Submitted by MONARA PEREIRA DA ROSA CARVALHO null (momoprc@gmail.com) on 2017-03-17T12:47:54Z No. of bitstreams: 1 MONARA_Dissertacao.pdf: 2926386 bytes, checksum: 52ab3ee5e454a3b74043a0bbef9630de (MD5) / Approved for entry into archive by Juliano Benedito Ferreira (julianoferreira@reitoria.unesp.br) on 2017-03-21T19:15:17Z (GMT) No. of bitstreams: 1 carvalho_mpr_me_ilha.pdf: 2926386 bytes, checksum: 52ab3ee5e454a3b74043a0bbef9630de (MD5) / Made available in DSpace on 2017-03-21T19:15:17Z (GMT). No. of bitstreams: 1 carvalho_mpr_me_ilha.pdf: 2926386 bytes, checksum: 52ab3ee5e454a3b74043a0bbef9630de (MD5) Previous issue date: 2017-01-20 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / This work analysis load data from desegregated levels that presented difficulties to load forecasting with several methods due to variation in electrical energy consumption. The application proposed in this work is short-term load forecasting to a university building by GRNN (General Regression Neural Network) considering the bottom up approach and using a moving average filter to deal with the missing or wrong data. It is presented the system that provides the data as well as the methods used for pre-processing and realize the forecasting. The results are evaluated by MAPE (Mean Absolute Perceptual Error) and are considered good when compared with other methods. / Este trabalho destaca a análise de dados provenientes de locais com níveis de consumo mais desagregados que apresentam dificuldades para previsões de demanda com vários métodos devido à alta variação no consumo de energia elétrica. Apresenta-se resultados de previsões de demanda de curto prazo da energia elétrica consumida em um bloco de uma universidade por meio da rede neural de regressão generalizada (GRNN), utilizando a abordagem de modelagem de dados de baixo para cima e tratamento de ruídos e dados faltantes no banco de dados através da aplicação de um filtro de médias móveis. É apresentado o local que fornece as informações para os estudos e a etapa de pré-processamentos dos dados. Foi possível analisar a assertividade das previsões de acordo com o cálculo do MAPE, mostrando vantagens ao se comparar a outros métodos utilizados para os mesmos fins.

Short-term traffic flow forecasting is a critical function in advanced traffic management systems (ATMS) and advanced traveler information systems (ATIS). Accurate forecasting results are useful to indicate future traffic conditions and assist traffic managers in seeking solutions to congestion problems on urban freeways and surface streets. There is new research interest in short-term traffic flow forecasting due to recent developments in ITS technologies. Previous research involves technologies in multiple areas, and a significant number of forecasting methods exist in literature. However, forecasting reliability is not properly addressed in existing studies. Most forecasting methods only focus on the expected value of traffic flow, assuming constant variance when perform forecasting. This method does not consider the volatility nature of traffic flow data. This paper demonstrated that the variance part of traffic flow data is not constant, and dependency exists. A volatility model studies the dependency among the variance part of traffic flow data and provides a prediction range to indicate the reliability of traffic flow forecasting. We proposed an ARIMA-GARCH (Autoregressive Integrated Moving Average- AutoRegressive Conditional Heteroskedasticity) model to study the volatile nature of traffic flow data. Another problem of existing studies is that most methods have limited forecasting abilities when there is missing data in historical or current traffic flow data. We developed a General Regression Neural Network(GRNN) based multivariate forecasting method to deal with this issue. This method uses upstream information to predict traffic flow at the studied site. The study results indicate that the ARIMA-GARCH model outperforms other methods in non-missing data situations, while the GRNN model performs better in missing data situations.

