Global climate models predict changes in precipitation patterns in many areas of the world. Extreme precipitation in particular is poorly represented in climate models and there are significant difficulties involved in assessing the frequency and severity of future extreme precipitation events. In this study, several methods have been reviewed and compared for estimating projected changes in Intensity-Duration-Frequency (IDF) curves, commonly used in urban hydrology. A theoretical approach based on geostatistical considerations is employed to derive reasonable areal-reduction factors that make it possible to compare gridded model data with observations. The mean value method and QQ-mapping have been used to remove biases from modeled data. A simple scaling model has been developed to construct IDF curves using the bias-corrected modeled data for the control and future climate. To investigate uncertainties in predicted changes, different simulations from the North American Regional Climate Change Assessment Program (NARCCAP) have been analyzed.

Global climate models predict changes in precipitation patterns in many areas of the world. Extreme precipitation in particular is poorly represented in climate models and there are significant difficulties involved in assessing the frequency and severity of future extreme precipitation events. In this study, several methods have been reviewed and compared for estimating projected changes in Intensity-Duration-Frequency (IDF) curves, commonly used in urban hydrology. A theoretical approach based on geostatistical considerations is employed to derive reasonable areal-reduction factors that make it possible to compare gridded model data with observations. The mean value method and QQ-mapping have been used to remove biases from modeled data. A simple scaling model has been developed to construct IDF curves using the bias-corrected modeled data for the control and future climate. To investigate uncertainties in predicted changes, different simulations from the North American Regional Climate Change Assessment Program (NARCCAP) have been analyzed.

Made available in DSpace on 2014-06-12T17:37:10Z (GMT). No. of bitstreams: 2 arquivo2398_1.pdf: 2577610 bytes, checksum: 42437f0a584393092b0393e11fcd3d48 (MD5) license.txt: 1748 bytes, checksum: 8a4605be74aa9ea9d79846c1fba20a33 (MD5) Previous issue date: 2009 / Conselho Nacional de Desenvolvimento Científico e Tecnológico / Para dimensionamento da estrutura dos sistemas de microdrenagem e macrodrenagem, deve-se saber qual a quantidade de precipitação na área a ser drenada, associada a um período de retorno. As equações de chuvas intensas são essenciais no processo de drenagem urbana e rural de águas pluviais, apresentando grande importância para os projetos de hidráulica dos centros urbanos e sistemas de drenagem agrícola que necessitam definir a chuva de projeto para estimar a vazão de projeto dos mesmos. Além disso, as relações IDF contribuem para a diminuição da margem de erros, minimizando-se o risco de enchentes causado por chuvas de alta intensidade, como também são indispensáveis para elaboração e controle dos sistemas de drenagem e irrigação. Em função da pouca disponibilidade de informações no estado de Pernambuco quanto às equações de chuvas intensas, este trabalho tem como objetivo principal a determinação da relação IDF para algumas localidades do território pernambucano. Para isso, foram utilizados os dados de 12 estações pluviográficas e 11 estações pluviométricas, com séries históricas de 8 a 14 anos de observações (pluviógrafos) e séries de 10 a 34 anos de registros (pluviômetros). Além disso, foram realizadas comparações entre as equações geradas por dados pluviográficos e pluviométricos, objetivando avaliar a qualidade das curvas IDF oriundas de dados pluviométricos. O critério adotado para o estabelecimento das séries históricas foi o de séries anuais. Tais séries históricas foram obtidas através de análise de pluviogramas digitalizados, no caso das estações pluviográficas, e por meio da metodologia de desagregação de chuvas de um dia em tempos menores, para os pluviômetros. Na análise de freqüência das séries anuais foram aplicados o modelo de distribuição de Gumbel e o de Weibull, sendo que a segunda distribuição, na maioria dos casos estudados, apresentou os melhores ajustes com o teste de Kolmogorov-Smirnov ao nível de significância de 5% e com R² variando de 0,9199 a 0,9907. Para determinação dos parâmetros da equação de chuvas intensas foram usadas as metodologias de Regressão Linear e a Regressão não-linear, onde o segundo método apresentou melhor ajuste dos parâmetros em todas as equações desenvolvidas. Por fim, as equações geradas por meio de registros pluviométricos apresentaram bons ajustes em comparação com as relações IDF obtidas com dados de pluviogramas, com R² variando de 97% a 99,9%

