Spatial dynamics of team sports exposed by Voronoi diagrams

Team sports represent complex systems: players interact continuously during a game, and exhibit intricate patterns of interaction, which can be identified and investigated at both individual and collective levels. We used Voronoi diagrams to identify and investigate the spatial dynamics of players' behavior in Futsal. Using this tool, we examined 19 plays of a sub-phase of a Futsal game played in a reduced area (20 m(2)) from which we extracted the trajectories of all players. Results obtained from a comparative analysis of player's Voronoi area (dominant region) and nearest teammate distance revealed different patterns of interaction between attackers and defenders, both at the level of individual players and teams. We found that, compared to defenders, larger dominant regions were associated with attackers. Furthermore, these regions were more variable in size among players from the same team but, at the player level, the attackers' dominant regions were more regular than those associated with each of the defenders. These findings support a formal description of the dynamic spatial interaction of the players, at least during the particular sub-phase of Futsal investigated. The adopted approach may be extended to other team behaviors where the actions taken at any instant in time by each of the involved agents are associated with the space they occupy at that particular time.

An unsupervised approach to feature discretization and selection

Many learning problems require handling high dimensional datasets with a relatively small number of instances. Learning algorithms are thus confronted with the curse of dimensionality, and need to address it in order to be effective. Examples of these types of data include the bag-of-words representation in text classification problems and gene expression data for tumor detection/classification. Usually, among the high number of features characterizing the instances, many may be irrelevant (or even detrimental) for the learning tasks. It is thus clear that there is a need for adequate techniques for feature representation, reduction, and selection, to improve both the classification accuracy and the memory requirements. In this paper, we propose combined unsupervised feature discretization and feature selection techniques, suitable for medium and high-dimensional datasets. The experimental results on several standard datasets, with both sparse and dense features, show the efficiency of the proposed techniques as well as improvements over previous related techniques.

SAKWeb – Spatial autocorrelation and kriging web, a W3 geocomputation perspective

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

Impact of the Meteorology Conditions on the Particulate Matter Dispersion from an Industrial Complex into Urban and Suburban Areas

The aim of this work was to assess the influence of meteorological conditions on the dispersion of particulate matter from an industrial zone into urban and suburban areas. The particulate matter concentration was related to the most important meteorological variables such as wind direction, velocity and frequency. A coal-fired power plant was considered to be the main emission source with two stacks of 225 m height. A middle point between the two stacks was taken as the centre of two concentric circles with 6 and 20 km radius delimiting the sampling area. About 40 sampling collectors were placed within this area. Meteorological data was obtained from a portable meteorological station placed at approximately 1.7 km to SE from the stacks. Additional data was obtained from the electrical company that runs the coal power plant. These data covers the years from 2006 to the present. A detailed statistical analysis was performed to identify the most frequent meteorological conditions concerning mainly wind speed and direction. This analysis revealed that the most frequent wind blows from Northwest and North and the strongest winds blow from Northwest. Particulate matter deposition was obtained in two sampling campaigns carried out in summer and in spring. For the first campaign the monthly average flux deposition was 1.90 g/m2 and for the second campaign this value was 0.79 g/m2. Wind dispersion occurred predominantly from North to South, away from the nearest residential area, located at about 6 km to Northwest from the stacks. Nevertheless, the higher deposition fluxes occurred in the NW/N and NE/E quadrants. This study was conducted considering only the contribution of particulate matter from coal combustion, however, others sources may be present as well, such as road traffic. Additional chemical analyses and microanalysis are needed to identify the source linkage to flux deposition levels.

ECG biometrics using a dissimilarity space representation

Electrocardiogram (ECG) biometrics are a relatively recent trend in biometric recognition, with at least 13 years of development in peer-reviewed literature. Most of the proposed biometric techniques perform classifi-cation on features extracted from either heartbeats or from ECG based transformed signals. The best representation is yet to be decided. This paper studies an alternative representation, a dissimilarity space, based on the pairwise dissimilarity between templates and subjects' signals. Additionally, this representation can make use of ECG signals sourced from multiple leads. Configurations of three leads will be tested and contrasted with single-lead experiments. Using the same k-NN classifier the results proved superior to those obtained through a similar algorithm which does not employ a dissimilarity representation. The best Authentication EER went as low as 1:53% for a database employing 503 subjects. However, the employment of extra leads did not prove itself advantageous.

