Temporal patterns of TV watching for Portuguese viewers

Audiometer systems provide enormous amounts of detailed TV watching data. Several relevant and interdependent factors may influence TV viewers' behavior. In this work we focus on the time factor and derive Temporal Patterns of TV watching, based on panel data. Clustering base attributes are originated from 1440 binary minute-related attributes, capturing the TV watching status (watch/not watch). Since there are around 2500 panel viewers a data reduction procedure is first performed. K-Means algorithm is used to obtain daily clusters of viewers. Weekly patterns are then derived which rely on daily patterns. The obtained solutions are tested for consistency and stability. Temporal TV watching patterns provide new insights concerning Portuguese TV viewers' behavior.

User modeling and recommendation strategies for tourism

Mestrado em Engenharia Inform��tica

A similarity measure for music signals

One of the goals in the field of Music Information Retrieval is to obtain a measure of similarity between two musical recordings. Such a measure is at the core of automatic classification, query, and retrieval systems, which have become a necessity due to the ever increasing availability and size of musical databases. This paper proposes a method for calculating a similarity distance between two music signals. The method extracts a set of features from the audio recordings, models the features, and determines the distance between models. While further work is needed, preliminary results show that the proposed method has the potential to be used as a similarity measure for musical signals.

Decisions with multiple simultaneous goals and uncertain causal effects

A key aspect of decision-making in a disaster response scenario is the capability to evaluate multiple and simultaneously perceived goals. Current competing approaches to build decision-making agents are either mental-state based as BDI, or founded on decision-theoretic models as MDP. The BDI chooses heuristically among several goals and the MDP searches for a policy to achieve a specific goal. In this paper we develop a preferences model to decide among multiple simultaneous goals. We propose a pattern, which follows a decision-theoretic approach, to evaluate the expected causal effects of the observable and non-observable aspects that inform each decision. We focus on yes-or-no (i.e., pursue or ignore a goal) decisions and illustrate the proposal using the RoboCupRescue simulation environment.

The usefulness of ultrasound in the classification of chronic liver disease

Chronic Liver Disease is a progressive, most of the time asymptomatic, and potentially fatal disease. In this paper, a semi-automatic procedure to stage this disease is proposed based on ultrasound liver images, clinical and laboratorial data. In the core of the algorithm two classifiers are used: a k nearest neighbor and a Support Vector Machine, with different kernels. The classifiers were trained with the proposed multi-modal feature set and the results obtained were compared with the laboratorial and clinical feature set. The results showed that using ultrasound based features, in association with laboratorial and clinical features, improve the classification accuracy. The support vector machine, polynomial kernel, outperformed the others classifiers in every class studied. For the Normal class we achieved 100% accuracy, for the chronic hepatitis with cirrhosis 73.08%, for compensated cirrhosis 59.26% and for decompensated cirrhosis 91.67%.

Diffuse liver disease classification from ultrasound surface characterization, clinical and laboratorial data

In this work liver contour is semi-automatically segmented and quantified in order to help the identification and diagnosis of diffuse liver disease. The features extracted from the liver contour are jointly used with clinical and laboratorial data in the staging process. The classification results of a support vector machine, a Bayesian and a k-nearest neighbor classifier are compared. A population of 88 patients at five different stages of diffuse liver disease and a leave-one-out cross-validation strategy are used in the classification process. The best results are obtained using the k-nearest neighbor classifier, with an overall accuracy of 80.68%. The good performance of the proposed method shows a reliable indicator that can improve the information in the staging of diffuse liver disease.