Generally urban drainage systems are built to protect urban property and control runoff. Moreover, these systems collect the runoff for storage purposes to serve society through sufficient water supply to meet the needs of demand, irrigation, and drainage. Urban environments are exposed to risks of extreme hydrological events. Therefore, urban water systems and their management are critical. Precipitation data are crucial, but may be prone to errors due to the lack of information e.g., short length of records. In this thesis, a Monte Carlo simulation and regional frequency analysis based on L-moments approach were utilized during the research in order to estimate the uncertainty in the Intensity Duration Frequency (IDF) curves by using historical precipitation data from Environment Canada (EC) weather stations and simulating a new series of data through a weather generator (WG) model. The simulations were then disaggregated from daily into hourly data for extraction of the annual maximum precipitation for different durations in hours (1, 2, 6, 10, 12, and 24). Regional frequency analysis was used to form the sites into groups based on homogeneity test results, and the quantile values were computed for various sites and durations with the return periods (T) in years (2, 10, 20, and 100). As a result, the regional frequency analysis was used to estimate the regional quantile values based on L-moment approach. Moreover, the box and whisker plots were utilized to display the results. When the return periods and durations increased, the uncertainty slightly increased. The historical IDF curves of London site falls within the regional simulated IDF curves. Furthermore, 1000 runs have been generated by using the weather generator.

Generally urban drainage systems are built to protect urban property and control runoff. Moreover, these systems collect the runoff for storage purposes to serve society through sufficient water supply to meet the needs of demand, irrigation, and drainage. Urban environments are exposed to risks of extreme hydrological events. Therefore, urban water systems and their management are critical. Precipitation data are crucial, but may be prone to errors due to the lack of information e.g., short length of records. In this thesis, a Monte Carlo simulation and regional frequency analysis based on L-moments approach were utilized during the research in order to estimate the uncertainty in the Intensity Duration Frequency (IDF) curves by using historical precipitation data from Environment Canada (EC) weather stations and simulating a new series of data through a weather generator (WG) model. The simulations were then disaggregated from daily into hourly data for extraction of the annual maximum precipitation for different durations in hours (1, 2, 6, 10, 12, and 24). Regional frequency analysis was used to form the sites into groups based on homogeneity test results, and the quantile values were computed for various sites and durations with the return periods (T) in years (2, 10, 20, and 100). As a result, the regional frequency analysis was used to estimate the regional quantile values based on L-moment approach. Moreover, the box and whisker plots were utilized to display the results. When the return periods and durations increased, the uncertainty slightly increased. The historical IDF curves of London site falls within the regional simulated IDF curves. Furthermore, 1000 runs have been generated by using the weather generator.

This thesis gives formal definitions of discourse-givenness, coreference and reference, and reports on experiments with computational models of discourse-givenness of noun phrases for English and German. Definitions are based on Bach's (1987) work on reference, Kibble and van Deemter's (2000) work on coreference, and Kamp and Reyle's Discourse Representation Theory (1993). For the experiments, the following corpora with coreference annotation were used: MUC-7, OntoNotes and ARRAU for Englisch, and TueBa-D/Z for German. As for classification algorithms, they cover J48 decision trees, the rule based learner Ripper, and linear support vector machines. New features are suggested, representing the noun phrase's specificity as well as its context, which lead to a significant improvement of classification quality. / Die vorliegende Arbeit gibt formale Definitionen der Konzepte Diskursgegebenheit, Koreferenz und Referenz. Zudem wird über Experimente berichtet, Nominalphrasen im Deutschen und Englischen hinsichtlich ihrer Diskursgegebenheit zu klassifizieren. Die Definitionen basieren auf Arbeiten von Bach (1987) zu Referenz, Kibble und van Deemter (2000) zu Koreferenz und der Diskursrepräsentationstheorie (Kamp und Reyle, 1993). In den Experimenten wurden die koreferenzannotierten Korpora MUC-7, OntoNotes und ARRAU (Englisch) und TüBa-D/Z (Deutsch) verwendet. Sie umfassen die Klassifikationsalgorithmen J48 (Entscheidungsbäume), Ripper (regelbasiertes Lernen) und lineare Support Vector Machines. Mehrere neue Klassifikationsmerkmale werden vorgeschlagen, die die Spezifizität der Nominalphrase messen, sowie ihren Kontext abbilden. Mit Hilfe dieser Merkmale kann eine signifikante Verbesserung der Klassifikation erreicht werden.

