Rede neural artificial para detecção de sobremortalidade atribuível à cólera no Ceará

OBJETIVO: Avaliar as redes neurais recorrentes enquanto técnica preditiva para séries temporais em saúde. MÉTODOS: O estudo foi realizado durante uma epidemia de cólera ocorrida no Estado do Ceará, em 1993 e 1994, a partir da sobremortalidade tendo como causa básica as infecções intestinais mal definidas (CID-9). O número mensal de óbitos por essa causa, referente ao período de 1979 a 1995 no Estado do Ceará, foram obtidos do Sistema de Informação de Mortalidade (SIM) do Ministério da Saúde. Estruturou-se uma rede com dois neurônios na camada de entrada, 12 na camada oculta, um neurônio na camada de saída e um na camada de memória. Todas as funções de ativação eram a função logística. O treinamento foi realizado pelo método de backpropagation, com taxa de aprendizado de 0,01 e momentum de 0,9, com dados de janeiro de 1979 a junho de 1991. O critério para fim do treinamento foi atingir 22.000 epochs. Compararam-se os resultados com os de um modelo de regressão binomial negativa. RESULTADOS: A predição da rede neural a médio prazo foi adequada, em dezembro de 1993 e novembro e dezembro de 1994. O número de óbitos registrados foi superior ao limite do intervalo de confiança. Já o modelo regressivo detectou sobremortalidade a partir de março de 1992. CONCLUSÕES: A rede neural se mostrou capaz de predição, principalmente no início do período, como também ao detectar uma alteração concomitante e posterior à ocorrência da epidemia de cólera. No entanto, foi menos precisa do que o modelo de regressão binomial, que se mostrou mais sensível para detectar aberrações concomitantes à circulação da cólera.

Estudo do recife artificial a construir na Leirosa: aplicação de modelação numérica para simulação da propagação da agitação marítima

A temática dos recifes artificiais multifuncionais é relativamente recente, sendo que o primeiro recife artificial multifuncional construído data do ano de 1999 (Perth, Austrália). A palavra multifuncional está associada aos múltiplos propósitos que se podem atingir com a construção de uma estrutura destas, sendo eles, a proteção costeira, o aumento da biodiversidade local, a melhoria da qualidade das ondas para o Surf e a promoção do turismo ligado aos desportos de ondas. Para dar resposta a um caso de proteção costeira, na zona marítima adjacente à praia de Leirosa, Portugal, foi pensada uma construção de um recife artificial que funcione como obra de proteção do sistema dunar local e que, adicionalmente melhore as condições locais para a prática de Surf. Este trabalho descreve a análise de duas soluções de recife (em forma de “V”, formando um ângulo de 45º e 66º, entre si), através dos valores das características das ondas (altura, período e direção) e parâmetros de surfabilidade (linha de rebentação, número de Iribarren e ângulo de rebentação), para uma gama alargada de condições de agitação frequente. Para tal, foi necessário caracterizar a agitação marítima, através do modelo numérico SWAN para determinação dos casos de agitação mais frequentes na zona marítima adjacente ao local de implantação do recife e para, posteriormente, se proceder à sua utilização no modelo numérico DREAMS, que permitiu a simulação da propagação das ondas sobre o recife. A comparação dos resultados do modelo numérico DREAMS para as situações de com e sem recife (para as duas soluções de recife) permitiu avaliar a influência do mesmo em termos de alturas de onda, linha de rebentação e ângulo de rebentação, tendo-se chegado a resultados satisfatórios do ponto de vista do melhoramento das condições locais para a prática do Surf.

Scheduling distributed energy resources in an isolated grid - an artificial neural network approach

Power Systems (PS), have been affected by substantial penetration of Distributed Generation (DG) and the operation in competitive environments. The future PS will have to deal with large-scale integration of DG and other distributed energy resources (DER), such as storage means, and provide to market agents the means to ensure a flexible and secure operation. Virtual power players (VPP) can aggregate a diversity of players, namely generators and consumers, and a diversity of energy resources, including electricity generation based on several technologies, storage and demand response. This paper proposes an artificial neural network (ANN) based methodology to support VPP resource schedule. The trained network is able to achieve good schedule results requiring modest computational means. A real data test case is presented.

Combining artificial intelligence techniques for the training of power system control centre operators

Control Centre operators are essential to assure a good performance of Power Systems. Operators’ actions are critical in dealing with incidents, especially severe faults, like blackouts. In this paper we present an Intelligent Tutoring approach for training Portuguese Control Centre operators in incident analysis and diagnosis, and service restoration of Power Systems, offering context awareness and an easy integration in the working environment.

Perspectives and progress of ecology and conservation science in the Azores: the possible contribution of Artificial Intelligence

EPIA 2013 - XVI Portuguese Conference on Artificial Intelligence Angra do Heroísmo, Azores, Portugal, 9 – 12 September.

Feasibility study on multifunctional artificial reefs for the Azores

Tese de Doutoramento, Geografia (Ordenamento do Território), 25 de Novembro de 2013, Universidade dos Açores.

