Ontology-based representation and generation of workflows for micro-task human-machine computation

Doctoral Program in Computer Science

A deslocalização produtiva em Portugal: o caso de uma empresa do setor das cutelarias

Projeto de estágio de mestrado em Economia Industrial e da Empresa

Using data mining for prediction of hospital length of stay: an application of the CRISP-DM Methodology

Hospitals are nowadays collecting vast amounts of data related with patient records. All this data hold valuable knowledge that can be used to improve hospital decision making. Data mining techniques aim precisely at the extraction of useful knowledge from raw data. This work describes an implementation of a medical data mining project approach based on the CRISP-DM methodology. Recent real-world data, from 2000 to 2013, were collected from a Portuguese hospital and related with inpatient hospitalization. The goal was to predict generic hospital Length Of Stay based on indicators that are commonly available at the hospitalization process (e.g., gender, age, episode type, medical specialty). At the data preparation stage, the data were cleaned and variables were selected and transformed, leading to 14 inputs. Next, at the modeling stage, a regression approach was adopted, where six learning methods were compared: Average Prediction, Multiple Regression, Decision Tree, Artificial Neural Network ensemble, Support Vector Machine and Random Forest. The best learning model was obtained by the Random Forest method, which presents a high quality coefficient of determination value (0.81). This model was then opened by using a sensitivity analysis procedure that revealed three influential input attributes: the hospital episode type, the physical service where the patient is hospitalized and the associated medical specialty. Such extracted knowledge confirmed that the obtained predictive model is credible and with potential value for supporting decisions of hospital managers.

Using customer lifetime value to improve the prediction of bank deposit subscription in telemarketing campaigns

Customer lifetime value (LTV) enables using client characteristics, such as recency, frequency and monetary (RFM) value, to describe the value of a client through time in terms of profitability. We present the concept of LTV applied to telemarketing for improving the return-on-investment, using a recent (from 2008 to 2013) and real case study of bank campaigns to sell long- term deposits. The goal was to benefit from past contacts history to extract additional knowledge. A total of twelve LTV input variables were tested, un- der a forward selection method and using a realistic rolling windows scheme, highlighting the validity of five new LTV features. The results achieved by our LTV data-driven approach using neural networks allowed an improvement up to 4 pp in the Lift cumulative curve for targeting the deposit subscribers when compared with a baseline model (with no history data). Explanatory knowledge was also extracted from the proposed model, revealing two highly relevant LTV features, the last result of the previous campaign to sell the same product and the frequency of past client successes. The obtained results are particularly valuable for contact center companies, which can improve pre- dictive performance without even having to ask for more information to the companies they serve.

Out-of-plane behavior of masonry infill walls

In order to investigate the out-of-plane behaviour of masonry infill walls, quasi-static testing was performed on a masonry infill walls built inside a reinforced concrete frame by means of an airbag system to apply the uniform out-of-plane load to each component of the infill. The main advantage of this testing setup is that the out-of-plane loading can be applied more uniformly in the walls, contrarily to point load configuration. The test was performed under displacement control by selecting the mid-point of the infill as control point. Input and output air in the airbag was controlled by using a software to apply a specific displacement in the control point of the infill wall. The effect of the distance between the reaction frame of the airbag and the masonry infill on the effective contact area was previously analysed. Four load cells were attached to the reaction frame to measure the out-of-plane force. The effective contact area of the airbag was calculated by dividing the load measured in load cells by the pressure inside the airbag. When the distance between the reaction walls and the masonry infill wall is smaller, the effective area is closer to the nominal area of the airbag. Deformation and crack patterns of the infill confirm the formation of arching mechanism and two-way bending of the masonry infill. Until collapse of the horizontal interface between infill and upper beam in RC frame, the infill bends in two directions but the failure of that interface which is known as weakest interface due to difficulties in filling the mortar between bricks of last row and upper beam results in the crack opening trough a well-defined path and the consequent collapse of the infill.

In-plane and out-of plane experimental characterization of rc masonry infilled frames

Seismic investigations of typical south European masonry infilled frames were performed by testing two reduced scale specimens: one in the in-plane direction and another in the out-ofplane direction. Information about geometry and reinforcement scheme of those structures constructed in 1980s were obtained by [1]. The specimen to be tested in the in-plane direction was constructed as double leaf masonry while the specimen for testing in the out-of-plane direction is constructed with only its exterior leaf since the recent earthquakes have highlighted the vulnerability of the external leaf of the infills in out-of-plane direction [2]. The tests were performed by applying the pre-defined values of displacements in the in-plane and out-of-plane directions in the control points. For in-plane testing it was done by hydraulic actuator and for out-of-plane testing through the application of an airbag. Input and output air in the airbag was controlled by using a software to apply a specific displacement in the control point of the infill wall. Mid-point of the infill was assumed as a control point for outof- plane testing. Deformation and crack patterns of the infill confirm the formation of two-way arching mechanism of the masonry infill until collapse of the upper horizontal interface between infill and frame which is known as weakest interface due to difficulties in filling the mortar between bricks of last row and upper beam. This results in the crack opening through a welldefined path and the consequent collapse of the infill.