Fatty liver characterization and classification by ultrasound

Steatosis, also known as fatty liver, corresponds to an abnormal retention of lipids within the hepatic cells and reflects an impairment of the normal processes of synthesis and elimination of fat. Several causes may lead to this condition, namely obesity, diabetes, or alcoholism. In this paper an automatic classification algorithm is proposed for the diagnosis of the liver steatosis from ultrasound images. The features are selected in order to catch the same characteristics used by the physicians in the diagnosis of the disease based on visual inspection of the ultrasound images. The algorithm, designed in a Bayesian framework, computes two images: i) a despeckled one, containing the anatomic and echogenic information of the liver, and ii) an image containing only the speckle used to compute the textural features. These images are computed from the estimated RF signal generated by the ultrasound probe where the dynamic range compression performed by the equipment is taken into account. A Bayes classifier, trained with data manually classified by expert clinicians and used as ground truth, reaches an overall accuracy of 95% and a 100% of sensitivity. The main novelties of the method are the estimations of the RF and speckle images which make it possible to accurately compute textural features of the liver parenchyma relevant for the diagnosis.

eTourismGUI - Adaptative Interface

Mestrado em Engenharia Inform��tica. ��rea de Especializa����o em Tecnologias do Conhecimento e Decis��o.

Sobreviv��ncia: jogo interactivo para estudo de arquitecturas de agentes aut��nomos

Trabalho de projeto para obten����o do grau de Mestre em Engenharia Inform��tica e de Computadores

Biblioteca de navega����o aut��noma baseada em campos de potencial

Trabalho de Final de Mestrado para obten����o do grau de Mestre em Engenharia Inform��tica e de Computadores

Adaptive learning in agents behaviour: a framework for electricity markets simulation

Electricity markets are complex environments with very particular characteristics. A critical issue regarding these specific characteristics concerns the constant changes they are subject to. This is a result of the electricity markets��� restructuring, which was performed so that the competitiveness could be increased, but it also had exponential implications in the increase of the complexity and unpredictability in those markets scope. The constant growth in markets unpredictability resulted in an amplified need for market intervenient entities in foreseeing market behaviour. The need for understanding the market mechanisms and how the involved players��� interaction affects the outcomes of the markets, contributed to the growth of usage of simulation tools. Multi-agent based software is particularly well fitted to analyze dynamic and adaptive systems with complex interactions among its constituents, such as electricity markets. This dissertation presents ALBidS ��� Adaptive Learning strategic Bidding System, a multiagent system created to provide decision support to market negotiating players. This system is integrated with the MASCEM electricity market simulator, so that its advantage in supporting a market player can be tested using cases based on real markets��� data. ALBidS considers several different methodologies based on very distinct approaches, to provide alternative suggestions of which are the best actions for the supported player to perform. The approach chosen as the players��� actual action is selected by the employment of reinforcement learning algorithms, which for each different situation, simulation circumstances and context, decides which proposed action is the one with higher possibility of achieving the most success. Some of the considered approaches are supported by a mechanism that creates profiles of competitor players. These profiles are built accordingly to their observed past actions and reactions when faced with specific situations, such as success and failure. The system���s context awareness and simulation circumstances analysis, both in terms of results performance and execution time adaptation, are complementary mechanisms, which endow ALBidS with further adaptation and learning capabilities.

Clustering and selecting categorical features

In data clustering, the problem of selecting the subset of most relevant features from the data has been an active research topic. Feature selection for clustering is a challenging task due to the absence of class labels for guiding the search for relevant features. Most methods proposed for this goal are focused on numerical data. In this work, we propose an approach for clustering and selecting categorical features simultaneously. We assume that the data originate from a finite mixture of multinomial distributions and implement an integrated expectation-maximization (EM) algorithm that estimates all the parameters of the model and selects the subset of relevant features simultaneously. The results obtained on synthetic data illustrate the performance of the proposed approach. An application to real data, referred to official statistics, shows its usefulness.

Sistema pericial para suporte �� decis��o na aloca����o de recursos de emerg��ncia

