Privacy-preserving topic model for tagging recommender systems

Tagging recommender systems provide users the freedom to explore tags and obtain recommendations. The releasing and sharing of these tagging datasets will accelerate both commercial and research work on recommender systems. However, releasing the original tagging datasets is usually confronted with serious privacy concerns, because adversaries may re-identify a user and her/his sensitive information from tagging datasets with only a little background information. Recently, several privacy techniques have been proposed to address the problem, but most of these lack a strict privacy notion, and rarely prevent individuals being re-identified from the dataset. This paper proposes a privacy- preserving tag release algorithm, PriTop. This algorithm is designed to satisfy differential privacy, a strict privacy notion with the goal of protecting users in a tagging dataset. The proposed PriTop algorithm includes three privacy-preserving operations: Private topic model generation structures the uncontrolled tags; private weight perturbation adds Laplace noise into the weights to hide the numbers of tags; while private tag selection finally finds the most suitable replacement tags for the original tags, so the exact tags can be hidden. We present extensive experimental results on four real-world datasets, Delicious, MovieLens, Last.fm and BibSonomy. While the recommendation algorithm is successful in all the cases, our results further suggest the proposed PriTop algorithm can successfully retain the utility of the datasets while preserving privacy.

The addition of depression to the Framingham Risk Equation model for predicting coronary heart disease risk in women

BACKGROUND: Depression is widely considered to be an independent and robust predictor of Coronary Heart Disease (CHD), however is seldom considered in the context of formal risk assessment. We assessed whether the addition of depression to the Framingham Risk Equation (FRE) improved accuracy for predicting 10-year CHD in a sample of women.

DESIGN: A prospective, longitudinal design comprising an age-stratified, population-based sample of Australian women collected between 1993 and 2011 (n=862).

METHODS: Clinical depressive disorder was assessed using the Structured Clinical Interview for Diagnostic and Statistical Manual of Mental Disorders (SCID-I/NP), using retrospective age-of-onset data. A composite measure of CHD included non-fatal myocardial infarction, unstable angina coronary intervention or cardiac death. Cox proportional-hazards regression models were conducted and overall accuracy assessed using area under receiver operating characteristic (ROC) curve analysis.

RESULTS: ROC curve analyses revealed that the addition of baseline depression status to the FRE model improved its overall accuracy (AUC:0.77, Specificity:0.70, Sensitivity:0.75) when compared to the original FRE model (AUC:0.75, Specificity:0.73, Sensitivity:0.67). However, when calibrated against the original model, the predicted number of events generated by the augmented version marginally over-estimated the true number observed.

CONCLUSIONS: The addition of a depression variable to the FRE equation improves the overall accuracy of the model for predicting 10-year CHD events in women, however may over-estimate the number of events that actually occur. This model now requires validation in larger samples as it could form a new CHD risk equation for women.

Affective dysfunction in a mouse model of Rett syndrome: therapeutic effects of environmental stimulation and physical activity

Rett syndrome (RTT) is a neurodevelopmental disorder associated with mutations in the X-linked gene encoding methyl-CpG-binding protein 2 (MeCP2) and consequent dysregulation of brain maturation. Patients suffer from a range of debilitating physical symptoms, however, behavioral and emotional symptoms also severely affect their quality of life. Here, we present previously unreported and clinically relevant affective dysfunction in the female heterozygous Mecp2tm1Tam mouse model of RTT (129sv and C57BL6 mixed background). The affective dysfunction and aberrant anxiety-related behavior of the Mecp2+ / - mice were found to be reversible with environmental enrichment (EE) from 4 weeks of age. The effect of exercise alone (via wheel running) was also explored, providing the first evidence that increased voluntary physical activity in an animal model of RTT is beneficial for some phenotypes. Mecp2+ / - mutants displayed elevated corticosterone despite decreased Crh expression, demonstrating hypothalamic-pituitary-adrenal axis dysregulation. EE of Mecp2+ / - mice normalized basal serum corticosterone and hippocampal BDNF protein levels. The enrichment-induced rescue appears independent of the transcriptional regulation of the MeCP2 targets Bdnf exon 4 and Crh. These findings provide new insight into the neurodevelopmental role of MeCP2 and pathogenesis of RTT, in particular the affective dysfunction. The positive outcomes of environmental stimulation and physical exercise have implications for the development of therapies targeting the affective symptoms, as well as behavioral and cognitive dimensions, of this devastating neurodevelopmental disorder.

