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.

Efficient feature selection filters for high-dimensional data

Feature selection is a central problem in machine learning and pattern recognition. On large datasets (in terms of dimension and/or number of instances), using search-based or wrapper techniques can be cornputationally prohibitive. Moreover, many filter methods based on relevance/redundancy assessment also take a prohibitively long time on high-dimensional. datasets. In this paper, we propose efficient unsupervised and supervised feature selection/ranking filters for high-dimensional datasets. These methods use low-complexity relevance and redundancy criteria, applicable to supervised, semi-supervised, and unsupervised learning, being able to act as pre-processors for computationally intensive methods to focus their attention on smaller subsets of promising features. The experimental results, with up to 10(5) features, show the time efficiency of our methods, with lower generalization error than state-of-the-art techniques, while being dramatically simpler and faster.

Modifications of MCM-22 zeolite through sequential post-synthesis treatments. Implications on the acidic and catalytic behaviour

Desilication and a combination of alkaline followed by acid treatment were applied to MCM-22 zeolite using two different base concentrations. The samples were characterised by powder X-ray diffraction, Al-27 and Si-29 MAS-NMR spectroscopy, SEM, TEM and low temperature N-2 adsorption. The acidity of the samples was study through pyridine adsorption followed by FTIR spectroscopy and by the analyses of the hydroxyl region. The catalytic behaviour, anticipated by the effect of post-synthesis treatments on the acidity and space available inside the two internal pore systems was evaluated by using the model reaction of m-xylene transformation. The generation of mesoporosity was achieved upon alkaline treatment with 0.05 M NaOH solution and practically no additional gain was obtained when the more concentrate solution, 0.1 M, was used. Instead, Al extraction takes place along with Si, as shown by Si-29 and Al-27 MAS-NMR data, followed by Al deposition as extraframework species. Samples submitted to alkaline plus acid treatments present distinct behaviour. When the lowest NaOH solution was used no relevant effect was observed on the textural characteristics. Additionally, when the acid treatment was performed on an already fragilized MCM-22 structure, due to previous desilication with 0.1 M NaOH solution, the extraction of Al from both internal pore systems promotes their interconnection, evolving from a 2-D to a 3-D porous structure. This transformation has a marked effect in the catalytic behaviour, allowing an increase of m-xylene conversion as a consequence of an easier and faster molecular traffic in the 3-D structure. On the other hand, the continuous deposition of extraframework Al species inside the pores leads to a shape selective effect that privileges the formation of the more valuable isomer p-xylene.

Paracoccidioidomycosis: a sequential histopathologic study of lesions in experimentally-infected rats

Female albino rats were used for the sequential histopathological study of experimental paracoccidioidomycosis. The animals were inoculated intraperitoneally with a strain of Paracoccidioides brasiliensis in the yeast-like phase, and sacrificed at given intervals from 1 to 168 days after inoculation; each animal received an inoculum of 4 x 10(6) cells in 0.8 ml of saline. The control group received saline containing scrapings of the culture medium. Tissue from the inoculation site was examined. The cellular population, the extracellular matrix, and the presence and characteristics of fungi were analysed in the inflammatory granulomatous process by light microscopy. The results allowed to separate the kinetic of the inflammatory response into three stages: 1) neutrophilic or macrophagic-neutrophilic; 2) pre-granulomatous; 3) granulomatous. Synthesis of the extracellular matrix began with the depositing of fibrin-like material, and increased gradually with deposits of collagen, proteoglycans, and glycoproteins. Parasites were present in all of the examined periods. Recurrences of the disease were clearly shown through the concurrence of recently-formed granulomas with older granulomas, implying that this type of granulomatous process does not eliminate the disease, nor is it able to limit fungal dissemination over a prolonged period of time.

Enhancing the selection of a model-based clustering with external categorical variables

In cluster analysis, it can be useful to interpret the partition built from the data in the light of external categorical variables which are not directly involved to cluster the data. An approach is proposed in the model-based clustering context to select a number of clusters which both fits the data well and takes advantage of the potential illustrative ability of the external variables. This approach makes use of the integrated joint likelihood of the data and the partitions at hand, namely the model-based partition and the partitions associated to the external variables. It is noteworthy that each mixture model is fitted by the maximum likelihood methodology to the data, excluding the external variables which are used to select a relevant mixture model only. Numerical experiments illustrate the promising behaviour of the derived criterion.

Materials selection for a set of multiple parts considering manufacturing costs and weight reduction with structural isoperformance using direct multisearch optimization

Materials selection is a matter of great importance to engineering design and software tools are valuable to inform decisions in the early stages of product development. However, when a set of alternative materials is available for the different parts a product is made of, the question of what optimal material mix to choose for a group of parts is not trivial. The engineer/designer therefore goes about this in a part-by-part procedure. Optimizing each part per se can lead to a global sub-optimal solution from the product point of view. An optimization procedure to deal with products with multiple parts, each with discrete design variables, and able to determine the optimal solution assuming different objectives is therefore needed. To solve this multiobjective optimization problem, a new routine based on Direct MultiSearch (DMS) algorithm is created. Results from the Pareto front can help the designer to align his/hers materials selection for a complete set of materials with product attribute objectives, depending on the relative importance of each objective.

