Determinantes do estado de saúde dos Portugueses

Dissertação apresentada como requisito parcial para obtenção do grau de Mestre em Estatística e Gestão de Informação

Análise de Desempenho e Recomendação para Soluções IVR

Mestrado em Engenharia Informática, Área de Especialização em Tecnologias do Conhecimento e da Decisão

A Data-mining-based Methodology to support MV Electricity Customers' Characterization

This paper presents an electricity medium voltage (MV) customer characterization framework supportedby knowledge discovery in database (KDD). The main idea is to identify typical load profiles (TLP) of MVconsumers and to develop a rule set for the automatic classification of new consumers. To achieve ourgoal a methodology is proposed consisting of several steps: data pre-processing; application of severalclustering algorithms to segment the daily load profiles; selection of the best partition, corresponding tothe best consumers’ segmentation, based on the assessments of several clustering validity indices; andfinally, a classification model is built based on the resulting clusters. To validate the proposed framework,a case study which includes a real database of MV consumers is performed.

Coalition of distributed generation units to Virtual Power Players - a game theory approach

Smart Grids (SGs) have emerged as the new paradigm for power system operation and management, being designed to include large amounts of distributed energy resources. This new paradigm requires new Energy Resource Management (ERM) methodologies considering different operation strategies and the existence of new management players such as several types of aggregators. This paper proposes a methodology to facilitate the coalition between distributed generation units originating Virtual Power Players (VPP) considering a game theory approach. The proposed approach consists in the analysis of the classifications that were attributed by each VPP to the distributed generation units, as well as in the analysis of the previous established contracts by each player. The proposed classification model is based in fourteen parameters including technical, economical and behavioural ones. Depending of the VPP strategies, size and goals, each parameter has different importance. VPP can also manage other type of energy resources, like storage units, electric vehicles, demand response programs or even parts of the MV and LV distribution network. A case study with twelve VPPs with different characteristics and one hundred and fifty real distributed generation units is included in the paper.

Variables efectives de l'empowerment

Aquest treball que porta per títol variables efectives de l’empowerment es el treball investigació del programa de doctorat Interuniversitari en Organització i Administració d’Empreses. El treball està composat per tres parts diferenciades. La primera part del treball consisteix amb el comentari de vint articles relacionats amb la motivació, el downsizing i l’empowerment. Els resums exposats han servit per establir els fonaments teòrics previs al model proposat de variables efectives d’empowerment. La segona part consisteix amb l’elaboració d’un article que resumeix les principals fonts consultades i proposa un model de classificació de les variables que poden contribuir a aconseguir amb èxit un procés d’empowerment. Les variables efectives es poden dividir en variables recíproques, variables unidireccionals, variables compartides i variables reflexives. La tercera part i amb l’objectiu de comprovar la validesa de model s’ha desenvolupat un qüestionari per mesurar l’estat de les variables anomenades efectives d’empowerment i la seva contribució amb l’èxit del procés. Es descriu l’eina desenvolupada, el tractament i la representació de les dades obtingudes. Finalment es pot trobar els primers resultats de la prova pilot realitzada per provar el model conceptual proposat.

Prediction of postpartum depression using multilayer perceptrons and pruning.

OBJECTIVE. The main goal of this paper is to obtain a classification model based on feed-forward multilayer perceptrons in order to improve postpartum depression prediction during the 32 weeks after childbirth with a high sensitivity and specificity and to develop a tool to be integrated in a decision support system for clinicians. MATERIALS AND METHODS. Multilayer perceptrons were trained on data from 1397 women who had just given birth, from seven Spanish general hospitals, including clinical, environmental and genetic variables. A prospective cohort study was made just after delivery, at 8 weeks and at 32 weeks after delivery. The models were evaluated with the geometric mean of accuracies using a hold-out strategy. RESULTS. Multilayer perceptrons showed good performance (high sensitivity and specificity) as predictive models for postpartum depression. CONCLUSIONS. The use of these models in a decision support system can be clinically evaluated in future work. The analysis of the models by pruning leads to a qualitative interpretation of the influence of each variable in the interest of clinical protocols.