A new transfer learning framework with application to model-agnostic multi-task learning

Learning from small number of examples is a challenging problem in machine learning. An effective way to improve the performance is through exploiting knowledge from other related tasks. Multi-task learning (MTL) is one such useful paradigm that aims to improve the performance through jointly modeling multiple related tasks. Although there exist numerous classification or regression models in machine learning literature, most of the MTL models are built around ridge or logistic regression. There exist some limited works, which propose multi-task extension of techniques such as support vector machine, Gaussian processes. However, all these MTL models are tied to specific classification or regression algorithms and there is no single MTL algorithm that can be used at a meta level for any given learning algorithm. Addressing this problem, we propose a generic, model-agnostic joint modeling framework that can take any classification or regression algorithm of a practitioner’s choice (standard or custom-built) and build its MTL variant. The key observation that drives our framework is that due to small number of examples, the estimates of task parameters are usually poor, and we show that this leads to an under-estimation of task relatedness between any two tasks with high probability. We derive an algorithm that brings the tasks closer to their true relatedness by improving the estimates of task parameters. This is achieved by appropriate sharing of data across tasks. We provide the detail theoretical underpinning of the algorithm. Through our experiments with both synthetic and real datasets, we demonstrate that the multi-task variants of several classifiers/regressors (logistic regression, support vector machine, K-nearest neighbor, Random Forest, ridge regression, support vector regression) convincingly outperform their single-task counterparts. We also show that the proposed model performs comparable or better than many state-of-the-art MTL and transfer learning baselines.

DeepCare: a deep dynamic memory model for predictive medicine

Personalized predictive medicine necessitates the modeling of patient illness and care processes, which inherently have long-term temporal dependencies. Healthcare observations, recorded in electronic medical records, are episodic and irregular in time. We introduce DeepCare, an end-to-end deep dynamic neural network that reads medical records, stores previous illness history, infers current illness states and predicts future medical outcomes. At the data level, DeepCare represents care episodes as vectors in space, models patient health state trajectories through explicit memory of historical records. Built on Long Short-Term Memory (LSTM), DeepCare introduces time parameterizations to handle irregular timed events by moderating the forgetting and consolidation of memory cells. DeepCare also incorporates medical interventions that change the course of illness and shape future medical risk. Moving up to the health state level, historical and present health states are then aggregated through multiscale temporal pooling, before passing through a neural network that estimates future outcomes. We demonstrate the efficacy of DeepCare for disease progression modeling, intervention recommendation, and future risk prediction. On two important cohorts with heavy social and economic burden -- diabetes and mental health -- the results show improved modeling and risk prediction accuracy.

Mhealth entrepreneurship: an exploratory research for a managerial model for mhealth start ups in low and middle income countries

Since some years, mobile technologies in healthcare (mHealth) stand for the transformational force to improve health issues in low- and middle-income countries (LMICs). Although several studies have identified the prevailing issue of inconsistent evidence and new evaluation frameworks have been proposed, few have explored the role of entrepreneurship to create disruptive change in a traditionally conservative sector. I argue that improving the effectiveness of mHealth entrepreneurs might increase the adoption of mHealth solutions. Thus, this study aims at proposing a managerial model for the analysis of mHealth solutions from the entrepreneurial perspective in the context of LMICs. I identified the Khoja–Durrani–Scott (KDS) framework as theoretical basis for the managerial model, due to its explicit focus on the context of LMICs. In the subsequent exploratory research I, first, used semi-structured interviews with five specialists in mHealth, local healthcare systems and investment to identify necessary adaptations to the model. The findings of the interviews proposed that especially the economic theme had to be clarified and an additional entrepreneurial theme was necessary. Additionally, an evaluation questionnaire was proposed. In the second phase, I applied the questionnaire to five start-ups, operating in Brazil and Tanzania, and conducted semi-structured interviews with the entrepreneurs to gain practical insights for the theoretical development. Three of five entrepreneurs perceived that the results correlated with the entrepreneurs' expectations of the strengths and weaknesses of the start-ups. Main shortcomings of the model related to the ambiguity of some questions. In addition to the findings for the model, the results of the scores were analyzed. The analysis suggested that across the participating mHealth start-ups the ‘behavioral and socio-technical’ outcomes were the strongest and the ‘policy’ outcomes were the weakest themes. The managerial model integrates several perspectives, structured around the entrepreneur. In order to validate the model, future research may link the development of a start-up with the evolution of the scores in longitudinal case studies or large-scale tests.