Submitted by Marina Alves Ferraz null (marinaaferraz@yahoo.com.br) on 2018-03-04T14:57:52Z No. of bitstreams: 1 Pedrini.M.AF.dissertação_final.pdf: 6365438 bytes, checksum: d72d8b494e5dd313a3dd15d4a45ba79a (MD5) / Approved for entry into archive by Maria Marlene Zaniboni null (zaniboni@bauru.unesp.br) on 2018-03-05T15:59:19Z (GMT) No. of bitstreams: 1 pedrini_maf_me_bauru.pdf: 6365438 bytes, checksum: d72d8b494e5dd313a3dd15d4a45ba79a (MD5) / Made available in DSpace on 2018-03-05T15:59:19Z (GMT). No. of bitstreams: 1 pedrini_maf_me_bauru.pdf: 6365438 bytes, checksum: d72d8b494e5dd313a3dd15d4a45ba79a (MD5) Previous issue date: 2018-02-01 / O debate sobre o aquecimento global nas últimas décadas e também o aumento da frequência e intensidade de acontecimentos extremos causados por eventos hidro-meteorológicos e climatológicos levou a uma maior ênfase em estudos de desastres naturais. No Brasil as secas e as enxurradas são as tipologias mais recorrentes, sendo que a chuva é o evento desencadeador com maior incidência de danos. A análise cruzada de danos e seus fatores geradores relacionados a eventos hidrológicos indica que a chuva intensa é o evento mais preponderante para a ocorrência de desastres naturais relacionados à drenagem. O objetivo do trabalho foi a elaboração do mapa de risco à inundação urbana (alagamento, enchente e enxurrada) da cidade de Bauru e também a comparação entre dados obtidos de pluviômetros automáticos e dados de radar. Os dados de precipitação foram obtidos de pluviômetros automáticos instalados na cidade de Bauru. A probabilidade de ocorrência dos eventos hidrológicos foi estimada pelas equações IDF (intensidade, duração, frequência). Os resultados desta pesquisa foram espacializados e classificados em função de sua severidade verificando-se que eventos com baixo período de retorno geram danos significativos. Foi elaborado um mapa de perigo atribuindo pesos aos fatores ambientais e sociais com o auxílio de SIG (Sistema de Informação Geográfica). O estudo de caso desta pesquisa pode auxiliar os órgãos competentes na implementação de sistemas de alerta precoce e políticas de prevenção. / The current debate about global warming in recent decades and also the increase in the frequency and intensity of extreme events caused by hydro-meteorological and climatological events led to a greater emphasis on natural disasters studies. In Brazil, droughts and floods are the most recurrent typologies, with rain being the main event that causes the majority of the damages. Cross-analysis of damage triggering events related to rain indicates that heavy rain is the most significant event in the occurrence of natural disasters related to drainage. The objective of this research was the elaboration of an urban flooding risk map for Bauru and also the comparison between data of rain gauges and radar. Precipitation data was obtained from rain gauges installed in the city of Bauru. The probability of occurrence of hydrological events was estimated by the IDF (intensity, duration, frequency) equations. The results of this research were spatialized and classified according to its severity and they showed low return events cause significant damage. A map of hazard areas was developed with the help of GIS (Geographic Information System) software. The case study of this research can help the local governments with the implementation of early warning systems and prevention policies.