Intelligent PCA contribution analysis for quality estimation in batch processes. Application in a sequencing batch reactor for wastewater treatment

En aquest treball, es proposa un nou mètode per estimar en temps real la qualitat del producte final en processos per lot. Aquest mètode permet reduir el temps necessari per obtenir els resultats de qualitat de les anàlisi de laboratori. S'utiliza un model de anàlisi de componentes principals (PCA) construït amb dades històriques en condicions normals de funcionament per discernir si un lot finalizat és normal o no. Es calcula una signatura de falla pels lots anormals i es passa a través d'un model de classificació per la seva estimació. L'estudi proposa un mètode per utilitzar la informació de les gràfiques de contribució basat en les signatures de falla, on els indicadors representen el comportament de les variables al llarg del procés en les diferentes etapes. Un conjunt de dades compost per la signatura de falla dels lots anormals històrics es construeix per cercar els patrons i entrenar els models de classifcació per estimar els resultas dels lots futurs. La metodologia proposada s'ha aplicat a un reactor seqüencial per lots (SBR). Diversos algoritmes de classificació es proven per demostrar les possibilitats de la metodologia proposada.

Structural and functional properties of genes involved in human cancer

Background: One of the main goals of cancer genetics is to identify the causative elements at the molecular level leading to cancer.Results: We have conducted an analysis of a set of genes known to be involved in cancer in order to unveil their unique features that can assist towards the identification of new candidate cancer genes. Conclusion: We have detected key patterns in this group of genes in terms of the molecular function or the biological process in which they are involved as well as sequence properties. Based on these features we have developed an accurate Bayesian classification model with which human genes have been scored for their likelihood of involvement in cancer.

Bayes factor for investigative assessment of selected handwriting features

This paper extends previous research [1] on the use of multivariate continuous data in comparative handwriting examinations, notably for gender classification. A database has been constructed by analyzing the contour shape of loop characters of type a and d by means of Fourier analysis, which allows characters to be described in a global way by a set of variables (e.g., Fourier descriptors). Sample handwritings were collected from right- and left-handed female and male writers. The results reported in this paper provide further arguments in support of the view that investigative settings in forensic science represent an area of application for which the Bayesian approach offers a logical framework. In particular, the Bayes factor is computed for settings that focus on inference of gender and handedness of the author of an incriminated handwritten text. An emphasis is placed on comparing the efficiency for investigative purposes of characters a and d.

Etude de la composition initiale et du vieillissement des traces digitales: vers le développement d'une méthode de datation?

