Vertex component analysis: a fast algorithm to extract endmembers spectra from hyperspectral data

Chapter in Book Proceedings with Peer Review First Iberian Conference, IbPRIA 2003, Puerto de Andratx, Mallorca, Spain, JUne 4-6, 2003. Proceedings

Vertex component analysis: a fast algorithm to unmix hyperspectral data

Given a set of mixed spectral (multispectral or hyperspectral) vectors, linear spectral mixture analysis, or linear unmixing, aims at estimating the number of reference substances, also called endmembers, their spectral signatures, and their abundance fractions. This paper presents a new method for unsupervised endmember extraction from hyperspectral data, termed vertex component analysis (VCA). The algorithm exploits two facts: (1) the endmembers are the vertices of a simplex and (2) the affine transformation of a simplex is also a simplex. In a series of experiments using simulated and real data, the VCA algorithm competes with state-of-the-art methods, with a computational complexity between one and two orders of magnitude lower than the best available method.

Exploração de dados de uma solução mobile wallet usando as técnicas de data mining

A descoberta de conhecimento em dados hoje em dia é um ponto forte para as empresas. Atualmente a CardMobili não dispõe de qualquer sistema de mineração de dados, sendo a existência deste uma mais-valia para as suas operações de marketing diárias, nomeadamente no lançamento de cupões a um grupo restrito de clientes com uma elevada probabilidade que os mesmos os utilizem. Para isso foi analisada a base de dados da aplicação tentando extrair o maior número de dados e aplicadas as transformações necessárias para posteriormente serem processados pelos algoritmos de mineração de dados. Durante a etapa de mineração de dados foram aplicadas as técnicas de associação e classificação, sendo que os melhores resultados foram obtidos com técnicas de associação. Desta maneira pretende-se que os resultados obtidos auxiliem o decisor na sua tomada de decisões.

Fatores determinantes da maturidade e composição da dívida das empresas portuguesas cotadas

Mestrado em Contabilidade e análise financeira

Three-dimensional modeling of tongue during speech using MRI data

The tongue is the most important and dynamic articulator for speech formation, because of its anatomic aspects (particularly, the large volume of this muscular organ comparatively to the surrounding organs of the vocal tract) and also due to the wide range of movements and flexibility that are involved. In speech communication research, a variety of techniques have been used for measuring the three-dimensional vocal tract shapes. More recently, magnetic resonance imaging (MRI) becomes common; mainly, because this technique allows the collection of a set of static and dynamic images that can represent the entire vocal tract along any orientation. Over the years, different anatomical organs of the vocal tract have been modelled; namely, 2D and 3D tongue models, using parametric or statistical modelling procedures. Our aims are to present and describe some 3D reconstructed models from MRI data, for one subject uttering sustained articulations of some typical Portuguese sounds. Thus, we present a 3D database of the tongue obtained by stack combinations with the subject articulating Portuguese vowels. This 3D knowledge of the speech organs could be very important; especially, for clinical purposes (for example, for the assessment of articulatory impairments followed by tongue surgery in speech rehabilitation), and also for a better understanding of acoustic theory in speech formation.

Association between grip strength, anthropometric data and functional capacity

The hand is one of the most important instruments of the human body, mainly due to the possibility of grip movements. Grip strength has been described as an important predictor of functional capacity. There are several factors that may influence it, such as gender, age and anthropometric characteristics. Functional capacity refers to the ability to perform daily activities which allow the individual to self-care and to live with autonomy. Composite Physical Function (CPF) scale is an evaluation tool for functional capacity that includes daily activities, self-care, sports activities, upper limb function and gait capacity. In 2011, Portugal had 15% of young population (0-14years) and 19% of elderly population (over 65 years). Considering the double-ageing phenomen, it is important to understand the effect of the grip strength in elderly individuals, considering their characteristics, as the need to maintainin dependency as long as possible.

