The wavelets have become increasingly popular in the field of bioinformatics due to their capacity in multiresolution analysis and space-frequency localization; the latter particularity is acquired due to a moving window that runs through the analyzed space. As a feature, they have a better ability to capture hidden components of biological data and an efficient link between biological systems and the mathematical objects used to describe them. The decomposition of signals/sequences at different levels of resolution allows obtaining distinct characteristics in each level. The energy (variance) obtained at each level provides a new set of information that can be used to search similarities between sequences. We show that the behavior of GC-content sequence can be succinctly described regarding the non-decimated wavelet transform, and we indicate how this characterization can be used to improve clustering of the similar strains of the genome of the Mycobacterium tuberculosis, having a very efficient level of detail. The clustering analysis using the energy obtained at each level of the analyzed sequences was essential to verify the dissimilarity of the sequences.

The integration of distributed generation (DG) sources in the electric energy systems may bring new problems that need attention, one of these problems is the occurrence of unintentional islanding. Islanding is a condition in which part of the distribution network is disconnected from the system, and consumer units are still powered by one or more DGs, which can cause damage to equipment and pose risks to the safety of technicians. This paper shows an islanding detection method (IDM) in Power Systems with DG based on statistical signal processing. We used a MathWorks Simulink model of a grid-connected 250 kW photovoltaic (PV) array to simulate the behavior of the three-phase voltage signal in the point of common coupling (PCC) under the nominal operation, islanding condition, and fault condition using different load compositions. Principal Component Analysis (PCA) was used to extract the transitory events from the voltage signals, and then we used second-, third-, and fourth-order cumulants to generate features and the best ones were selected using the Fisher’s Discriminant Ratio (FDR). A Radial Basis Function Network (RBFN) makes the classification of the events. We found that, for this setup, we can achieve detection rates of 99% for both islanding condition detection and fault occurrence classification, no matter the power mismatch between the load and the DG.

Voltage disturbances are the most frequent cause of a large range of disruption in industrial, commercial, and residential power supply systems. These disturbances are often referred to as power quality problems and affect the Power Systems causing substantial losses. To avoid the storage of a large amount of data, the first task in monitoring the power quality is the realtime detection of disturbances, which must be performed by an accurate and low-complexity system. This paper proposes a low-complexity system for power quality disturbance detection. The method makes innovative use of simple features extracted from reduced segments of the monitored voltage waveform. The extract features (the mean value, variance, energy, and the maximum and minimum values of the filtered voltage signals) require low computational effort and allow a considerable dimensional reduction of the signals, leading to simple detection algorithms. The proposed method achieves high detection rates on both simulated and real signals.

Resumo. Este trabalho propõe a utilização de técnicas computacionais baseadas em Inteligência Computacional (IC) para a estimação dos requerimentos energéticos em gado bovino. As técnicas abordadas são: Redes Neurais Perceptron Multicamadas (MLP) e Sistemas de Inferência Neuro-Fuzzy Adaptativos (ANFIS). Ambas técnicas foram utilizadas para a estimação da Energia Metabolizável Ingerida (MEI) a partir de um banco de dados de 840 animais. Os parâmetros utilizados para a modelagem foram: gênero, raça, sistema de alimentação, peso corporal vazio médio (AEBW) e ganho de corpo vazio (EBG). Para os modelos MLP foram utilizadas arquiteturas contendo uma ou duas camadas escondidas com um número máximo de neurônios em cada camada de 10. Para a elaboração dos modelos ANFIS, foram utilizadas duas técnicas de clusterização, Fuzzy C-Means (ANFIS-FCM) e Clusterização Subtrativa (ANFIS-SC), que foram responsáveis pela geração do Sistema de Inferência Fuzzy (FIS) inicial, posteriormente reestruturado para a arquitetura ANFIS. Os modelos MLP obtiveram correlação média superior a 80%, enquanto os modelos ANFIS obtiveram resultados da ordem de 74%. A técnica MLP supera a abordagem clássica para a estimação da MEI, baseada em regressão linear múltipla (MLR). As técnicas de IC se mostraram promissoras e boas alternativas aos modelos comumente usados.

No abstract