Earthquakes in western Iberia: improving the understanding of lithospheric deformation in a slowly deforming region

Mainland Portugal, on the southwestern edge of the European continent, is located directly north of the boundary between the Eurasian and Nubian plates. It lies in a region of slow lithospheric deformation (< 5 mm yr(-1)), which has generated some of the largest earthquakes in Europe, both intraplate (mainland) and interplate (offshore). Some offshore earthquakes are nucleated on old and cold lithospheric mantle, at depths down to 60 km. The seismicity of mainland Portugal and its adjacent offshore has been repeatedly classified as diffuse. In this paper, we analyse the instrumental earthquake catalogue for western Iberia, which covers the period between 1961 and 2013. Between 2010 and 2012, the catalogue was enriched with data from dense broad-band deployments. We show that although the plate boundary south of Portugal is diffuse, in that deformation is accommodated along several distributed faults rather than along one long linear plate boundary, the seismicity itself is not diffuse. Rather, when located using high-quality data, earthquakes collapse into well-defined clusters and lineations. We identify and characterize the most outstanding clusters and lineations of epicentres and correlate them with geophysical and tectonic features (historical seismicity, topography, geologically mapped faults, Moho depth, free-air gravity, magnetic anomalies and geotectonic units). Both onshore and offshore, clusters and lineations of earthquakes are aligned preferentially NNE-SSW and WNW-ESE. Cumulative seismic moment and epicentre density decrease from south to north, with increasing distance from the plate boundary. Only few earthquake lineations coincide with geologically mapped faults. Clusters and lineations that do not match geologically mapped faults may correspond to previously unmapped faults (e.g. blind faults), rheological boundaries or distributed fracturing inside blocks that are more brittle and therefore break more easily than neighbour blocks. The seismicity map of western Iberia presented in this article opens important questions concerning the regional seismotectonics. This work shows that the study of low-magnitude earthquakes using dense seismic deployments is a powerful tool to study lithospheric deformation in slowly deforming regions, such as western Iberia, where high-magnitude earthquakes occur with long recurrence intervals.

Classificação e agregação automática de notícias desportivas

Mestrado em Engenharia Informática - Área de Especialização em Arquiteturas, Sistemas e Redes

Text classification using compression-based dissimilarity measures

Arguably, the most difficult task in text classification is to choose an appropriate set of features that allows machine learning algorithms to provide accurate classification. Most state-of-the-art techniques for this task involve careful feature engineering and a pre-processing stage, which may be too expensive in the emerging context of massive collections of electronic texts. In this paper, we propose efficient methods for text classification based on information-theoretic dissimilarity measures, which are used to define dissimilarity-based representations. These methods dispense with any feature design or engineering, by mapping texts into a feature space using universal dissimilarity measures; in this space, classical classifiers (e.g. nearest neighbor or support vector machines) can then be used. The reported experimental evaluation of the proposed methods, on sentiment polarity analysis and authorship attribution problems, reveals that it approximates, sometimes even outperforms previous state-of-the-art techniques, despite being much simpler, in the sense that they do not require any text pre-processing or feature engineering.

Using data mining techniques to support breast cancer diagnosis

More than ever, there is an increase of the number of decision support methods and computer aided diagnostic systems applied to various areas of medicine. In breast cancer research, many works have been done in order to reduce false-positives when used as a double reading method. In this study, we aimed to present a set of data mining techniques that were applied to approach a decision support system in the area of breast cancer diagnosis. This method is geared to assist clinical practice in identifying mammographic findings such as microcalcifications, masses and even normal tissues, in order to avoid misdiagnosis. In this work a reliable database was used, with 410 images from about 115 patients, containing previous reviews performed by radiologists as microcalcifications, masses and also normal tissue findings. Throughout this work, two feature extraction techniques were used: the gray level co-occurrence matrix and the gray level run length matrix. For classification purposes, we considered various scenarios according to different distinct patterns of injuries and several classifiers in order to distinguish the best performance in each case described. The many classifiers used were Naïve Bayes, Support Vector Machines, k-nearest Neighbors and Decision Trees (J48 and Random Forests). The results in distinguishing mammographic findings revealed great percentages of PPV and very good accuracy values. Furthermore, it also presented other related results of classification of breast density and BI-RADS® scale. The best predictive method found for all tested groups was the Random Forest classifier, and the best performance has been achieved through the distinction of microcalcifications. The conclusions based on the several tested scenarios represent a new perspective in breast cancer diagnosis using data mining techniques.

Sistema de localização indoor com base no protocolo ZIGBEE

Actualmente, os sistemas de localização são uma área em forte expansão sendo que para espaços exteriores existe uma grande variedade de sistemas de localização enquanto que para espaços interiores as soluções são mais escassas. Este trabalho apresenta o estudo e implementação de um sistema de localização indoor baseado no protocolo ZigBee, utilizando a informação da intensidade de sinal recebido (RSSI - Received Signal Strength Indication). Para a realização deste projecto foi necessário iniciar uma pesquisa mais pormenorizada do protocolo ZigBee. O dispositivo móvel a ser localizado é o módulo XBee Serie 2 que se baseia no mesmo protocolo. Posto isto, foi necessário efectuar um estudo sobre sistemas de localização existentes e analisar as técnicas de localização utilizadas para ambientes interiores. Desta forma utiliza-se neste projecto uma técnica que consiste na análise de fingerprinting, onde é criado um mapa com os valores RSSI para diferentes coordenadas do espaço físico. As intensidades de sinal recebido são relativas a dispositivos XBee instalados em pontos fixos de referência. Para calcular a localização do dispositivo móvel é utilizado o algoritmo K-NN (K- Nearest Neighbors) que permite estimar a posição aproximada do dispositivo móvel. Por último é descrito todo o desenvolvimento do projecto assim como a apresentação e discussão de resultados.

