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. © 2014 Springer-Verlag Berlin Heidelberg.

Machine learning Vasicek model calibration with Gaussian processes

In this article, we calibrate the Vasicek interest rate model under the risk neutral measure by learning the model parameters using Gaussian processes for machine learning regression. The calibration is done by maximizing the likelihood of zero coupon bond log prices, using mean and covariance functions computed analytically, as well as likelihood derivatives with respect to the parameters. The maximization method used is the conjugate gradients. The only prices needed for calibration are zero coupon bond prices and the parameters are directly obtained in the arbitrage free risk neutral measure.

Assessing organizations collaboration readiness: a behavioral approach

Dissertation presented at the Faculty of Sciences and Technology of the New University of Lisbon to obtain the degree of Doctor in Electrical Engineering, specialty of Robotics and Integrated Manufacturing

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.

McCoy cell line as a possible model containing CD4+ receptors for the study of HIV-1 replication

Several studies have recently shown the use of recombinant rabies virus as potential vector-viral vaccine for HIV-1. The sequence homology between gp 120 and rabies virus glycoprotein has been reported. The McCoy cell line has therefore been used to show CD4+ or CD4+ like receptors. Samples of HIV-1 were isolated, when plasma of HIV-1 positive patients was inoculated in the McCoy cell line. The virus infection was then studied during successive virus passages. The proteins released in the extra cellular medium were checked for protein activity, by exposure to SDS Electrophoresis and blotting to nitro-cellulose filter, then reacting with sera of HIV positive and negative patients. Successive passages were performed, and showed viral replication, membrane permeabilization, the syncytium formation, and the cellular lysis (cytopathic effect). Flow cytometry analysis shows clear evidence that CD4+ receptors are present in this cell line, which enhances the likelihood of easy isolation and replication of HIV. The results observed allow the use of this cell line as a possible model for isolating HIV, as well as for carrying out studies of the dynamics of viral infection in several situations, including exposure to drugs in pharmacological studies, and possibly studies and analyses of the immune response in vaccine therapies.

ezGo: A voice operated wheelchair with biosignal monitoring for home environments

In this paper we present ezGo, an electric powered wheelchair with a speech based interface and biosignal monitoring instrumentation. The user can use the voice, a natural communication method, for controlling the chair movement and obtain information about his health. Additionally a set of semi-autonomous modes with macro recording enable the execution of navigation tasks with little effort and improved precision. The main purpose of the system is to provide severely disabled persons with an assistive device that can improve their confidence and daily independence. The obtained results on usability tests showed that users consider ezGo a valuable help on their daily tasks and a very desirable addition to standard wheelchairs.

Predição de Pré-Eclâmpsia no Primeiro Trimestre em Gravidezes de Baixo Risco: Determinação do Cut-Off numa Amostra da População Portuguesa

Objective:We aimed to identify the cut-off for risk of pre-eclampsia (PE) in Portuguese population by applying the first trimester prediction model from Fetal Medicine Foundation (FMF) in a prospective enrolled cohort of low risk pregnant women. Population and methods: A prospective cohort of low risk singleton pregnancies underwent routine first-trimester scree - ning from 2011 through 2013. Maternal characteristics, blood pressure, uterine artery Doppler, levels of pregnancy-associated plasma protein-A (PAPP-A) and free b-human chorionic gonadotropin were evaluated. The prediction of PE in first trimester was calculated through software Astraia, the outcome obtained from medical records and the cutoff value was subse quently calculated. Results:Of the 273 enrolled patients, 7 (2.6%) developed PE. In first trimester women who developed PE presented higher uterine arteries resistance, represented by higher values of lowest and mean uterine pulsatility index, p <0.005. There was no statistical significance among the remaining maternal characteristics, body mass index, blood pressure and PAPP-A. Using the FMF first trimester PE algorithm, an ideal cut-off of 0.045 (1/22) would correctly detect 71% women who developed PE for a 12% false positive rate and a likelihood ratio of 12.98 (area under the curve: 0.69; confidence interval 95%: 0.39-0.99). By applying the reported cutoff to our cohort, we would obtain 71.4% true positives, 88.3% true negatives, 11.4% false positives and 28.6% false negatives. Conclusion: By applying a first trimester PE prediction model to low risk pregnancies derived from a Portuguese population, a significant proportion of patients would have been predicted as high risk. New larger studies are required to confirm the present findings.