«Quel est l'âge de cette trace digitale?» Cette question est relativement souvent soulevée au tribunal ou lors d'investigations, lorsque la personne suspectée admet avoir laissé ses empreintes digitales sur une scène de crime mais prétend l'avoir fait à un autre moment que celui du crime et pour une raison innocente. Toutefois, aucune réponse ne peut actuellement être donnée à cette question, puisqu'aucune méthodologie n'est pour l'heure validée et acceptée par l'ensemble de la communauté forensique. Néanmoins, l'inventaire de cas américains conduit dans cette recherche a montré que les experts fournissent tout de même des témoignages au tribunal concernant l'âge de traces digitales, même si ceux-­‐ci sont majoritairement basés sur des paramètres subjectifs et mal documentés. Il a été relativement aisé d'accéder à des cas américains détaillés, ce qui explique le choix de l'exemple. Toutefois, la problématique de la datation des traces digitales est rencontrée dans le monde entier, et le manque de consensus actuel dans les réponses données souligne la nécessité d'effectuer des études sur le sujet. Le but de la présente recherche est donc d'évaluer la possibilité de développer une méthode de datation objective des traces digitales. Comme les questions entourant la mise au point d'une telle procédure ne sont pas nouvelles, différentes tentatives ont déjà été décrites dans la littérature. Cette recherche les a étudiées de manière critique, et souligne que la plupart des méthodologies reportées souffrent de limitations prévenant leur utilisation pratique. Néanmoins, certaines approches basées sur l'évolution dans le temps de composés intrinsèques aux résidus papillaires se sont montrées prometteuses. Ainsi, un recensement détaillé de la littérature a été conduit afin d'identifier les composés présents dans les traces digitales et les techniques analytiques capables de les détecter. Le choix a été fait de se concentrer sur les composés sébacés détectés par chromatographie gazeuse couplée à la spectrométrie de masse (GC/MS) ou par spectroscopie infrarouge à transformée de Fourier. Des analyses GC/MS ont été menées afin de caractériser la variabilité initiale de lipides cibles au sein des traces digitales d'un même donneur (intra-­‐variabilité) et entre les traces digitales de donneurs différents (inter-­‐variabilité). Ainsi, plusieurs molécules ont été identifiées et quantifiées pour la première fois dans les résidus papillaires. De plus, il a été déterminé que l'intra-­‐variabilité des résidus était significativement plus basse que l'inter-­‐variabilité, mais que ces deux types de variabilité pouvaient être réduits en utilisant différents pré-­‐ traitements statistiques s'inspirant du domaine du profilage de produits stupéfiants. Il a également été possible de proposer un modèle objectif de classification des donneurs permettant de les regrouper dans deux classes principales en se basant sur la composition initiale de leurs traces digitales. Ces classes correspondent à ce qui est actuellement appelé de manière relativement subjective des « bons » ou « mauvais » donneurs. Le potentiel d'un tel modèle est élevé dans le domaine de la recherche en traces digitales, puisqu'il permet de sélectionner des donneurs représentatifs selon les composés d'intérêt. En utilisant la GC/MS et la FTIR, une étude détaillée a été conduite sur les effets de différents facteurs d'influence sur la composition initiale et le vieillissement de molécules lipidiques au sein des traces digitales. Il a ainsi été déterminé que des modèles univariés et multivariés pouvaient être construits pour décrire le vieillissement des composés cibles (transformés en paramètres de vieillissement par pré-­‐traitement), mais que certains facteurs d'influence affectaient ces modèles plus sérieusement que d'autres. En effet, le donneur, le substrat et l'application de techniques de révélation semblent empêcher la construction de modèles reproductibles. Les autres facteurs testés (moment de déposition, pression, température et illumination) influencent également les résidus et leur vieillissement, mais des modèles combinant différentes valeurs de ces facteurs ont tout de même prouvé leur robustesse dans des situations bien définies. De plus, des traces digitales-­‐tests ont été analysées par GC/MS afin d'être datées en utilisant certains des modèles construits. Il s'est avéré que des estimations correctes étaient obtenues pour plus de 60 % des traces-­‐tests datées, et jusqu'à 100% lorsque les conditions de stockage étaient connues. Ces résultats sont intéressants mais il est impératif de conduire des recherches supplémentaires afin d'évaluer les possibilités d'application de ces modèles dans des cas réels. Dans une perspective plus fondamentale, une étude pilote a également été effectuée sur l'utilisation de la spectroscopie infrarouge combinée à l'imagerie chimique (FTIR-­‐CI) afin d'obtenir des informations quant à la composition et au vieillissement des traces digitales. Plus précisément, la capacité de cette technique à mettre en évidence le vieillissement et l'effet de certains facteurs d'influence sur de larges zones de traces digitales a été investiguée. Cette information a ensuite été comparée avec celle obtenue par les spectres FTIR simples. Il en a ainsi résulté que la FTIR-­‐CI était un outil puissant, mais que son utilisation dans l'étude des résidus papillaires à des buts forensiques avait des limites. En effet, dans cette recherche, cette technique n'a pas permis d'obtenir des informations supplémentaires par rapport aux spectres FTIR traditionnels et a également montré des désavantages majeurs, à savoir de longs temps d'analyse et de traitement, particulièrement lorsque de larges zones de traces digitales doivent être couvertes. Finalement, les résultats obtenus dans ce travail ont permis la proposition et discussion d'une approche pragmatique afin d'aborder les questions de datation des traces