Modelação do transporte e dispersão de um marcador corado no Rio Febros

O rio Febros é um pequeno curso de água, situado no concelho de Vila Nova de Gaia, com cerca de 15 km de extensão, cuja bacia hidrográfica ocupa uma área de aproximadamente 35,4 km2. Nasce em Seixezelo e desagua na margem esquerda do Rio Douro no Cais do Esteiro, em Avintes. Em Maio de 2008, um acidente de viação teve como consequência o derrame de cerca de quatro toneladas de ácido clorídrico que rapidamente convergiu às águas do rio. Apenas um dia depois, o pH desceu para três e muitos foram os peixes que morreram. A solução adoptada para evitar o desaire foi introduzir milhares de litros de água de modo a diluir o ácido presente, ao longo de todo o curso de água. Tal facto não evitou a destruição de parte de um ecossistema, que ainda nos dias de hoje se encontra em recuperação. De forma a avaliar-se o impacto destas possíveis perturbações sejam estas de origem antropogénica ou natural é necessário possuir conhecimentos dos processos químicos tais como a advecção, a mistura devida à dispersão e a transferência de massa ar/água. Estes processos irão determinar o movimento e destino das substâncias que poderão ser descarregadas no rio. Para tal, recorrer-se-á ao estudo hidrogeométrico do curso de água assim como ao estudo do comportamento de um marcador, simulando uma possível descarga. A rodamina WT será o marcador a ser utilizado devido à panóplia de características ambientalmente favoráveis. Os estudos de campo com este corante, realizados em sequência de descarga previamente estudada, fornecem uma das melhores fontes de informação para verificação e validação de modelos hidráulicos utilizados em estudos de qualidade de águas e protecção ambiental. Escolheram-se dois pontos de descarga no Febros, um em Casal Drijo e outro no Parque Biológico de Gaia, possuindo cada um deles, a jusante, duas estações de monitorização. Pelo modelo ADE os valores obtidos para o coeficiente de dispersão longitudinal para as estações Pontão d’ Alheira, Pinheiral, Menesas e Giestas foram, respectivamente, 0,3622; 0,5468; 1,6832 e 1,7504 m2/s. Para a mesma sequência de estações, os valores da velocidade de escoamento obtidos neste trabalho experimental foram de 0,0633; 0,0684; 0,1548 e 0,1645 m/s. Quanto ao modelo TS, os valores obtidos para o coeficiente de dispersão longitudinal para as estações Pontão d’ Alheira, Pinheiral, Menesas e Giestas foram, respectivamente, 0,2339; 0,1618; 0,5057e 1,1320 m2/s. Para a mesma sequência de estações, os valores da velocidade de escoamento obtidos neste trabalho experimental foram de 0,0652; 0,0775; 0,1891 e 0,1676 m/s. Os resultados foram ajustados por um método directo, o método dos momentos, e por dois métodos indirectos, os modelos ADE e TS. O melhor ajuste corresponde ao modelo TS onde os valores do coeficiente de dispersão longitudinal e da velocidade de escoamento são aqueles que melhor se aproximam da realidade. Quanto ao método dos momentos, o valor estimado para a velocidade é de 0,162 m/s e para o coeficiente de dispersão longitudinal de 9,769 m2/s. Não obstante, a compreensão da hidrodinâmica do rio e das suas características, bem como a adequação de modelos matemáticos no tratamento de resultados formam uma estratégia de protecção ambiental inerente a futuros impactos que possam suceder.

Contributo para a estimativa da idade fetal à data da morte na população portuguesa

A estimativa da idade gestacional (IG) em restos cadavéricos fetais é importante em contextos forenses. Para esse efeito, os especialistas forenses recorrem à avaliação do padrão de calcificação dentária e/ou ao estudo do esqueleto. Neste último, o comprimento das diáfises de ossos longos é um dos métodos mais utilizados, sendo utilizadas equações de regressão de obras pouco atuais ou baseadas em dados ecográficos, cujas medições diferem das efetuadas diretamente no osso. Este trabalho tem como objetivo principal a obtenção de equações de regressão para a população Portuguesa, com base na medição das diáfises de fémur, tíbia e úmero, utilizando radiografias postmortem. A amostra é constituída por 80 fetos de IG conhecida. Tratando-se de um estudo retrospectivo, os casos foram selecionados com base nas informações clínicas e anatomopatológicas, excluindo-se aqueles cujo normal crescimento se encontrava efetiva ou potencialmente comprometido. Os resultados confirmaram uma forte correlação entre o comprimento das diáfises estudadas e a IG, apresentando o fémur a correlação mais forte (r=0.967; p <0,01). Assim, foi possível obter uma equação de regressão para cada um dos ossos estudados. Concluindo, os objetivos do estudo foram atingidos com a obtenção das equações de regressão para os ossos estudados. Pretende-se, futuramente, alargar a amostra para validar e consolidar os resultados obtidos neste estudo.

Assessment of data-driven modeling strategies for water delivery canals

The aim of this paper is to develop models for experimental open-channel water delivery systems and assess the use of three data-driven modeling tools toward that end. Water delivery canals are nonlinear dynamical systems and thus should be modeled to meet given operational requirements while capturing all relevant dynamics, including transport delays. Typically, the derivation of first principle models for open-channel systems is based on the use of Saint-Venant equations for shallow water, which is a time-consuming task and demands for specific expertise. The present paper proposes and assesses the use of three data-driven modeling tools: artificial neural networks, composite local linear models and fuzzy systems. The canal from Hydraulics and Canal Control Nucleus (A parts per thousand vora University, Portugal) will be used as a benchmark: The models are identified using data collected from the experimental facility, and then their performances are assessed based on suitable validation criterion. The performance of all models is compared among each other and against the experimental data to show the effectiveness of such tools to capture all significant dynamics within the canal system and, therefore, provide accurate nonlinear models that can be used for simulation or control. The models are available upon request to the authors.