Orientador: Anna Diva Plasencia Lotufo / Resumo: This work analysis load data from desegregated levels that presented difficulties to load forecasting with several methods due to variation in electrical energy consumption. The application proposed in this work is short-term load forecasting to a university building by GRNN (General Regression Neural Network) considering the bottom up approach and using a moving average filter to deal with the missing or wrong data. It is presented the system that provides the data as well as the methods used for pre-processing and realize the forecasting. The results are evaluated by MAPE (Mean Absolute Perceptual Error) and are considered good when compared with other methods. / Mestre

This thesis examines the determinants of capital structure in the MENA coun- tries. The main interest is to investigate both financial firms specially banks and non-financial firms. This study test the main theories of capital structure, namely: trade off theory and pecking order theory. The countries included in this thesis are Saudi Arabia, United Arab Emirates (Include both Abo-Dhabi and Dubai stock indexes), Bahrain, Qatar, Kuwait, Oman, Egypt, Morocco, Tunisia, Palestine and Jor- dan. The characteristics it covers as suggested by previous literature are tangibility, profitability, risk, debt tax shield, growth, dividends,size, cash flow and liquidity. It will also investigate the effect of the industry, credit rating and ownership structure on the capital structure This study also investigates the determinants of capital structure in Islamic and conventional banks. This is one of the first attempts to empirically examine the determinants of capital structure in Islamic and conventional banks in general and in MENA countries in particular. This study fills the gap in this important area of research and can provide a base for future research on capital structure in Islamic banks. This thesis use different models to test the capital structure and these are Panel data models (OLS, Fixed, and Random); Tobit and Dynamical model (Arellano-Bover Blundell-Bond), Structural Equation Modeling (SEM) and Generalised Regression Neural Networks (GRNN). The results suggest that the three methods used in this study lead to similar re- sults with a few exceptions in some countries. This thesis finds that the relation between leverage and the determinants of capital structure is different when using the market or the book leverage. It also finds that the determinants of capital struc- ture between the MENA countries are different. For example, profitability attribute relation with leverage follow the trade-off theory in some countries and follow the picking order theory in other countries. Also, liquidity is significant in all the countries in the sample and have a negative relation to leverage. In addition, tangibility is found to have a mixed results with some countries following the trade-off theory and other countries which follow the trade-off theory but overall it is a key determinant of capital structure. Additionally, the findings show that although that the majority of firms in the MENA countries don’t pay dividends the relation between the long term debt and leverage is negative in all the countries in the sample. The growth opportunities have a negative relation in Bahrain, Egypt, Jordan, Kuwait, Morocco, Palestine, Qatar and Tunisia but positive in rest of the countries. The cash flow attribute have a negative relation with leverage in all the countries in the sample except Saudi Arabia and Qatar when using the short and long term debt. Furthermore, the ownership variable is expected to have a negative relation when the ultimate owner is an institution. The results show that overall when there is an ultimate owner the leverage will have a negative relation. Suggesting that ultimate owners will force managers to keep a low debt in firms capital structure. This PhD also attempt to investigate the capital structure in banks within the MENA countries. A special focus is on the differences between the Islamic banks and conventional banks capital structure. First, the findings show that the banks follow the same determinants of capital structure as non-financial firms and that regulations are not the main determinant of capital structure in banks. Then, This study show that there is a difference in capital structure of Islamic banks in com- parison with conventional banks. The findings for the dividends variable show that Islamic banks do not follow the pecking order theory but conventional banks don’t. The results of the size variable show that when Islamic banks are large they use less debt in their capital structure. Growth variable show mixed results depending on the use of book or market leverage. Ownership structure show that when there is an ultimate owner leverage increase which is the reverse of the relation in the non-financial firms. The age variable is negative in relation to the book leverage and positive with the market leverage. Also, credit rating relation is different between the two banks, as it is positive with the conventional banks and negative with Islamic banks. Therefore, this study conclude that the main capital structure theories are applicable to MENA countries. Also indicate that Islamic banks have a different capital structure to conventional banks.