NovaIntell - projecto de text-Mining para a língua portuguesa numa empresa de Gestão de Informação e Conhecimento

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

Selection Constructive based Hyper-heuristic for Dynamic Scheduling

A função de escalonamento desempenha um papel importante nos sistemas de produção. Os sistemas de escalonamento têm como objetivo gerar um plano de escalonamento que permite gerir de uma forma eficiente um conjunto de tarefas que necessitam de ser executadas no mesmo período de tempo pelos mesmos recursos. Contudo, adaptação dinâmica e otimização é uma necessidade crítica em sistemas de escalonamento, uma vez que as organizações de produção têm uma natureza dinâmica. Nestas organizações ocorrem distúrbios nas condições requisitos de trabalho regularmente e de forma inesperada. Alguns exemplos destes distúrbios são: surgimento de uma nova tarefa, cancelamento de uma tarefa, alteração na data de entrega, entre outros. Estes eventos dinâmicos devem ser tidos em conta, uma vez que podem influenciar o plano criado, tornando-o ineficiente. Portanto, ambientes de produção necessitam de resposta imediata para estes eventos, usando um método de reescalonamento em tempo real, para minimizar o efeito destes eventos dinâmicos no sistema de produção. Deste modo, os sistemas de escalonamento devem de uma forma automática e inteligente, ser capazes de adaptar o plano de escalonamento que a organização está a seguir aos eventos inesperados em tempo real. Esta dissertação aborda o problema de incorporar novas tarefas num plano de escalonamento já existente. Deste modo, é proposta uma abordagem de otimização – Hiper-heurística baseada em Seleção Construtiva para Escalonamento Dinâmico- para lidar com eventos dinâmicos que podem ocorrer num ambiente de produção, a fim de manter o plano de escalonamento, o mais robusto possível. Esta abordagem é inspirada em computação evolutiva e hiper-heurísticas. Do estudo computacional realizado foi possível concluir que o uso da hiper-heurística de seleção construtiva pode ser vantajoso na resolução de problemas de otimização de adaptação dinâmica.

Automatic syllabification for danish text-to-speech systems

In this paper, a rule-based automatic syllabifier for Danish is described using the Maximal Onset Principle. Prior success rates of rule-based methods applied to Portuguese and Catalan syllabification modules were on the basis of this work. The system was implemented and tested using a very small set of rules. The results gave rise to 96.9% and 98.7% of word accuracy rate, contrary to our initial expectations, being Danish a language with a complex syllabic structure and thus difficult to be rule-driven. Comparison with data-driven syllabification system using artificial neural networks showed a higher accuracy rate of the former system.

Enterobacteriaceae ISOLATES FROM THE ORAL CAVITY OF WORKERS IN A BRAZILIAN ONCOLOGY HOSPITAL

The evaluation of workers as potential reservoirs and disseminators of pathogenic bacteria has been described as a strategy for the prevention and control of healthcare-associated infections (HAIs). The aim of this study was to evaluate the presence of Enterobacteriaceae in the oral cavity of workers at an oncology hospital in the Midwest region of Brazil, as well as to characterize the phenotypic profile of the isolates. Saliva samples of 294 workers from the hospital’s healthcare and support teams were collected. Microbiological procedures were performed according to standard techniques. Among the participants, 55 (18.7%) were colonized by Enterobacteriaceae in the oral cavity. A total of 64 bacteria were isolated, including potentially pathogenic species. The most prevalent species was Enterobacter gergoviae (17.2%). The highest rates of resistance were observed for β-lactams, and 48.4% of the isolates were considered multiresistant. Regarding the enterobacteria isolated, the production of ESBL and KPC was negative. Nevertheless, among the 43 isolates of the CESP group, 51.2% were considered AmpC β-lactamase producers by induction, and 48.8% were hyper-producing mutants. The significant prevalence of carriers of Enterobacteriaceae and the phenotypic profile of the isolates represents a concern, especially due to the multiresistance and production of AmpC β-lactamases.