Creating a business model for a pharmaceutical bank in Lisbon

A Work Project, presented as part of the requirements for the Award of a Masters Degree in Management from the NOVA – School of Business and Economics

Men’s purchase intention of shampoo: a model based on the theory of planned behavior and self-concept

A Work Project, presented as part of the requirements for the Award of a Masters Degree in Management from the NOVA – School of Business and Economics

Protest attitudes and stated preferences: Evidence on scale usage heterogeneity

We contribute to the stated preference literature by addressing scale usage heterogeneity regarding how individuals answer attitudinal questions capturing lack of trust in institutions and fairness issues. Using a latent class model, we conduct a contingent valuation study to elicit the willingness-to-pay to preserve a recreational site. We find evidence that respondents within the same class, that is, with similar preferences and attitudes, interpret the Likert scale differently when answering the attitudinal questions. We identify different patterns of scale usage heterogeneity within and across classes and associate them with individual characteristics. Our approach contributes to better a understanding of individual behavior in the presence of protest attitudes.

Protesting or justifying? A latent class model for contingent valuation with attitudinal data

This article develops a latent class model for estimating willingness-to-pay for public goods using simultaneously contingent valuation (CV) and attitudinal data capturing protest attitudes related to the lack of trust in public institutions providing those goods. A measure of the social cost associated with protest responses and the consequent loss in potential contributions for providing the public good is proposed. The presence of potential justification biases is further considered, that is, the possibility that for psychological reasons the response to the CV question affects the answers to the attitudinal questions. The results from our empirical application suggest that psychological factors should not be ignored in CV estimation for policy purposes, allowing for a correct identification of protest responses.

Adaptação cultural e contributo para a validação da pain attitudes and believes scale for physiotherapists (PABS-PT) para a população de fisioterapeutas portugueses