Mixed integer non linear programming and artificial neural network based approach to ancillary services dispatch in competitive electricity markets

Ancillary services represent a good business opportunity that must be considered by market players. This paper presents a new methodology for ancillary services market dispatch. The method considers the bids submitted to the market and includes a market clearing mechanism based on deterministic optimization. An Artificial Neural Network is used for day-ahead prediction of Regulation Down, regulation-up, Spin Reserve and Non-Spin Reserve requirements. Two test cases based on California Independent System Operator data concerning dispatch of Regulation Down, Regulation Up, Spin Reserve and Non-Spin Reserve services are included in this paper to illustrate the application of the proposed method: (1) dispatch considering simple bids; (2) dispatch considering complex bids.

Controlo e optimização de sistemas AVAC recorrendo a técnicas de inteligência artificial

Trabalho Final de Mestrado para obtenção do grau de Mestre em Engenharia Mecânica

Environmental Decision Support: a Distributed Artificial Intelligence Approach

Decision making in any environmental domain is a complex and demanding activity, justifying the development of dedicated decision support systems. Every decision is confronted with a large variety and amount of constraints to satisfy as well as contradictory interests that must be sensibly accommodated. The first stage of a project evaluation is its submission to the relevant group of public (and private) agencies. The individual role of each agency is to verify, within its domain of competence, the fulfilment of the set of applicable regulations. The scope of the involved agencies is wide and ranges from evaluation abilities on the technical or economical domains to evaluation competences on the environmental or social areas. The second project evaluation stage involves the gathering of the recommendations of the individual agencies and their justified merge to produce the final conclusion. The incorporation and accommodation of the consulted agencies opinions is of extreme importance: opinions may not only differ, but can be interdependent, complementary, irreconcilable or, simply, independent. The definition of adequate methodologies to sensibly merge, whenever possible, the existing perspectives while preserving the overall legality of the system, will lead to the making of sound justified decisions. The proposed Environmental Decision Support System models the project evaluation activity and aims to assist developers in the selection of adequate locations for their projects, guaranteeing their compliance with the applicable regulations.

Multiple Linear Regression and Artificial Neural Networks to Predict Time and Efficiency of Soil Vapor Extraction

The prediction of the time and the efficiency of the remediation of contaminated soils using soil vapor extraction remain a difficult challenge to the scientific community and consultants. This work reports the development of multiple linear regression and artificial neural network models to predict the remediation time and efficiency of soil vapor extractions performed in soils contaminated separately with benzene, toluene, ethylbenzene, xylene, trichloroethylene, and perchloroethylene. The results demonstrated that the artificial neural network approach presents better performances when compared with multiple linear regression models. The artificial neural network model allowed an accurate prediction of remediation time and efficiency based on only soil and pollutants characteristics, and consequently allowing a simple and quick previous evaluation of the process viability.

Solar Intensity Forecasting using Artificial Neural Networks and Support Vector Machines

This paper presents several forecasting methodologies based on the application of Artificial Neural Networks (ANN) and Support Vector Machines (SVM), directed to the prediction of the solar radiance intensity. The methodologies differ from each other by using different information in the training of the methods, i.e, different environmental complementary fields such as the wind speed, temperature, and humidity. Additionally, different ways of considering the data series information have been considered. Sensitivity testing has been performed on all methodologies in order to achieve the best parameterizations for the proposed approaches. Results show that the SVM approach using the exponential Radial Basis Function (eRBF) is capable of achieving the best forecasting results, and in half execution time of the ANN based approaches.

Artificial Intelligence in geospatial analysis: applications of self-organizing maps in the context of geographic information science.

A thesis submitted in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Information Systems.

Trypanosoma cruzi parasitemia observed in immunocompromised patients: the importance of the artificial xenodiagnosis

Trypanosoma cruzi parasitemia observed in immunocompromised patients (transplant or positive HIV) occurred more frequently by the artificial xenodiagnosis method (10/38) compared with hemoculture (2/38), given the same quantity of blood. Other ways of diagnosis, like mice inoculation (5/38), QBC and buffy coat (2/38), were evaluated also. This result showed the importance of the artificial xenodiagnosis. The other techniques increased only one more patient positive.

Artificial antibodies for troponin T by its imprinting on the surface of multiwalled carbon nanotubes: Its use as sensory surfaces

A novel artificial antibody for troponin T (TnT) was synthesized by molecular imprint (MI) on the surface of multiwalled carbon nanotubes (MWCNT). This was done by attaching TnT to the MWCNT surface, and filling the vacant spaces by polymerizing under mild conditions acrylamide (monomer) in N,N′-methylenebisacrylamide (cross-linker) and ammonium persulphate (initiator). After removing the template, the obtained biomaterial was able to rebind TnT and discriminate it among other interfering species. Stereochemical recognition of TnT was confirmed by the non-rebinding ability displayed by non-imprinted (NI) materials, obtained by imprinting without a template. SEM and FTIR analysis confirmed the surface modification of the MWCNT. The ability of this biomaterial to rebind TnT was confirmed by including it as electroactive compound in a PVC/plasticizer mixture coating a wire of silver, gold or titanium. Anionic slopes of 50 mV decade−1 were obtained for the gold wire coated with MI-based membranes dipped in HEPES buffer of pH 7. The limit of detection was 0.16 μg mL−1. Neither the NI-MWCNT nor the MWCNT showed the ability to recognize the template. Good selectivity was observed against creatinine, sucrose, fructose, myoglobin, sodium glutamate, thiamine and urea. The sensor was tested successfully on serum samples. It is expected that this work opens new horizons on the design of new artificial antibodies for complex protein structures.