Hyper-IgD and Periodic Fever Syndrome: a New MVK Mutation (p.R277G) Associated with a Severe Phenotype

Hyperimmunoglobulinemia D and periodic fever syndrome (HIDS; MIM#260920) is a rare recessively-inherited autoinflammatory condition caused bymutations in the MVK gene, which encodes for mevalonate kinase, an essential enzyme in the isoprenoid pathway. HIDS is clinically characterized by recurrent episodes of fever and inflammation. Herewe report on the case of a 2 year-old Portuguese boy with recurrent episodes of fever, malaise, massive cervical lymphadenopathy and hepatosplenomegaly since the age of 12 months. Rash, arthralgia, abdominal pain and diarrhea were also seen occasionally. During attacks a vigorous acute-phase response was detected, including elevated erythrocyte sedimentation rate, C-reactive protein, serum amyloid A and leukocytosis. Clinical and laboratory improvement was seen between attacks. Despite normal serum IgD level, HIDS was clinically suspected. Mutational MVK analysis revealed the homozygous genotype with the novel p.Arg277Gly (p.R277G) mutation, while the healthy non consanguineous parents were heterozygous. Short nonsteroidal anti-inflammatory drugs and corticosteroid courses were given during attacks with poor benefits, where as anakinra showed positive responses only at high doses. The p.R277Gmutation here described is a novel missense MVK mutation, and it has been detected in this casewith a severe HIDS phenotype. Further studies are needed to evaluate a co-relation genotype, enzyme activity and phenotype, and to define the best therapeutic strategies.

Text Mining: análise de sentimentos na classificação de notícias

Trabalho de Projeto apresentado como requisito parcial para obtenção do grau de Mestre em Estatística e Gestão de Informação

Automatic information retrieval through text-mining

The dissertation presented for obtaining the Master’s Degree in Electrical Engineering and Computer Science, at Universidade Nova de Lisboa, Faculdade de Ciências e Tecnologia

Spatial analysis of new cases of leprosy in the State of Espírito Santo, Brazil, between 2004 and 2009

INTRODUCTION: Leprosy in Brazil is a public health issue, and there are many regions in the State of Espírito Santo with high endemic incidence levels of leprosy, characterizing this state as a priority for leprosy programs. The aim of this study was to determine the spatial distribution of coefficients of new cases of leprosy in the State of Espírito Santo, Brazil. METHODS: We conducted a descriptive and ecologic study based on the spatial distribution of leprosy in the State of Espírito Santo between 2004 and 2009. Data were gathered from the available records of the Espírito Santo State Health Secretary. The global and local Bayesian empirical methods were used to produce an estimate of leprosy risk, smoothing the fluctuation effects of the detection coefficients. RESULTS: The study resulted in a coefficient adjustment of new cases in 10 towns that changed their classification, among which, 2 went from low to medium, 4 from medium to high, 3 from high to very high, and 1 from very high to hyper-endemic. An average variation of 1.02, fluctuating between 0 and 12.39 cases/100,000 inhabitants, was found in a comparative calculation between the Local Ebest value and the average coefficient of new leprosy cases in the State of Espírito Santo. CONCLUSIONS: The spatial analysis of leprosy favors the establishment of control strategies with a better cost-benefit relationship since it reveals specific and priority regions, thereby enabling the development of actions that can interfere in the transmission chain.

Musées et casinos dans une ville "patrimoine mondial": authenticité et hyperréalité, deux formes culturelles de l'espace urbain de Macao

Texto submetido para publicação em 2011.