RESUMO: Introdução: Vários estudos têm encontrado evidência para a relação entre as crenças e atitudes dos profissionais de saúde, a sua abordagem de tratamento, referida como orientada segundo o modelo biomédico ou modelo biopsicossocial (Bishop, 2008; Mutsaers, 2012), e os resultados obtidos. É sugerido que, no tratamento da dor lombar crónica, os profissionais que orientam o seu raciocínio e prática segundo o modelo biomédico, tendem a obter piores resultados quando comparados com os obtidos pelos profissionais que orientam o seu raciocínio e prática segundo uma abordagem biopsicossocial. Esta área de estudo tem salientado a importância de desenhar instrumentos capazes de identificar a orientação preferencial dos profissionais de saúde no tratamento da dor crónica de natureza músculo-esquelética, entre os quais se encontra a “Pain Attitudes and Beliefs Scale for Physiotherapists” (PABS-PT). Objetivo: Adaptar culturalmente a PABS-PT para a população de fisioterapeutas portugueses que intervêm em utentes com dor (lombar) crónica, e contribuir para a sua validação Metodologia: A versão original (língua inglesa) do PABS-PT foi adaptada para a língua e cultura portuguesas, através das etapas definidas nas normas orientadoras estabelecidas para este processo (Beaton et al., 2002; MAPI Institute, 2001). A avaliação das propriedades psicométricas da versão portuguesa foi realizada com recurso a uma amostra de 202 fisioterapeutas e estudantes finalistas do curso de licenciatura em Fisioterapia. Inicialmente foi realizada a análise fatorial exploratória da escala através do método das componentes principais. Posteriormente avaliou-se a consistência interna das componentes obtidas com recurso ao alpha de Cronbach (α). Para analisar a validade de constructo foram correlacionadas as componentes obtidas com a versão portuguesa da “Health Care Providers’ Pain and Impairment Relationship Scale” (HC-PAIRS), recorrendo ao cálculo do coeficiente de Spearman. Resultados: O processo de adaptação cultural não revelou dificuldades importantes sendo considerado que a versão portuguesa da PABS-PT é de fácil compreensão e preenchimento, e os seus itens adequados para avaliar as crenças e atitudes dos fisioterapeutas portugueses relativas à intervenção na dor crónica músculo-esquelética. Os resultados revelaram uma estrutura fatorial de duas componentes, identificadas com as componentes da escala original que explicam 30,96% da variância total. A consistência interna encontrada é boa, para a componente biomédica (α de Cronbach = 0,826), mas muito fraca para a componente biopsicossocial (α de Cronbach= 0,589). Relativamente à validade convergente e discriminativa, foi encontrada uma associação estatisticamente significativa e positiva, entre as componentes 1 (biomédica) da versão nportuguesa da PABS e a HC-PAIRS (Rs = 0,481, p≤ 0,005), e negativa, fraca e significativa entre a pontuação total da componente 2 (biopsicossocial) e a HC-PAIRS (Rs = -0,038, p=0,612). Conclusão: A versão portuguesa do PABS-PT é de fácil compreensão e aparenta ser um instrumento válido para a medição da orientação preferencial dos fisioterapeutas, relativamente às suas atitudes e crenças na avaliação e tratamento de utentes com dor crónica de natureza músculo-esquelética. No entanto, a componente biopsicossocial requer uma análise mais aprofundada para que possa, com rigor, ser utilizada na definição de uma orientação preferencialmente biopsicossocial.--------------ABSTRACT: Introduction: Previous studies have found a relation between the beliefs and attitudes of health professionals, their treatment approach, which can follows a biomedical or a biopsychosocial orientation (Bishop, 2008; Mutsaers, 2012), and the outcomes obtained. Therefore, is suggested that the professionals who tend to approach chronic low back pain patients according to the biomedical model have worse outcomes than professionals who use a biopsychosocial approach in there clinical reasoning. This research field has highlighted the importance of developing measures capable of identifying the preferred orientation of health professionals in the treatment of chronic pain of musculoskeletal nature, including the “Pain Attitudes and Beliefs Scale for Physiotherapists” (PABS-PT). Objective: To cross culturally adapt and validate the PABS-PT for the Portuguese population of physiotherapists. Methodology: The original version (English version) of the PABS-PT was adapted to the Portuguese language and culture, through the guidelines established for these processes (Beaton et al., 2002; MAPI Institute, 2001). The psychometric evaluation of the Portuguese version was carried out on a sample of 202 p physiotherapist and final year students of the physiotherapy course. Initially, an exploratory factorial analysis was performed through the method of the main components. Then, the internal consistence of the main components was evaluated using the Cronbach’s alpha (α). The convergent construct validity was analysed through the correlation between the obtained components of PABS-PT and the Health Care Provider’s Pain and Impairment Relationship Scale (HC-PAIRS), using the Spearmen correlation coefficient. Results: No major difficulties were found during the cultural adaptation process of PABS-PT to Portugal, which means that the Portuguese version is easy to understand and fulfill, and items are appropriated to evaluate the beliefs and attitudes of the Portuguese physiotherapists who treat chronic pain of musculoskeletal origins. The results revealed a factorial structure of two components, as the original scale, explaining 30,96% of the total variance. Internal consistence results were good, for the biomedical component (Cronbach’s α = 0,826), but very weak for the biopsychosocial componente (Cronbach’s α = 0,589). Relatively to convergent and discriminative validity, a statistically significant association was found, between the components 1 (biomedical) of the Portuguese version of PABS-PT and the HC-PAIRS (Rs= 0,481, p≤ 0,005) and negative, weak and significant between the total score of component 2 (biopsychosocial) and the HC-PAIRS (Rs = -0,038, p=0,612). Conclusion: The Portuguese Version of PABS-PT is easy to understand and seems to be a valid instrument to measure the attitudes and beliefs of physiotherapists in the management of patients with chronic low back pain. However, the biopsychosocial component requires a further deep analysis to examine a preferable biopsychosocial orientation.

An analog model for global macro investing

This study proposes a systematic model that is able to fit the Global Macro Investing universe. The Analog Model tests the possibility of capturing the likelihood of an optimal investment allocation based on similarity across different periods in history. Instead of observing Macroeconomic data, the model uses financial markets’ variables to classify unknown short-term regimes. This methodology is particularly relevant considering that asset classes and investment strategies react differently to specific macro environment shifts.

A mediated moderation study to explain the negative relationship between job insecurity and employee supportive voice

Double degree

Ethical leadership and unethical pro-organizational behavior: A moderated mediation model of affective commitment and authenticity at work

Recently, unethical conduct in the workplace has been a focus of literature and media. Unethical pro-organizational behavior (UPB) refers to unethical conduct that employees engage in to benefit the organization. Given the complexity of UPB, there is an increasing need to understand how and under what conditions this attitude originates within organizations. Based on a sample of 167 employees and seven organizations, results support the moderated mediation model. An ethical leader increases employees’ organizational affective commitment which increases the likelihood to engage in UPB. However, the indirect relationship diminishes when employees feel authentic at work.