Exploiting the bin-class histograms for feature selection on discrete data

In machine learning and pattern recognition tasks, the use of feature discretization techniques may have several advantages. The discretized features may hold enough information for the learning task at hand, while ignoring minor fluctuations that are irrelevant or harmful for that task. The discretized features have more compact representations that may yield both better accuracy and lower training time, as compared to the use of the original features. However, in many cases, mainly with medium and high-dimensional data, the large number of features usually implies that there is some redundancy among them. Thus, we may further apply feature selection (FS) techniques on the discrete data, keeping the most relevant features, while discarding the irrelevant and redundant ones. In this paper, we propose relevance and redundancy criteria for supervised feature selection techniques on discrete data. These criteria are applied to the bin-class histograms of the discrete features. The experimental results, on public benchmark data, show that the proposed criteria can achieve better accuracy than widely used relevance and redundancy criteria, such as mutual information and the Fisher ratio.

The selection of senior civil servants in the context of the different public administration models: an issue of accountability

Comunicação apresentada no Congresso do IIAS-IISA no âmbito do IX Grupo de Estudo: Serviço público e política, realizado em Ifrane, Marrocos de 13 a 17 de junho de 2014

Finding a school technological partner: a multicriteria method for school information system producer selection

The choice of an information systems is a critical factor of success in an organization's performance, since, by involving multiple decision-makers, with often conflicting objectives, several alternatives with aggressive marketing, makes it particularly complex by the scope of a consensus. The main objective of this work is to make the analysis and selection of a information system to support the school management, pedagogical and administrative components, using a multicriteria decision aid system – MMASSITI – Multicriteria Method- ology to Support the Selection of Information Systems/Information Technologies – integrates a multicriteria model that seeks to provide a systematic approach in the process of choice of Information Systems, able to produce sustained recommendations concerning the decision scope. Its application to a case study has identi- fied the relevant factors in the selection process of school educational and management information system and get a solution that allows the decision maker’ to compare the quality of the various alternatives.

Group decision making for selection of an information system in a business context

The main objective of this work is to report on the development of a multi-criteria methodology to support the assessment and selection of an Information System (IS) framework in a business context. The objective is to select a technological partner that provides the engine to be the basis for the development of a customized application for shrinkage reduction on the supply chains management. Furthermore, the proposed methodology di ers from most of the ones previously proposed in the sense that 1) it provides the decision makers with a set of pre-defined criteria along with their description and suggestions on how to measure them and 2)it uses a continuous scale with two reference levels and thus no normalization of the valuations is required. The methodology here proposed is has been designed to be easy to understand and use, without a specific support of a decision making analyst.

Expatriates recruitment and selection for long-term international assignments in portuguese companies

Long-term international assignments’ increase requires more attention being paid for the preparation of these foreign assignments, especially on the recruitment and selection process of expatriates. This article explores how the recruitment and selection process of expatriates is developed in Portuguese companies, examining the main criteria on recruitment and selection of expatriates’ decision to send international assignments. The paper is based on qualitative case studies of companies located in Portugal. The data were collected through semi-structured interviews of 42 expatriates and 18 organisational representatives as well from nine Portuguese companies. The findings show that the most important criteria are: (1) trust from managers, (2) years in service, (3) previous technical and language competences, (4) organisational knowledge and, (5) availability. Based on the findings, the article discusses in detail the main theoretical and managerial implications. Suggestions for further research are also presented.

A proteomic approach toward the selection of proteins with enhanced intrinsic conformational stability

Journal of Proteome Research (2006)5: 2720-2726

A study on parallel versus sequential relational fuzzy clustering methods

Dissertação para obtenção do Grau de Mestre em Engenharia Informática

Enterococcus and biocides: mechanisms of tolerance and selection for vancomycin resistance

Dissertation presented to obtain a Doctoral degree in Biology, Instituto de Tecnologia Química e Biológica, Universidade Nova de Lisboa.

An approach for automated service selection and ranking using services choreography

In today’s highly competitive market, it is critical to provide customers services with a high level of configuration to answer their business needs. Knowing in advance the performance associated with a specific choreography of services (e.g., by taking into account the expected results of each component service) represents an important asset that allows businesses to provide a global service tailored to customers’ specific requests. This research work aims at advancing the state-of-the-art in this area by proposing an approach for service selection and ranking using services choreography, predicting the behavior of the services considering customers’ requirements and preferences, business process constraints and characteristics of the execution environment.