Text mining e twitter : o poder das redes sociais num mercado competitivo

Actualmente, com a massificação da utilização das redes sociais, as empresas passam a sua mensagem nos seus canais de comunicação, mas os consumidores dão a sua opinião sobre ela. Argumentam, opinam, criticam (Nardi, Schiano, Gumbrecht, & Swartz, 2004). Positiva ou negativamente. Neste contexto o Text Mining surge como uma abordagem interessante para a resposta à necessidade de obter conhecimento a partir dos dados existentes. Neste trabalho utilizámos um algoritmo de Clustering hierárquico com o objectivo de descobrir temas distintos num conjunto de tweets obtidos ao longo de um determinado período de tempo para as empresas Burger King e McDonald’s. Com o intuito de compreender o sentimento associado a estes temas foi feita uma análise de sentimentos a cada tema encontrado, utilizando um algoritmo Bag-of-Words. Concluiu-se que o algoritmo de Clustering foi capaz de encontrar temas através do tweets obtidos, essencialmente ligados a produtos e serviços comercializados pelas empresas. O algoritmo de Sentiment Analysis atribuiu um sentimento a esses temas, permitindo compreender de entre os produtos/serviços identificados quais os que obtiveram uma polaridade positiva ou negativa, e deste modo sinalizar potencias situações problemáticas na estratégia das empresas, e situações positivas passíveis de identificação de decisões operacionais bem-sucedidas.

Business intelligence in banking: A literature analysis from 2002 to 2013 using Text Mining and latent Dirichlet allocation

telligence applications for the banking industry. Searches were performed in relevant journals resulting in 219 articles published between 2002 and 2013. To analyze such a large number of manuscripts, text mining techniques were used in pursuit for relevant terms on both business intelligence and banking domains. Moreover, the latent Dirichlet allocation modeling was used in or- der to group articles in several relevant topics. The analysis was conducted using a dictionary of terms belonging to both banking and business intelli- gence domains. Such procedure allowed for the identification of relationships between terms and topics grouping articles, enabling to emerge hypotheses regarding research directions. To confirm such hypotheses, relevant articles were collected and scrutinized, allowing to validate the text mining proce- dure. The results show that credit in banking is clearly the main application trend, particularly predicting risk and thus supporting credit approval or de- nial. There is also a relevant interest in bankruptcy and fraud prediction. Customer retention seems to be associated, although weakly, with targeting, justifying bank offers to reduce churn. In addition, a large number of ar- ticles focused more on business intelligence techniques and its applications, using the banking industry just for evaluation, thus, not clearly acclaiming for benefits in the banking business. By identifying these current research topics, this study also highlights opportunities for future research.

Reconstructing transcriptional regulatory networks using data integration and text mining

Transcriptional Regulatory Networks (TRNs) are powerful tool for representing several interactions that occur within a cell. Recent studies have provided information to help researchers in the tasks of building and understanding these networks. One of the major sources of information to build TRNs is biomedical literature. However, due to the rapidly increasing number of scientific papers, it is quite difficult to analyse the large amount of papers that have been published about this subject. This fact has heightened the importance of Biomedical Text Mining approaches in this task. Also, owing to the lack of adequate standards, as the number of databases increases, several inconsistencies concerning gene and protein names and identifiers are common. In this work, we developed an integrated approach for the reconstruction of TRNs that retrieve the relevant information from important biological databases and insert it into a unique repository, named KREN. Also, we applied text mining techniques over this integrated repository to build TRNs. However, was necessary to create a dictionary of names and synonyms associated with these entities and also develop an approach that retrieves all the abstracts from the related scientific papers stored on PubMed, in order to create a corpora of data about genes. Furthermore, these tasks were integrated into @Note, a software system that allows to use some methods from the Biomedical Text Mining field, including an algorithms for Named Entity Recognition (NER), extraction of all relevant terms from publication abstracts, extraction relationships between biological entities (genes, proteins and transcription factors). And finally, extended this tool to allow the reconstruction Transcriptional Regulatory Networks through using scientific literature.

Text mining na análise de sentimentos em contextos de big data

Dissertação de mestrado integrado em Engenharia e Gestão de Sistemas de Informação