digitales. Cette approche permet ainsi d'identifier quel type d'information le scientifique serait capable d'apporter aux enquêteurs et/ou au tribunal à l'heure actuelle. De plus, le canevas proposé décrit également les différentes étapes itératives de développement qui devraient être suivies par la recherche afin de parvenir à la validation d'une méthodologie de datation des traces digitales objective, dont les capacités et limites sont connues et documentées. -- "How old is this fingermark?" This question is relatively often raised in trials when suspects admit that they have left their fingermarks on a crime scene but allege that the contact occurred at a time different to that of the crime and for legitimate reasons. However, no answer can be given to this question so far, because no fingermark dating methodology has been validated and accepted by the whole forensic community. Nevertheless, the review of past American cases highlighted that experts actually gave/give testimonies in courts about the age of fingermarks, even if mostly based on subjective and badly documented parameters. It was relatively easy to access fully described American cases, thus explaining the origin of the given examples. However, fingermark dating issues are encountered worldwide, and the lack of consensus among the given answers highlights the necessity to conduct research on the subject. The present work thus aims at studying the possibility to develop an objective fingermark dating method. As the questions surrounding the development of dating procedures are not new, different attempts were already described in the literature. This research proposes a critical review of these attempts and highlights that most of the reported methodologies still suffer from limitations preventing their use in actual practice. Nevertheless, some approaches based on the evolution of intrinsic compounds detected in fingermark residue over time appear to be promising. Thus, an exhaustive review of the literature was conducted in order to identify the compounds available in the fingermark residue and the analytical techniques capable of analysing them. It was chosen to concentrate on sebaceous compounds analysed using gas chromatography coupled with mass spectrometry (GC/MS) or Fourier transform infrared spectroscopy (FTIR). GC/MS analyses were conducted in order to characterize the initial variability of target lipids among fresh fingermarks of the same donor (intra-­‐variability) and between fingermarks of different donors (inter-­‐variability). As a result, many molecules were identified and quantified for the first time in fingermark residue. Furthermore, it was determined that the intra-­‐variability of the fingermark residue was significantly lower than the inter-­‐variability, but that it was possible to reduce both kind of variability using different statistical pre-­‐ treatments inspired from the drug profiling area. It was also possible to propose an objective donor classification model allowing the grouping of donors in two main classes based on their initial lipid composition. These classes correspond to what is relatively subjectively called "good" or "bad" donors. The potential of such a model is high for the fingermark research field, as it allows the selection of representative donors based on compounds of interest. Using GC/MS and FTIR, an in-­‐depth study of the effects of different influence factors on the initial composition and aging of target lipid molecules found in fingermark residue was conducted. It was determined that univariate and multivariate models could be build to describe the aging of target compounds (transformed in aging parameters through pre-­‐ processing techniques), but that some influence factors were affecting these models more than others. In fact, the donor, the substrate and the application of enhancement techniques seemed to hinder the construction of reproducible models. The other tested factors (deposition moment, pressure, temperature and illumination) also affected the residue and their aging, but models combining different values of these factors still proved to be robust. Furthermore, test-­‐fingermarks were analysed with GC/MS in order to be dated using some of the generated models. It turned out that correct estimations were obtained for 60% of the dated test-­‐fingermarks and until 100% when the storage conditions were known. These results are interesting but further research should be conducted to evaluate if these models could be used in uncontrolled casework conditions. In a more fundamental perspective, a pilot study was also conducted on the use of infrared spectroscopy combined with chemical imaging in order to gain information about the fingermark composition and aging. More precisely, its ability to highlight influence factors and aging effects over large areas of fingermarks was investigated. This information was then compared with that given by individual FTIR spectra. It was concluded that while FTIR-­‐ CI is a powerful tool, its use to study natural fingermark residue for forensic purposes has to be carefully considered. In fact, in this study, this technique does not yield more information on residue distribution than traditional FTIR spectra and also suffers from major drawbacks, such as long analysis and processing time, particularly when large fingermark areas need to be covered. Finally, the results obtained in this research allowed the proposition and discussion of a formal and pragmatic framework to approach the fingermark dating questions. It allows identifying which type of information the scientist would be able to bring so far to investigators and/or Justice. Furthermore, this proposed framework also describes the different iterative development steps that the research should follow in order to achieve the validation of an objective fingermark dating methodology, whose capacities and limits are well known and properly documented.

