda

Use of consensus clustering to flare explosion data

Angela Rosa Locateli de Godoy

Recommendation system based on machine learning applied to air pollutants

Gustavo Bartz Guedes

Machine learning techniques to support search and ranking of scientific articles

Supervised by Prof. A. L. S. Gradvohl

Felipe Carreiro Marchi

Indoor positioning and tracking algorithms using Bluetooth low energy beacons

Luís Fenando Lopes Grim

Automatic analysis of magnetogram sequences for solar flares forecasting

Matheus Ferraz da Silveira

Analysis of security aspects of the Internet of Things and proposals for improvements

Natascha Sander Abreu

Embedding cryptographic algorithms in devices for Internet of Things or Wearable Computing.

Supervised by Prof. G. P. Coelho

Arthur Emanuel de Oliveira Carosia

Decision Support for Financial Market based on Machine Learning and Sentiment Analysis.

Ederson Borges

Information Theoretic-based Machine Learning applied to Bioinspired Algorithms

Luis Fernando Panicachi Cocovilo Filho

Improving Stock Market Forecasting with Concept Drift Detection

Matheus Bernardelli de Moraes

Using Multi-Objective Optimization and Multi-Criteria Decision-Making for Exploitation Strategy Definition

Undergraduate students

Supervised by Prof. A. E. A. Silva

Luciano Souza Gomes do Nascimento

Association rules applied to bus traffic problems

Supervised by Prof. A. L. S. Gradvohl

Ana Luísa Fogarin de Sousa Lima

Analysis of solar images at diferent wavelengths to identify active regions in the Sun

Supervised by Prof. G. P. Coelho

Mateus Molinari Ferraz

A Study on the Relationship between Air Pollution and Hospitalizations via Association Rules

Former graduate students

Supervised by Prof. A. L. S. Gradvohl

Tiago Cinto

Solar flare forecasting: a methodology to automate the design of classifiers for events of diverse classes

Supervised by Prof. A. E. A. Silva

Pedro Nunes

Management of strategic surprises based on conceptual systems: detecting threats and opportunities from weak web signals

Former undergraduate students

Supervised by Prof. A. E. A. Silva

Ismael Caldana

Solar explosion forecasting through time-series of X-ray fluxes via MLP neural nets

Jéssica Pereira

Exploration of association rules for attributes of solar explosions via agglomerative hierarchical clustering

Lucas Tonon

Desenvolvimento de Rastreador Web de Sinais Fracos

Pedro Artico

Development of Web Crawler for weak signals search using specific sources

Thiago Viotto

Discovery of association between parameters of M and X solar flares

Supervised by Prof. A. L. S. Gradvohl

Larissa Benevides Vieira

Dynamic activities for the teaching information security in the Internet of Things devices

Letícia Sousa de Oliveira

Automatic analysis of Magnetograms for identification and classification of sunspots

Matheus Dias Queiroz

Study of the man in the middle attack in IPv4/IPv6 transition scenarios

Matheus Evers Rodrigues Fernandes

Using Deep Learning to predict solar explosions using magnetograms

Nahara Calcidoni Pacheco

Review of the SEA System and analysis of Space Weather phenomena related to Sunpy subpackages

Thaís Teche

Analysis of the Multi-layer Perceptrons applied to solar flares forecasting

Vinícius dos Santos

Study of the potential of SunPy for of solar data analysis in Python

Supervised by Prof. G. P. Coelho

Brenda Alexsandra Januário

Sentiment analysis applied to decision support in financial market

Bruno Antunes Carneiro da Silva

Self-organizing Maps for Solar Flare Forecasting

Douglas Araújo

Feature Selection for Solar Flare Forecast based on Evolutionary Algorithms

Gabriel Barros

Study about the Impact of Meteorological Variables in the Prediction of the Concentration of Particulated Matter

João Victor Ignácio

Study of solar flares using data clustering

Leonardo Ferrari Soares

A Comparative Study of Optimization Metaheuristics for Production Strategy Definition in Oil Fields

Marcelo Laendle Junior

Forecasting hospitalizations due to respiratory diseases through atmospheric pollutant concentration data

Mirelle Candida Bueno

A Comparative Study on Hierarchical Clustering for Solar Flare Forecasting

Victor Pedrazzi

Predicting Solar Activity via Artificial Neural Networks Applied to Flow Data of X-Rays

Software Development

The following software were registered at the Brazilian National Institute of Industrial Property (INPI).

  • Astronomus (Register 13168-3)

  • B2-4CEP - Benchmark Tool for Complex Event Processing systems (Register BR512016001623-0)

  • CONsensus (Register BR512014000362-0)

  • FlareCast (Register BR512019002140-1)

  • Gaspra - Gerador de experimentos para imagens astronômicas (Register BR512013000117-0)

  • Navy (Register BR512020002716-4)

  • StockMOS - Stock Market Operation System (Register BR5120160009-0)

  • SEA - SunPy Environment Application (Register BR512019002135-5)

Software for solar flare forecasting

Check our solar flare forecasting system: Guaraci.


Project Partners

Center for Petroleum Studies

Energy Production Innovation Center

São Paulo Astronomy Network - (SPAnet)

Smart Campus - Unicamp

José Roberto Cecatto

National Institute for Space Research - INPE

Sponsor Institutions

We would like to thank the following institutions that donated computing resources to our research

Powered by AWS Cloud Computing

AWS Cloud Credits for Research program

Intel DevCloud


Contact Us

If you have any questions, please contact us in the form in the following link!

Contact form.