Brazil

Floresta Aleatória é um algoritmo popular e efetivo na resolução de problemas de classificação e regressão. As predições de uma Floresta Aleatória são feitas considerando que cada árvore possui igual contribuição no resultado final. Este trabalho propõe um novo método de ponderação de árvores de regressão com o objetivo de melhorar o poder de predição do modelo. Nossa estratégia é motivada em utilizar medidas de dispersão estatística, como desvio padrão ou erro padrão da média, como indicadores de qualidade da predição na folha. A estratégia de ponderação proposta foi comparada com outros métodos de ponderação. Nessa comparação observou-se que a mesma reduziu o Erro Absoluto Médio em cerca de 30% dos conjuntos de dados estudados.

BRACIS 2020

Optimizing Random Forest from the pondering of regression tree leaves

ponderação em árvores

aprendizado de máquina

random forest

The 9th Brazilian Conference on Intelligent Systems (BRACIS), the 16h Encontro Nacional de Inteligência Artificial e Computacional (ENIAC), and the 12th Concurso de Teses e Dissertações em Inteligência Artificial e Computacional (CTDIAC) was held from 20 to 23 October, 2020.

BRACIS originated from the combination of the two most important scientific events in Brazil in Artificial Intelligence (AI) and Computational Intelligence (CI): the Brazilian Symposium on Artificial Intelligence - SBIA (22 editions), and the Brazilian Symposium on Neural Networks - SBRN (13 editions). ENIAC is a national forum for researchers, practitioners, educators and students to present and discuss innovations, trends, experiences, developments and work in progress in the fields of Artificial and Computational Intelligence. CTDIAC is a satellite event whose goals are to disseminate and award the best dissertations and theses (MSc and PhDs concluded in Brazilian universities from June 2018 to May 2020) in Artificial and Computational Intelligence areas.

In this year, the Symposium on Knowledge Discovery, Mining and Learning (KDMiLe) will be held in conjunction with BRACIS 2020. This sysposium aims at integrating researchers, practitioners, developers, students and users to present their research results, to discuss ideas, and to exchange techniques, tools, and practical experiences – related to the Data Mining and Machine Learning areas.

technical paper

We introduce robot journalists that cover two pressing topics in
Brazilian society: COVID-19 spread and Legal Amazon deforestation. Our approach is able to automatically analyse structured domain data, select relevant
content, generate news texts and publish them on the Web. We provide a thorough description of our system architecture, report on the results of automatic evaluation, discuss some of the advantages of robot-journalism in society, and point out further steps in our work. Corpus and code are publicly available.

Towards Fully Automated News Reporting in Brazilian Portuguese

Even though social networks can provide free space for discussing ideas, people can also use them to propagate Hate Speech and, given the amount of written material in such networks, it becomes necessary to rely on automatic methods for identifying this problem. In this work, we set out to verify the use of some classic Machine Learning algorithms for the task of Hate Speech detection in tweets written in Portuguese, by testing four different models (SVM, MLP, Logistic Regression and Naïve Bayes) with different configurations

Hate Speech Detection in Portuguese with Naïve Bayes, SVM, MLP and Logistic Regression

The Brazilian Digital Library of Thesis and Dissertations (BDTD) provides essential data to support many social sciences investigations. Nevertheless, there is still a lack of computation tools tailored for helping extract and analyze the necessary information from the BDTD library. In this paper, we discuss the development of computational solutions to answer questions from a particular social sciences research using metadata from BDTD. The solutions involve the integration of data processing and presentation techniques, such as string-processing algorithms, knowledge graphs and information visualizations. All programming codes implemented at the scope of the project are available for helping other researchers. The paper also highlights the importance of having researchers from Social Science and Computer Science working together, what motivates future collaborations in these areas.

Computational Mining on IBICT BDTD’s Thesis and Dissertation Metadata for Supporting Social Science Research

Glaucoma is a significant cause of blindness in the world. Doctors use computerized images to detect these diseases. Early detection of the disease increases the chances of treatment, reducing the adverse effects. This work proposes an evaluation of texture maps combinations as input to Convolutional Neural Networks for glaucoma classification in retinal images. In our experiments, we used three textures maps, three CNN architectures, and three classifiers. We achieve a Kappa = 0.708+-0.054 and a Accuracy = 0.859+-0.021. We conclude that using the combination of texture maps can improve the automatic detection of glaucoma compared to single-channel inputs, and could be used by state-of-the-art methods to improve their classification rates.