O panorama atual da emerg��ncia e socorro de primeira linha em Portugal, carateriza-se por uma grande aposta ao longo dos ��ltimos anos num incremento cont��nuo da qualidade e da efici��ncia que estes servi��os prestam ��s popula����es locais. Com vista �� prossecu����o do objetivo de melhoria cont��nua dos servi��os, foram realizados ao longo dos ��ltimos anos investimentos avultados ao n��vel dos recursos t��cnicos e ao n��vel da contrata����o e forma����o de recursos humanos altamente qualificados. Atualmente as institui����es que prestam socorro e emerg��ncia de primeira linha est��o bem dotadas ao n��vel f��sico e ao n��vel humano dos recursos necess��rios para fazerem face aos mais diversos tipos de ocorr��ncias. Contudo, ao n��vel dos sistemas de informa����o de apoio �� emerg��ncia e socorro de primeira linha, verifica-se uma inadequa����o (e por vezes inexist��ncia) de sistemas inform��ticos capazes de suportar convenientemente o atual contexto de exig��ncia e complexidade da emerg��ncia e socorro. Foi feita ao longo dos ��ltimos anos, uma forte aposta na melhoria dos recursos f��sicos e dos recursos humanos encarregues da resposta ��semerg��ncia de primeira linha, mas descurou-se a ��rea da gest��o e an��lise da informa����o sobre as ocorr��ncias, assim como, o delinear de poss��veis estrat��gias de preven����o que uma an��lise sistematizada da informa����o sobre as ocorr��ncias possibilita. Nas institui����es de emerg��ncia e socorro de primeira linha em Portugal (bombeiros, prote����o civil municipal, PSP, GNR, pol��cia municipal), prevalecem ainda hoje os sistemas inform��ticos apenas para o registo das ocorr��ncias �� posteriori e a total inexist��ncia de sistemas de registo de informa����o e de apoio �� decis��o na aloca����o de recursos que operem em tempo real. A generalidade dos sistemas inform��ticos atualmente existentes nas institui����es s��o unicamente de sistemas de backoffice, que n��o aproveitam a todas as potencialidades da informa����o operacional neles armazenada. Verificou-se tamb��m, que a geo-localiza����o por via inform��tica dos recursos f��sicos e de pontos de interesse relevantes em situa����es cr��ticas �� inexistente a este n��vel. Neste contexto, consideramos ser poss��vel e importante alinhar o n��vel dos sistemas inform��ticos das institui����es encarregues da emerg��ncia e socorro de primeira linha, com o n��vel dos recursos f��sicos e humanos que j�� disp��em atualmente. Dado que a emerg��ncia e socorro de primeira linha �� um dom��nio claramente eleg��vel para a aplica����o de tecnologias provenientes dos dom��nios da intelig��ncia artificial (nomeadamente sistemas periciais para apoio �� decis��o) e da geo-localiza����o, decidimos no ��mbito desta tese desenvolver um sistema inform��tico capaz de colmatar muitas das lacunas por n��s identificadas ao n��vel dos sistemas inform��ticos destas institui����es. Pretendemos colocar as suas plataformas inform��ticas num n��vel similar ao dos seus recursos f��sicos e humanos. Assim, foram por n��s identificadas duas ��reas chave onde a implementa����o de sistemas inform��ticos adequados ��s reais necessidades das institui����es podem ter um impacto muito proporcionar uma melhor gest��o e otimiza����o dos recursos f��sicos e humanos. As duas ��reas chave por n��s identificadas s��o o suporte �� decis��o na aloca����o dos recursos f��sicos e a geolocaliza����o dos recursos f��sicos, das ocorr��ncias e dos pontos de interesse. Procurando fornecer uma resposta v��lida e adequada a estas duas necessidades prementes, foi desenvolvido no ��mbito desta tese o sistema CRITICAL DECISIONS. O sistema CRITICAL DECISIONS incorpora um conjunto de funcionalidades t��picas de um sistema pericial, para o apoio na decis��o de aloca����o de recursos f��sicos ��s ocorr��ncias. A infer��ncia autom��tica dos recursos f��sicos, assenta num conjunto de regra de infer��ncia armazenadas numa base de conhecimento, em constante crescimento e atualiza����o, com base nas respostas bem sucedidas a ocorr��ncias passadas. Para suprimir as car��ncias aos n��vel da geo-localiza����o dos recursos f��sicos, das ocorr��ncias e dos pontos de interesse, o sistema CRITICAL DECISIONS incorpora tamb��m um conjunto de funcionalidades de geo-localiza����o. Estas permitem a geo-localiza����o de todos os recursos f��sicos da institui����o, a geo-localiza����o dos locais e as ��reas das v��rias ocorr��ncias, assim como, dos v��rios tipos de pontos de interesse. O sistema CRITICAL DECISIONS visa ainda suprimir um conjunto de outras car��ncias por n��s identificadas, ao n��vel da gest��o documental (planos de emerg��ncia, plantas dos edif��cios) , da comunica����o, da partilha de informa����o entre as institui����es de socorro e emerg��ncia locais, da contabiliza����o dos tempos de servi��o, entre outros. O sistema CRITICAL DECISIONS �� o culminar de um esfor��o colaborativo e cont��nuo com v��rias institui����es, respons��veis pela emerg��ncia e socorro de primeira linha a n��vel local. Esperamos com o sistema CRITICAL DECISIONS, dotar estas institui����es de uma plataforma inform��tica atual, inovadora, evolutiva, com baixos custos de implementa����o e de opera����o, capaz de proporcionar melhorias cont��nuas e significativas ao n��vel da qualidade da resposta ��s ocorr��ncias, das capacidades de preven����o e de uma melhor otimiza����o de todos os tipos de recursos que t��m ao dispor.