Asiakkuuksien strateginen arvottaminen ja luokittelu

Tämän tutkimuksen kohdeorganisaatio on suuren teollisuusyrityksen sisäinen raaka-aineen hankkija ja toimittaja. Tutkimuksessa selvitetään, mistä kohdeorganisaation hankinta-asiakkuuksien arvo muodostuu ja kuinka olemassa olevan liiketoimintadatan perusteella voidaan tutkia, arvioida ja luokitella kauppojen ja asiakkuuksien arvokkuutta aikaan sitomatta, objektiivisesti ja luotettavasti. Tutkimuksen teoriaosiossa esitellään lähestymistapoja ja menetelmiä, joiden avulla voidaan jalostaa olemassa olevasta datasta uutta sidosryhmätietämystä liiketoiminnan käyttöön, sekä tarkastellaan asiakaskannattavuusanalyysin, portfolioanalyysin, sekä asiakassegmentoinnin perusteita ja malleja. Näiden teorioiden ja mallien pohjalta rakennetaan kohdeorganisaatiolle räätälöity, indeksoituihin hinta-, määrä- ja kauppojen toistuvuus-muuttujiin perustuva, asiakkuuksien arvottamis- ja luokittelumalli. Arvottamis- ja luokittelumalli testataan vuosien 2003–2007 liiketoimintadatasta muodostetulla 389 336 kaupparivin otoksella, joka sisältää 42 186 arvioitavaa asiakkuussuhdetta. Merkittävin esille nouseva havainto on noin 5 000:n keskimääräistä selkeästi kalliimman asiakkuuden ryhmä. Aineisto ja sen poikkeavuudet testataan tilastollisin menetelmin, jotta saadaan selville asiakkuuden arvoon vaikuttavat ja arvoa selittävät tekijät. Lopuksi pohditaan arvottamismallin merkitystä analyyttisemman ostotoiminnan ja asiakkuudenhallinnan välineenä, sekä esitetään muutamia parannusehdotuksia.

Distinção entre cachaças destiladas em alambiques e em colunas usando quimiometria

One hundred fifteen cachaça samples derived from distillation in copper stills (73) or in stainless steels (42) were analyzed for thirty five itens by chromatography and inductively coupled plasma optical emission spectrometry. The analytical data were treated through Factor Analysis (FA), Partial Least Square Discriminant Analysis (PLS-DA) and Quadratic Discriminant Analysis (QDA). The FA explained 66.0% of the database variance. PLS-DA showed that it is possible to distinguish between the two groups of cachaças with 52.8% of the database variance. QDA was used to build up a classification model using acetaldehyde, ethyl carbamate, isobutyl alcohol, benzaldehyde, acetic acid and formaldehyde as chemical descriptors. The model presented 91.7% of accuracy on predicting the apparatus in which unknown samples were distilled.