Detecting topics by extracting keywords from written text using TF-IDF has been studied and successfully used in many applications. Adding a semantic layer to TF-IDF-based topic detection using WordNet synonyms and hypernyms has been explored in document clustering by assigning concepts that describe texts or by adding all synonyms and hypernyms that occurring words have to a list of keywords. A new method where TF-IDF scores are calculated and WordNet synset members’ TF-IDFscores are added together to all occurring synonyms and/or hypernyms is explored in this paper. Here, such an approach is evaluated by comparing extracted keywords using TF-IDF and the new proposed method, SynPlusTF-IDF, against manually assigned keywords in a database of scientific abstracts. As topic detection is widely used in many contexts and applications, improving current methods is of great value as the methods can become more accurate at extracting correct and relevant keywords from written text. An experiment was conducted comparing the two methods and their accuracy measured using precision and recall and by calculating F1-scores.The F1-scores ranged from 0.11131 to 0.14264 for different variables and the results show that SynPlusTF-IDF is not better at topic detection compared to TF-IDF and both methods performed poorly at topic detection with the chosen dataset.

In this thesis, we present a new clustering algorithm we call <em>Significance Feature Clustering</em>, which is designed to cluster text documents. Its central premise is the mapping of raw frequency count vectors to discrete-valued significance vectors which contain values of -1, 0, or 1. These values represent whether a word is <em>significantly negative</em>, <em>neutral</em>, or <em>significantly positive</em>, respectively. Initially, standard tf-idf vectors are computed from raw frequency vectors, then these tf-idf vectors are transformed to significance vectors using a parameter alpha, where alpha controls the mapping -1, 0, or 1 for each vector entry. SFC clusters agglomeratively, with each document's significance vector representing a cluster of size one containing just the document, and iteratively merges the two clusters that exhibit the most similar average using cosine similarity. We show that by using a good alpha value, the significance vectors produced by SFC provide an accurate indication of which words are significant to which documents, as well as the type of significance, and therefore correspondingly yield a good clustering in terms of a well-known definition of clustering quality. We further demonstrate that a user need not manually select an alpha as we develop a new definition of clustering quality that is highly correlated with text clustering quality. Our metric extends the family of metrics known as <em>internal similarity</em>, so that it can be applied to a tree of clusters rather than a set, but it also factors in an aspect of recall that was absent from previous internal similarity metrics. Using this new definition of internal similarity, which we call <em>maximum tree internal similarity</em>, we show that a close to optimal text clustering may be picked from any number of clusterings created by different alpha's. The automatically selected clusterings have qualities that are close to that of a well-known and powerful hierarchical clustering algorithm.

In this thesis, we present a new clustering algorithm we call <em>Significance Feature Clustering</em>, which is designed to cluster text documents. Its central premise is the mapping of raw frequency count vectors to discrete-valued significance vectors which contain values of -1, 0, or 1. These values represent whether a word is <em>significantly negative</em>, <em>neutral</em>, or <em>significantly positive</em>, respectively. Initially, standard tf-idf vectors are computed from raw frequency vectors, then these tf-idf vectors are transformed to significance vectors using a parameter alpha, where alpha controls the mapping -1, 0, or 1 for each vector entry. SFC clusters agglomeratively, with each document's significance vector representing a cluster of size one containing just the document, and iteratively merges the two clusters that exhibit the most similar average using cosine similarity. We show that by using a good alpha value, the significance vectors produced by SFC provide an accurate indication of which words are significant to which documents, as well as the type of significance, and therefore correspondingly yield a good clustering in terms of a well-known definition of clustering quality. We further demonstrate that a user need not manually select an alpha as we develop a new definition of clustering quality that is highly correlated with text clustering quality. Our metric extends the family of metrics known as <em>internal similarity</em>, so that it can be applied to a tree of clusters rather than a set, but it also factors in an aspect of recall that was absent from previous internal similarity metrics. Using this new definition of internal similarity, which we call <em>maximum tree internal similarity</em>, we show that a close to optimal text clustering may be picked from any number of clusterings created by different alpha's. The automatically selected clusterings have qualities that are close to that of a well-known and powerful hierarchical clustering algorithm.