Automated image tagging through tag propagation

Trabalho apresentado no âmbito do Mestrado em Engenharia Informática, como requisito parcial Para obtenção do grau de Mestre em Engenharia Informática

Desenvolvimento de sequenciador para um Problema de Roteamento de Veículos

Este projecto tem como objectivo a optimização das rotas dos técnicos de serviço após venda da Schmitt+Sohn Elevadores, associadas à realização das manutenções preventivas a cada elemento contratado à empresa (elevadores, escadas rolantes, etc). Como tal, é necessário fazer uma distribuição dos equipamentos que se encontram em carteira, por um dos técnicos que assegura a manutenção, pelos vários dias úteis de cada mês, e pelas horas de trabalho de cada dia. Apesar do técnico ter disponíveis, por dia, 8h de trabalho, apenas 6h podem ser preenchidas com manutenções preventivas. As 2h restantes são essencialmente para possíveis manutenções correctivas para as quais o técnico seja solicitado. Caso o técnico não seja contactado para resolver nenhuma avaria, essas horas podem ser utilizadas pelo mesmo para adiantar trabalho do dia seguinte, isto é, visitar já alguns dos próximos pontos de manutenção preventiva do dia seguinte, ou para compensar trabalho que esteja atrasado. De salientar que, para cada dia, as deslocações do técnico de qualquer local ao primeiro ponto de uma rota ou de regresso do último ponto de uma rota não são contabilizadas. O trabalho desenvolvido nesta dissertação pretende dar resposta ao problema apresentado pela Schmitt+Sohn Elevadores. Para isso foi desenvolvida uma heurística para a optimização das rotas dos técnicos. Esta é baseada no conceito de “vizinho mais próximo” que procura sempre o ponto que se apresenta mais perto do último ponto que foi adicionado à rota. Com base nesta metodologia, nos processos de escolha dos pontos que formam clusters, e na selecção dos pontos iniciais de cada uma das rotas diárias, a ferramenta de optimização resultante define as rotas diárias para que o percurso efectuado por cada técnico num mês seja o menor possível. São feitas alterações às rotas definidas inicialmente quando encontrados pontos de uma mesma entrada a serem visitados em dias diferentes. Isto obrigaria o técnico a fazer duas viagens ao mesmo local. Por fim, o resultado é apresentado num documento Word a ser utilizado pelo técnico como guia diário das suas deslocações aos equipamentos que necessitam de verificações periódicas. Os resultados obtidos foram comparados com as rotas que estavam a ser usadas pela empresa, tendo apresentado resultados de melhor qualidade, constatando-se a eficiência da solução criada pelo algoritmo proposto neste trabalho.

N170 Asymmetry as an Index of Inferior Occipital Dysfunction in Patients With Symptomatic Occipital Lobe Epilepsy

Objective: Localizing epileptic foci in posterior brain epilepsy remains a difficult exercise in surgery for epilepsy evaluation. Neither clinical manifestations, neurological, EEG nor neuropsychological evaluations provide strong information about the area of onset, and fast spread of paroxysms often produces mixed features of occipital, temporal and parietal symptoms. We investigated the usefulness of the N170 event-related potential to map epileptic activity in these patients. Methods: A group of seven patients with symptomatic posterior cortex epilepsy were submitted to a high-resolution EEG (78 electrodes), with recordings of interictal spikes and face-evoked N170. Generators of spikes and N170 were localized by source analysis. Range of normal N170 asymmetry was determined in 30 healthy volunteers. Results: In 3 out of 7 patients the N170 inter-hemispheric asymmetry was outside control values. Those were the patients whose spike sources were nearest (within 3 cm) to the fusiform gyrus, while foci further away did not affect the N170 ratio. Conclusions: N170 event-related potential provides useful information about focal cortical dysfunction produced by epileptic foci located in the close neighborhood of the fusiform gyrus, but are unaffected by foci further away. Significance: The N170 evoked by faces can improve the epileptic foci localization in posterior brain epilepsy.

Urban sprawl analysis and modeling in Asmara, Eritreia: Application of Geospatial Tools

Dissertation submitted in partial fulfillment of the requirements for the Degree of Master of Science in Geospatial Technologies.

Localização de Precisão Usando Redes Sem Fios Enterradas

As Redes Sem Fios Enterradas (Wireless Underground Networks - WUN) são formadas por nós que comunicam entre si através de ligações sem fios e têm como meio de propagação o solo. Os sistemas de localização mais utilizados atualmente têm desvantagens ao nível da precisão e o custo. Nesta tese é proposta uma solução de localização de precisão que recorre à utilização de redes sem fios enterradas e um algoritmo de posicionamento baseados em Wi-Fi. O objetivo é estimar a localização de objetos, utilizando dispositivos Wi-Fi de baixo custo. Os resultados experimentais obtidos demonstram que o erro de localização é inferior a 0,40 m, e que esta solução é viável para, por exemplo, localizar jogadores num campo de futebol ou localizar um objeto num campo agrícola.