Evaluation of Texture Maps as Input to Extract Deep Features in Glaucoma Diagnosis

Approximate Computing is an emerging paradigm that takes advantage of inherently error resilient digital circuits to design circuits with higher energetic efficiency, lower delay, or a smaller occupied area on the chips. Traditional approaches do not handle multiple objectives and metaheuristics appear as a proper alternative. In particular, Multiobjective Cartesian Genetic Programming (MOCGP) can find good solutions to the design and optimization of approximate circuits. The performance of CGP depends on the mutation adopted, as normally CGP only uses mutation for creating new solutions. However, to the best of our knowledge, just the traditional point mutation (PM) was used by the previously proposed MOCGP. Thus, the literature lacks an analysis of the best mutation operators of MOCGP. We propose here the analysis of PM and Single Active Mutation(SAM) on the multiobjective optimization of 15 heterogeneous combinational logic circuits from scratch and starting with a feasible solution. The results indicate that SAM obtained better results than PM.

On the Analysis of Mutation Operators in Multiobjective Cartesian Genetic Programming for designing Combinational Logic Circuits

This presentation brings the application of the Fuzzy Triangular Naive Bayes method for assessment of users in a simulator for training in gynecological examination. This method is interesting for data that can be modeled by triangular statistical distribution. The method was analyzed with simulated data that represent two variables present in one of the stages of a gynecological examination. The performance of Fuzzy Triangular Naive Bayes are considered satisfactory when compared to a version that does not use fuzzy logic.


Using the Fuzzy Triangular Naive Bayes to Assess Users in Gynecological Examination Training

In probabilistic planning, an usual problem is the Stochastic Shortest Path (SSP) in a Markov Decision Process (MDP), where the objective is to find policies that reach a goal with the minimum expected accumulated cost. In the presence of states from where it is not possible to reach a goal, this criterion is not well-defined in these states. Because of that, there are extensions and other formulations for this particular case, like the GUBS criterion. Although algorithms to solve SSPs under this criterion exist, they need to make sweeps through the entire state space. With that in mind, this work proposes UCT-GUBS: an online planning algorithm based on UCT to solve SSPs under the GUBS criterion. The results of the experiments that were performed show empirically that the algorithm makes the desired trade-off between cost and probability to goal when its parameters are varied, with the ability of making a decision in less or more time depending on the time limit provided to the algorithm, an important characteristic of UCT.

Monte Carlo Tree Search Algorithm for SSPs Under the GUBS Criterion

COVID-19 has caused thousands of deaths worldwide and undermined the economy of those countries which have been ravaged by the pandemic.
In this paper, we present an agent-based simulation to study the propagation of
COVID-19 in Ibirama (SC). The simulation uses real population, educational,
and business data, as well as the actual map of the city containing its traffic
network and location of educational institutions and businesses. The simulation
considers students and workers agents, which move around the environment and
interact with each other. Different social distancing scenarios of these agents
were simulated. The results obtained indicate that social distancing policies are
effective in reducing the spread of the virus.

An Agent-based Simulation to Study the Spread of COVID-19in Ibirama(SC)

K-Shortest Loopless Paths (KSLP) consiste em encontrar K | K > 1 caminhos, sem repetições de vértices, classificados em ordem crescente de distância, em um grafo G(V,E), no qual V representa o conjunto de vértices e E o conjunto dos arestas que os conecta. Dada a dificuldade de obter boa performance para solução do problema, este trabalho propõe a paralelização em GPUs CUDA do algoritmo mais veloz existente na literatura para o problema.

Proposta de paralelização em GPUs CUDA do algoritmo MPS para resolução do problema de encontrar os K-Caminhos Mais Curtos Sem Repetições de Vértices em grafos direcionados e não direcionados

This paper describes the application of supervised and unsupervised
machine learning in large volumes of open governmental data from INEP.
This work uses the following algorithms: K-Nearest Neighbors, Logistic
Regression, Decision Tree, Random Forest and K-means. The methodology is
based on the CRISP-DM and KDD processes, requiring the use of the
DataBricks cloud platform. In addition, the Hadoop and Apache Spark cluster
Technologies were also used. Such technologies provided high processing
power for the execution of the experiments. This enabled the performance
evaluation of the models and the discovery of knowledge about Brazilian basic
education.

Downloads

Next from BRACIS 2020

Towards Fully Automated News Reporting in Brazilian Portuguese

Similar lecture

ML-MDLText: um método de classificação de textos multirrótulo de aprendizado incremental

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from BRACIS 2020

Towards Fully Automated News Reporting in Brazilian Portuguese

Similar lecture

ML-MDLText: um método de classificação de textos multirrótulo de aprendizado incremental

Downloads