Open set recognition is a classification-like task. It accomplished not only by the identification of observations which belong to targeted classes (i.e., among those represented in training sample should be later recognized) but also rejection inputs from other problem domain. The need for proper handling elements beyond interest frequently ignored, even works found literature. This leads improper development learning systems, may obtain misleading results when evaluated their test beds,...

Open set recognition is, more than an interesting research subject, a component of various machine learning applications which is sometimes neglected: it not unusual the existence systems developed on top closed-set assumptions, ignoring error risk involved in prediction. This strictly related to location feature space where prediction has be made, compared training data: distant observations are, less known, higher risk. Proper handling this can necessary situation classification and its...

This paper presents an approximative online algorithm to perform the Kolmogorov–Smirnov test. There is a ubiquitous need for evaluating fitness between statistical distributions and data samples, which this test conveniently meets. Taking some inspiration from challenges of detecting concept drifts in streams, our methodology shows how goodness-of-fit can be used detect such events, taking advantage fact that it non-parametric could adapted handle streams while keeping its original...

<title>Abstract</title> Spectral clustering refers to a class of techniques that depend on the eigenstructure similarity matrix for purpose dividing data points into disjoint clusters, where within same cluster exhibit high and those in different clusters have lower similarity. The objective this work was develop spectral method could be compared algorithms which represent current state art. This investigation conceived novel method, as well 5 policies guide its execution, based graph theory...

The motivation that leads researchers to engage in patenting and licensing activities is a complex topic needs be further investigated, especially emerging countries. In Brazil, universities stand out as the main patent depositors, being important actors country's technological development process. objective of this study present profile from Brazilian motivational factors lead them results their research. Based on literature review, 27 were identified, which classified according nature...

Data assimilation (DA) is an essential issue for operational prediction centers, where a computer code applied to simulate physical phenomena by solving differential equations. The procedure determine the best initial condition combining data from observation and previous forecasting (background) carried out method. Kalman filter (KF) technique assimilation, but it computationally expensive. An approach reduce computational effort DA emulate KF neural network. multi-layer perceptron network...

To cluster a data stream is more challenging task than its regular batch version, having stricter performance constraints. In this paper an approach to problem presented, based on WiSARD, memory-based artificial neural network (ANN) model. This model functioning was reviewed and improved, in order adapt it task. The experimental results obtained support the use of system for analysis streams informative way.

Abstract Spectral clustering techniques depend on the eigenstructure of a similarity matrix to assign data points clusters, so that within same cluster exhibit high and are compared those in different clusters. This work aimed develop spectral method could be algorithms represent current state art. investigation conceived novel method, as well five policies guide its execution, based graph theory embodying hierarchical principles. Computational experiments comparing proposed with six...

<title>Abstract</title> Wireless Sensor Networks (WSNs) have become ubiquitous across various domains, from household applications to industrial grounds, and critical areas like emergency response environmental monitoring. While numerous optimization techniques emerged in the literature enhance network performance by strategically allocating sensor nodes, most studies this field rely on heuristic methods. Designing topology of WSNs is crucial for optimizing node energy consumption,...

Fraudes em transações com cartões de crédito são um desafio global, resultando grandes prejuízos financeiros. Este trabalho propõe simulador dados sintéticos para replicar a dinâmica reais. Esses foram usados criar modelos baseados algoritmos classificação e detecção anomalias, capazes identificar fraudes. Desafios como modelagem sequencial, mudança contexto, feedback atrasado peculiaridades dos abordados. O algoritmo Random Forest destacou-se, detectando 76,7% das fraudes 96,4% precisão.

This work investigates the effect of different data structures on performance and accuracy VG-RAM-based classifiers. weightless neural model is based RAM nodes having very large address input, what suggests use special in order to deal with space time computational costs. Four are explored, including classical one used recent VG-RAM related literature, resulting a novel accurate yet fast setup.

A popular computer puzzle, the game of <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Minesweeper</i> requires its human players to have a mix both luck and strategy succeed. Analyzing these aspects more formally, in our research, we assessed feasibility novel methodology based on reinforcement learning as an adequate approach tackle problem presented by this game. For purpose, employed multi-armed bandit algorithms which were carefully...

Gateways científicos trazem enormes benefícios para usuários finais, simplificando o acesso e ocultando a complexidade da infraestrutura de computação distribuída subjacente. O gateway científico bioinformática, BioinfoPortal, por meio do seu middleware CSGrid, usufrui dos recursos heterogêneos Santos Dumont. No entanto, submissão tarefas ainda exige um esforço significativo, no que tange à decisão melhor configuração leve uma execução eficiente. framework aprendizado máquina, em...