A Business Incubator, Accelerator, or Coworking Space? Case Health Innovation Village at GE

Health Innovation Village at GE is one of the new communities targeted for startup and growth-oriented companies. It has been established at the premises of a multinational conglomerate that will promote networking and growth of startup companies. The concept combines features from traditional business incubators, accelerators, and coworking spaces. This research compares Health Innovation Village to these concepts regarding its goals, target clients, source of income, organization, facilities, management, and success factors. In addition, a new incubator classification model is introduced. On the other hand, Health Innovation Village is examined from its tenants’ perspective and improvements are suggested. The work was implemented as a qualitative case study by interviewing GE staff with connections to Health Innovation Village as well as startup entrepreneurs and employees’ working there. The most evident features of Health Innovation Village correspond to those of business incubators although it is atypical as a non-profit corporate business incubator. Strong network orientation and connections to venture capitalists are common characteristics of these new types of accelerators. The design of the premises conforms to the principles of coworking spaces, but the services provided to the startup companies are considerably more versatile than the services offered by coworking spaces. The advantages of Health Innovation Village are that there are first-class premises and exceptionally good networking possibilities that other types of incubators or accelerators are not able to offer. A conglomerate can also provide multifaceted special knowledge for young firms. In addition, both GE and the startups gained considerable publicity through their cooperation, indeed a characteristic that benefits both parties. Most of the expectations of the entrepreneurs were exceeded. However, communication and the scope of cooperation remain challenges. Micro companies spend their time developing and marketing their products and acquiring financing. Therefore, communication should be as clear as possible and accessible everywhere. The startups would prefer to cooperate significantly more, but few have the time available to assume the responsibility of leadership. The entrepreneurs also expected to have more possibilities for cooperation with GE. Wider collaboration might be accomplished by curation in the same way as it is used in the well-functioning coworking spaces where curators take care of practicalities and promote cooperation. Communication issues could be alleviated if the community had its own Intranet pages where all information could be concentrated. In particular, a common calendar and a room reservation system could be useful. In addition, it could be beneficial to have a section of the Intranet open for both the GE staff and the startups so that those willing to share their knowledge and those having project offers could use it for advertising.

Detecting Faces in Impoverished Images

The ability to detect faces in images is of critical ecological significance. It is a pre-requisite for other important face perception tasks such as person identification, gender classification and affect analysis. Here we address the question of how the visual system classifies images into face and non-face patterns. We focus on face detection in impoverished images, which allow us to explore information thresholds required for different levels of performance. Our experimental results provide lower bounds on image resolution needed for reliable discrimination between face and non-face patterns and help characterize the nature of facial representations used by the visual system under degraded viewing conditions. Specifically, they enable an evaluation of the contribution of luminance contrast, image orientation and local context on face-detection performance.

Autonomia da Perícia Oficial Criminal no Brasil: identificação e classificação de stakeholders

Este estudo teve por objetivo identificar e classificar os stakeholders que influenciam e/ou são influenciados pela perícia oficial de natureza criminal e que, direta ou indiretamente, impactam a autonomia desse arranjo estatal, bem como, obter um conjunto de ações aos peritos oficiais como forma de exercer algum grau de influência no processo de autonomia. Foi abordada a Teoria dos Stakeholders e destacadas algumas propostas de classificação de partes interessadas. Foi eleita a proposta de classificação desenvolvida por Mainardes, Alves, et al. em trabalho apresentado no V Encontro de Estudos em Estratégia/3Es, realizado pela Associação Nacional de Pós-Graduação e Pesquisa em Administração (ANPAD), na cidade de Porto Alegre/RS, em maio de 2013, com o intuito de executar o teste empírico recomendado pelos autores da nova proposta de classificação. Os stakeholders foram identificados como resultado de entrevista coletiva a grupo focal composto por seis peritos oficiais criminais. Para fins de aplicação do teste empírico do modelo da proposta de classificação de stakeholders de Mainardes, Alves, et al. foi considerado um contexto empírico onde a organização, para fins da Teoria dos Stakeholders e deste trabalho, é a perícia oficial criminal e sua respectiva autonomia. Como resultado desse teste empírico, os stakeholders identificados e entrevistados foram classificados em dependente, passivo, parceiro, controlador ou regulador. Aqueles não abrangidos pela classificação são tidos como não stakeholders. Por fim, foi sugerido por esses stakeholders entrevistados um conjunto de ações aos peritos oficiais criminais como forma de exercerem algum grau de influência no processo de autonomia da perícia oficial de natureza criminal.