Analysis and visualization of seismic data using mutual information

Seismic data is difficult to analyze and classical mathematical tools reveal strong limitations in exposing hidden relationships between earthquakes. In this paper, we study earthquake phenomena in the perspective of complex systems. Global seismic data, covering the period from 1962 up to 2011 is analyzed. The events, characterized by their magnitude, geographic location and time of occurrence, are divided into groups, either according to the Flinn-Engdahl (F-E) seismic regions of Earth or using a rectangular grid based in latitude and longitude coordinates. Two methods of analysis are considered and compared in this study. In a first method, the distributions of magnitudes are approximated by Gutenberg-Richter (G-R) distributions and the parameters used to reveal the relationships among regions. In the second method, the mutual information is calculated and adopted as a measure of similarity between regions. In both cases, using clustering analysis, visualization maps are generated, providing an intuitive and useful representation of the complex relationships that are present among seismic data. Such relationships might not be perceived on classical geographic maps. Therefore, the generated charts are a valid alternative to other visualization tools, for understanding the global behavior of earthquakes.

Feature selection for clustering categorical data with an embedded modelling approach

Research on the problem of feature selection for clustering continues to develop. This is a challenging task, mainly due to the absence of class labels to guide the search for relevant features. Categorical feature selection for clustering has rarely been addressed in the literature, with most of the proposed approaches having focused on numerical data. In this work, we propose an approach to simultaneously cluster categorical data and select a subset of relevant features. Our approach is based on a modification of a finite mixture model (of multinomial distributions), where a set of latent variables indicate the relevance of each feature. To estimate the model parameters, we implement a variant of the expectation-maximization algorithm that simultaneously selects the subset of relevant features, using a minimum message length criterion. The proposed approach compares favourably with two baseline methods: a filter based on an entropy measure and a wrapper based on mutual information. The results obtained on synthetic data illustrate the ability of the proposed expectation-maximization method to recover ground truth. An application to real data, referred to official statistics, shows its usefulness.

Categorical data clustering using a minimum message length criterion

Research on cluster analysis for categorical data continues to develop, new clustering algorithms being proposed. However, in this context, the determination of the number of clusters is rarely addressed. We propose a new approach in which clustering and the estimation of the number of clusters is done simultaneously for categorical data. We assume that the data originate from a finite mixture of multinomial distributions and use a minimum message length criterion (MML) to select the number of clusters (Wallace and Bolton, 1986). For this purpose, we implement an EM-type algorithm (Silvestre et al., 2008) based on the (Figueiredo and Jain, 2002) approach. The novelty of the approach rests on the integration of the model estimation and selection of the number of clusters in a single algorithm, rather than selecting this number based on a set of pre-estimated candidate models. The performance of our approach is compared with the use of Bayesian Information Criterion (BIC) (Schwarz, 1978) and Integrated Completed Likelihood (ICL) (Biernacki et al., 2000) using synthetic data. The obtained results illustrate the capacity of the proposed algorithm to attain the true number of cluster while outperforming BIC and ICL since it is faster, which is especially relevant when dealing with large data sets.

Determining the number of clusters in categorical data

Cluster analysis for categorical data has been an active area of research. A well-known problem in this area is the determination of the number of clusters, which is unknown and must be inferred from the data. In order to estimate the number of clusters, one often resorts to information criteria, such as BIC (Bayesian information criterion), MML (minimum message length, proposed by Wallace and Boulton, 1968), and ICL (integrated classification likelihood). In this work, we adopt the approach developed by Figueiredo and Jain (2002) for clustering continuous data. They use an MML criterion to select the number of clusters and a variant of the EM algorithm to estimate the model parameters. This EM variant seamlessly integrates model estimation and selection in a single algorithm. For clustering categorical data, we assume a finite mixture of multinomial distributions and implement a new EM algorithm, following a previous version (Silvestre et al., 2008). Results obtained with synthetic datasets are encouraging. The main advantage of the proposed approach, when compared to the above referred criteria, is the speed of execution, which is especially relevant when dealing with large data sets.

Disseminating data using broadcast when topology is unknown

Consider the problem of disseminating data from an arbitrary source node to all other nodes in a distributed computer system, like Wireless Sensor Networks (WSNs). We assume that wireless broadcast is used and nodes do not know the topology. We propose new protocols which disseminate data faster and use fewer broadcasts than the simple broadcast protocol.

Transfering data from a server to an Android Mobile application: a case study

Nowadays, due to the incredible grow of the mobile devices market, when we want to implement a client-server applications we must consider mobile devices limitations. In this paper we discuss which can be the more reliable and fast way to exchange information between a server and an Android mobile application. This is an important issue because with a responsive application the user experience is more enjoyable. In this paper we present a study that test and evaluate two data transfer protocols, socket and HTTP, and three data serialization formats (XML, JSON and Protocol Buffers) using different environments and mobile devices to realize which is the most practical and fast to use.