Invariant tests in an instrumental variables model with unknown data generating process

In this work we focus on tests for the parameter of an endogenous variable in a weakly identi ed instrumental variable regressionmodel. We propose a new unbiasedness restriction for weighted average power (WAP) tests introduced by Moreira and Moreira (2013). This new boundary condition is motivated by the score e ciency under strong identi cation. It allows reducing computational costs of WAP tests by replacing the strongly unbiased condition. This latter restriction imposes, under the null hypothesis, the test to be uncorrelated to a given statistic with dimension given by the number of instruments. The new proposed boundary condition only imposes the test to be uncorrelated to a linear combination of the statistic. WAP tests under both restrictions to perform similarly numerically. We apply the di erent tests discussed to an empirical example. Using data from Yogo (2004), we assess the e ect of weak instruments on the estimation of the elasticity of inter-temporal substitution of a CCAPM model.

An evaluation of the Business Model Canvas by technology-oriented startups in Brazil : a qualitative study about the usage of the tool by Brazilian entrepreneurs

The following paper was conducted with the support of several entrepreneurs and startups from Brazil. The aim of the research was to find out which impact the Business Model Canvas, further abbreviated as BMC, has on technology-oriented startups in Brazil. The first step of the study was identify some general concepts of entrepreneurship, as well as the conditions and environment of the country. Afterwards, it was focused on defining and comparing different business model tools and concepts to the BMC. After the literature review and meeting with several professionals in the area of entrepreneurship and startups, a questionnaire was formulated in order to conduct the qualitative study and identify the main impact of the tool. The questionnaire was answered by ten startups. In order to check the validity and credibility of the research outcomes, theory and investigator triangulation was used. As a result, the usage of the BMC could be evaluated by obtaining the outcomes and the theory, which showed that Brazilian tech startups are using Osterwalder’s model for the reason of idea creation and testing, validating and pivoting their business model. Interestingly, the research revealed that the entrepreneurs are using the tool often not in the traditional way of printing it, but rather applying it as a thinking approach. Besides, the entrepreneurs are focusing mostly on developing a strong Value Proposition, Customer Segment and sustainable Revenue Streams, while afterwards the remaining building blocks are built. Moreover, the research showed that the startups are using also other concepts, such as the Customer Development Process or Build-Measure-Learn Feedback Loop. These methodologies are often applied together with the BMC and helps to identify the most sustainable components of the business idea. Keywords: Business

An analysis of the perceived service quality of hospitality industry in Rio de Janeiro through the SERVQUAL model: a multiple case study

Brazil is under political and financial crises where the end seems far away. Because of that, researchers argue that the hotel rooms offered by Rio de Janeiro, built to host the Olympic Games 2016, will be difficult to occupy after the event. It is then necessary for the hotels to understand how guests perceive the service quality in order to adapt to this new era. If guests’ perceptions meet or exceed their expectations, they will be satisfied and will probably return. Thus based on the SERVQUAL approach, this paper aims to study the impact of the service dimensions on the guests’ overall satisfaction at hotels of Rio de Janeiro. Two hotels were considered representative of the city in terms of service quality and customers’ profile. Interviews to the hotel managers were performed, and questionnaires to the guests were administered. Among the five SERVQUAL dimensions – Reliability, Tangibles, Responsiveness, Assurance, and Empathy – the Empathy dimension appears to be the only one that affects the guests’ overall satisfaction. The study could also identify that gender, country of residence, home country and family income have an impact on guests’ satisfaction. This study has no intention of generalization, but rather of refining the theory about services and the SERVQUAL model.