Six thinking hats for electricity markets simulation: adapting meeting tools to agent decision

In almost all industrialized countries, the energy sector has suffered a severe restructuring that originated a greater complexity in market players��� interactions. The complexity that these changes brought made way for the creation of decision support tools that facilitate the study and understanding of these markets. MASCEM ��� ���Multiagent Simulator for Competitive Electricity Markets��� arose in this context providing a framework for evaluating new rules, new behaviour, and new participants in deregulated electricity markets. MASCEM uses game theory, machine learning techniques, scenario analysis and optimisation techniques to model market agents and to provide them with decision-support. ALBidS is a multiagent system created to provide decision support to market negotiating players. Fully integrated with MASCEM it considers several different methodologies based on very distinct approaches. The Six Thinking Hats is a powerful technique used to look at decisions from different perspectives. This tool���s goal is to force the thinker to move outside his habitual thinking style. It was developed to be used mainly at meetings in order to ���run better meetings, make faster decisions���. This dissertation presents a study about the applicability of the Six Thinking Hats technique in Decision Support Systems, particularly with the multiagent paradigm like the MASCEM simulator. As such this work���s proposal is of a new agent, a meta-learner based on STH technique that organizes several different ALBidS��� strategies and combines the distinct answers into a single one that, expectedly, out-performs any of them.

An��lise e s��ntese de m��sica atrav��s de sistemas computacionais

Este trabalho, realizado no ��mbito da unidade curricular de Tese/Disserta����o, procura mostrar de que forma a Computa����o Evolucion��ria se pode aplicar no mundo da M��sica. Este ��, de resto, um tema sobejamente aliciante dentro da ��rea da Intelig��ncia Artificial. Come��a-se por apresentar o mundo da M��sica com uma perspetiva cronol��gica da sua hist��ria, dando especial relevo ao estilo musical do Fado de Coimbra. Abordam-se tamb��m os conceitos fundamentais da teoria musical. Relativamente �� Computa����o Evolucion��ria, exp��em-se os elementos associados aos Algoritmos Evolucion��rios e apresentam-se os principais modelos, nomeadamente os Algoritmos Gen��ticos. Ainda no ��mbito da Computa����o Evolucion��ria, foi elaborado um pequeno estudo do ���estado da arte��� da aplica����o da Computa����o Evolucion��ria na M��sica. A implementa����o pr��tica deste trabalho baseia-se numa aplica����o ��� AG Fado ��� que comp��e melodias de Fado de Coimbra, utilizando Algoritmos Gen��ticos. O trabalho foi dividido em duas partes principais: a primeira parte consiste na recolha de informa����es e posterior levantamento de dados estat��sticos sobre o g��nero musical escolhido, nomeadamente fados em tonalidade maior e fados em tonalidade menor; a segunda parte consiste no desenvolvimento da aplica����o, com a conce����o do respetivo algoritmo gen��tico para composi����o de melodias. As melodias obtidas atrav��s da aplica����o desenvolvida s��o bastante aud��veis e boas melodicamente. No entanto, destaca-se o facto de a avalia����o ser efetuada por seres humanos o que implica sensibilidades musicais distintas levando a resultados igualmente distintos.