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Protection systems used in distribution networks of electricity are not able to detect short circuits with high contact resistance due to the low currents generated, endangering the population and degrading the quality of the energy supplied. In this sense, this paper presents a new methodology for detecting high-impedance faults (HIF) in distribution networks. The developed algorithm has the main advantage the fact also detect and classify other types of transient as, for example, switching capacitor banks, transformers and loads. This characteristic decreases the improper operation caused by transient switching. Another feature of the developed method is no need to install additional equipment on the network which greatly reduces the cost of implementation. Additionally, the paper also discusses the causes, consequences and characteristics of HIF in order to evidence the difficulties related to their detection. A brief review of the models proposed in the literature for computational simulation of HIF is also discussed, and the model used in this work is described in detail. To evaluate the performance of the algorithm developed a series of tests with different fault scenarios high impedance were made. Furthermore, other various types of transients that are normal in the feeders were tested. The good results obtained, combined the simplicity of the method and does not need to install additional equipment feeders, makes a promising technique for real applications. / Os sistemas de proteção utilizados em redes de distribuição de energia elétrica não são capazes de detectar curtos-circuitos com alta resistência de contato devido às reduzidas correntes geradas, colocando em risco a população e degradando a qualidade da energia fornecida. Neste sentido, este trabalho visa apresentar uma nova metodologia para detecção de faltas de alta impedância (FAI) em redes de distribuição de energia elétrica. O algoritmo desenvolvido possui como principal vantagem o fato de também detectar e classificar outros tipos de transitórios como, por exemplo, chaveamento de banco de capacitores, transformadores e manobras de ramais. Essa característica diminui consideravelmente as atuações indevidas causadas por transitórios oriundos de manobras. Outra característica do método desenvolvido é a não necessidade de instalação de equipamentos adicionais na rede, o que reduz consideravelmente o custo de sua implementação. Além disso, o trabalho também aborda as causas, consequências e características das FAI de forma a evidenciar as dificuldades relacionadas à sua detecção. Uma breve revisão sobre os modelos propostos na literatura para simulação computacional de FAI também é abordado, sendo que o modelo utilizado nesse trabalho é descrito em detalhes. Para avaliar o desempenho do algoritmo desenvolvido, uma série de testes com diferentes cenários de falta de alta impedância foram realizados. Além disso, outros tipos de transitórios que são normais nos alimentadores foram testados. Os bons resultados obtidos, aliado a simplicidade do método e a não necessidade de instalação de equipamentos adicionais nos alimentadores, torna a técnica promissora para aplicações reais.

The relationship between Objective Metrics (OM) and Subjective Assessments (SA) has been analyzed by people using different methods. This paper continues Gaspar Gil Gómez’s research over test method development by use of SOM-GRNN, aiming to find correlations between OM and SA. In this paper, CAE simulation is performed to analyze the relationship between OM and vehicle parameters. First impression test is refined and one more dataset has been added in order to populate SOM-GRNN map. This paper also conducts analysis over SOM and GRNN algorithms, and explores several possible applications using SOM-GRNN map. Finally the whole SOM-GRNN system is integrated and a User Interface is built in GUI for future research and application. The system can still be improved by populating database, refining SA evaluation method and optimizing SOM-GRNN training algorithms. / Relationen mellan målmetri (OM) och subjektiv bedömning (SA) har analyserats av personer som använder olika metoder. I detta dokument fortsätter Gaspar Gil Gómezs forskning kring testmetodutveckling med hjälp av SOM-GRNN, som syftar till att finna korrelationer mellan OM och SA. I detta papper utförs CAE-simulering för att analysera förhållandet mellan OM och fordonsparametrar. Första intryckstestet är raffinerat och ytterligare en dataset har lagts till för att fylla i SOM-GRNN-kartan. I detta dokument analyseras även SOM- och GRNN-algoritmer, och undersöker flera möjliga tillämpningar med hjälp av SOM-GRNN kartan. Slutligen är hela SOM-GRNN-systemet integrerat och ett användargränssnitt är byggt i GUI för framtida forskning och tillämpning. Systemet kan fortfarande förbättras genom att fylla i databasen, förädla SA utvärderingsmetod och optimera SOM-GRNN träningsalgoritmer.