Indústria de máquinas agrícolas no Brasil: um modelo para estimação da demanda de tratores para o triênio 2016 – 2018

O objetivo desta dissertação foi estimar a demanda de tratores agrícolas para o mercado brasileiro no triênio 2016-2018, utilizando-se para isto de técnicas de econometria de séries temporais, neste caso, modelos univariados da classe ARIMA e SARIMA e ou multivariados SARIMAX. Justifica-se esta pesquisa quando se observa a indústria de máquinas agrícolas no Brasil, dados os ciclos econômicos e outros fatores exógenos aos fundamentos econômicos da demanda, onde esta enfrenta muitos desafios. Dentre estes, a estimação de demanda se destaca, pois exerce forte impacto, por exemplo, no planejamento e custo de produção de curto e médio prazo, níveis de inventários, na relação com fornecedores de materiais e de mão de obra local, e por consequência na geração de valor para o acionista. Durante a fase de revisão bibliográfica foram encontrados vários trabalhos científicos que abordam o agronegócio e suas diversas áreas de atuação, porém, não foram encontrados trabalhos científicos publicados no Brasil que abordassem a previsão da demanda de tratores agrícolas no Brasil, o que serviu de motivação para agregar conhecimento à academia e valor ao mercado através deste. Concluiu-se, após testes realizados com diversos modelos que estão dispostos no texto e apêndices, que o modelo univariado SARIMA (15,1,1) (1,1,1) cumpriu as premissas estabelecidas nos objetivos específicos para escolha do modelo que melhor se ajusta aos dados, e foi escolhido então, como o modelo para estimação da demanda de tratores agrícolas no Brasil. Os resultados desta pesquisa apontam para uma demanda de tratores agrícolas no Brasil oscilando entre 46.000 e 49.000 unidades ano entre os anos de 2016 e 2018.

Bank networks and firm credit: an agent based model approach

Starting from the idea that economic systems fall into complexity theory, where its many agents interact with each other without a central control and that these interactions are able to change the future behavior of the agents and the entire system, similar to a chaotic system we increase the model of Russo et al. (2014) to carry out three experiments focusing on the interaction between Banks and Firms in an artificial economy. The first experiment is relative to Relationship Banking where, according to the literature, the interaction over time between Banks and Firms are able to produce mutual benefits, mainly due to reduction of the information asymmetry between them. The following experiment is related to information heterogeneity in the credit market, where the larger the bank, the higher their visibility in the credit market, increasing the number of consult for new loans. Finally, the third experiment is about the effects on the credit market of the heterogeneity of prices that Firms faces in the goods market.

Private label supply chains: a model for competitiveness?

The search for efficiency in supply chains has usually focused on logistic optimization aspects. Initiatives like the ECR are an example. This research questions the appropriateness of this focus comparing detailed cost structures of fifteen consumer products, covering five different product categories. It compares supply chains of private label products, presumably more efficient due to closer collaboration between chain members, to national brands supply chains. The major source of cost differences lies in other indirect costs incurred by the national brands and not directly assignable to advertising. Results indicate that a complete reconception of the supply chain, exploring different governance structures offers greater opportunities for cost savings than the logistic aspect in isolation. Research was done in the UK in 1995-1997, but results are only now publishable due to confidentiality agreements

Speculative bubbles and contagion: analysis of volatility's clusters during the DotCom bubble based on the dynamic conditional correlation model

Reviewing the de nition and measurement of speculative bubbles in context of contagion, this paper analyses the DotCom bubble in American and European equity markets using the dynamic conditional correlation (DCC) model proposed by (Engle and Sheppard 2001) as on one hand as an econometrics explanation and on the other hand the behavioral nance as an psychological explanation. Contagion is de ned in this context as the statistical break in the computed DCCs as measured by the shifts in their means and medians. Even it is astonishing, that the contagion is lower during price bubbles, the main nding indicates the presence of contagion in the di¤erent indices among those two continents and proves the presence of structural changes during nancial crisis

A model of competition and interbank network formation

This work aims to understand the interaction between competition and network formation in the banking market. Combining Matutes and Padilla (1994) and Matutes and Vives (2000), we build a model of imperfect bank competition for deposits in which an interbank relationship network is a key strategic decision: it affects banks’ profit and risk position. The competition level exerts influence in the banking network structure since it affects the network outcomes. As result, we have that different competition levels imply different network topologies. Specifically, greater competition imply denser networks. Finally, when we allow for the possibility of collusion, the denser network can come out in the least competitive environment.

A bidimensional model of matching in the marriage market with women labor decision

We construct a frictionless matching model of the marriage market where women have bidimensional attributes, one continuous (income) and the other dichotomous (home ability). Equilibrium in the marriage market determines intrahousehold allocation of resources and female labor participation. Our model is able to predict partial non-assortative matching, with rich men marrying women with low income but high home ability. We then perform numerical exercises to evaluate the impacts of income taxes in individual welfare and find that there is considerable divergence in the female labor participation response to taxes between the short run and the long run.