Interaction paradigms versus age-related user profiles: an evaluation on content selection

Novel input modalities such as touch, tangibles or gestures try to exploit human's innate skills rather than imposing new learning processes. However, despite the recent boom of different natural interaction paradigms, it hasn't been systematically evaluated how these interfaces influence a user's performance or whether each interface could be more or less appropriate when it comes to: 1) different age groups; and 2) different basic operations, as data selection, insertion or manipulation. This work presents the first step of an exploratory evaluation about whether or not the users' performance is indeed influenced by the different interfaces. The key point is to understand how different interaction paradigms affect specific target-audiences (children, adults and older adults) when dealing with a selection task. 60 participants took part in this study to assess how different interfaces may influence the interaction of specific groups of users with regard to their age. Four input modalities were used to perform a selection task and the methodology was based on usability testing (speed, accuracy and user preference). The study suggests a statistically significant difference between mean selection times for each group of users, and also raises new issues regarding the “old” mouse input versus the “new” input modalities.

Analysis of output properties of the Peroxiredoxin/Thioredoxin/Thioredoxin reductase system

Dissertação de mestrado em Bioinformática

High strain rate constitutive modeling forhistorical structures subjected to blast loading

Doctoral Thesis Civil Engineering

Structure-property relationships in polymer nanocomposites

Tese de Doutoramento em Ciência e Engenharia de Polímeros e Compósitos

Measuring end-to-end delay in real-time auralisation systems

One of the major challenges in the development of an immersive system is handling the delay between the tracking of the user’s head position and the updated projection of a 3D image or auralised sound, also called end-to-end delay. Excessive end-to-end delay can result in the general decrement of the “feeling of presence”, the occurrence of motion sickness and poor performance in perception-action tasks. These latencies must be known in order to provide insights on the technological (hardware/software optimization) or psychophysical (recalibration sessions) strategies to deal with them. Our goal was to develop a new measurement method of end-to-end delay that is both precise and easily replicated. We used a Head and Torso simulator (HATS) as an auditory signal sensor, a fast response photo-sensor to detect a visual stimulus response from a Motion Capture System, and a voltage input trigger as real-time event. The HATS was mounted in a turntable which allowed us to precisely change the 3D sound relative to the head position. When the virtual sound source was at 90º azimuth, the correspondent HRTF would set all the intensity values to zero, at the same time a trigger would register the real-time event of turning the HATS 90º azimuth. Furthermore, with the HATS turned 90º to the left, the motion capture marker visualization would fell exactly in the photo-sensor receptor. This method allowed us to precisely measure the delay from tracking to displaying. Moreover, our results show that the method of tracking, its tracking frequency, and the rendering of the sound reflections are the main predictors of end-to-end delay.

Gd2O3: Eu Nanoparticle-Based Poly(vinylidene fluoride) Composites for Indirect X-ray Detection

Polymer based scintillator composites have been fabricated by combining poly(vinylidene fluoride) (PVDF) and Gd2O3:Eu nanoparticles (50nm). PVDF has been used since it is a flexible and stable binder matrix and highly resistance to thermal and light deterioration. Gd2O3:Eu has been selected as scintillator material due to its wide band gap, high density and suitable visible light yield. The structural, mechanical, thermal and electrical characteristics of the composites were studied as a function of filler content, together with their performance as scintillator material. The introduction of Gd2O3:Eu nanoparticles into the PVDF matrix does not influence the morphology of the polymer or the degree of crystallinity. On the other hand, an increase of the Young´s modulus with respect to PVDF matrix is observed for filler contents of 0.1-0.75 wt.%. The introduction of Gd2O3:Eu into the PVDF matrix increases dielectric constant and DC electrical conductivity as well as the visible light yield in the nanocomposite, being this increase dependent upon Gd2O3:Eu content and X-ray input power. In this way, Gd2O3:Eu/PVDF composites shows suitable characteristics to be used as X-ray radiation transducers, in particular for large area applications.

Aquisição de verbos da língua portuguesa, no primeiro ciclo de escolaridade, por crianças bilingues luso-alemãs e crianças monolingues alemãs

Tese de Doutoramento em Ciências da Linguagem (área de especialização em Linguística Aplicada).

Simulation of a rectifier malfunction on a offshore wind system with HVDC transmission

A new integrated mathematical model for the simulation of an offshore wind system having a rectifier input voltage malfunction at one phase is presented in this paper. The mathematical model considers an offshore variable-speed wind turbine on a floating platform, equipped with a permanent magnet synchronous generator using full-power three-level converter to inject energy into the electric network, through a high voltage direct current transmission submarine cable. The model for the drive train is a discrete three mass, incorporating the dynamic of the moving surface. A case study is presented to access conclusion about the malfunction.

Impacto do azoto no processo de soldadura reflow

Dissertação de mestrado em Engenharia Industrial (área de especialização em Qualidade, Segurança e Manutenção)