Estimativa dos parâmetros de compactação de solos usando técnicas de data mining

Dissertação de mestrado integrado em Engenharia Civil

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.

Intelligent decision support to predict patient barotrauma risk in intensive care units

The occurrence of Barotrauma is identified as a major concern for health professionals, since it can be fatal for patients. In order to support the decision process and to predict the risk of occurring barotrauma Data Mining models were induced. Based on this principle, the present study addresses the Data Mining process aiming to provide hourly probability of a patient has Barotrauma. The process of discovering implicit knowledge in data collected from Intensive Care Units patientswas achieved through the standard process Cross Industry Standard Process for Data Mining. With the goal of making predictions according to the classification approach they several DM techniques were selected: Decision Trees, Naive Bayes and Support Vector Machine. The study was focused on identifying the validity and viability to predict a composite variable. To predict the Barotrauma two classes were created: “risk” and “no risk”. Such target come from combining two variables: Plateau Pressure and PCO2. The best models presented a sensitivity between 96.19% and 100%. In terms of accuracy the values varied between 87.5% and 100%. This study and the achieved results demonstrated the feasibility of predicting the risk of a patient having Barotrauma by presenting the probability associated.

Generic system for human-computer gesture interaction: applications on sign language recognition and robotic soccer refereeing

Hand gestures are a powerful way for human communication, with lots of potential applications in the area of human computer interaction. Vision-based hand gesture recognition techniques have many proven advantages compared with traditional devices, giving users a simpler and more natural way to communicate with electronic devices. This work proposes a generic system architecture based in computer vision and machine learning, able to be used with any interface for human-computer interaction. The proposed solution is mainly composed of three modules: a pre-processing and hand segmentation module, a static gesture interface module and a dynamic gesture interface module. The experiments showed that the core of visionbased interaction systems could be the same for all applications and thus facilitate the implementation. For hand posture recognition, a SVM (Support Vector Machine) model was trained and used, able to achieve a final accuracy of 99.4%. For dynamic gestures, an HMM (Hidden Markov Model) model was trained for each gesture that the system could recognize with a final average accuracy of 93.7%. The proposed solution as the advantage of being generic enough with the trained models able to work in real-time, allowing its application in a wide range of human-machine applications. To validate the proposed framework two applications were implemented. The first one is a real-time system able to interpret the Portuguese Sign Language. The second one is an online system able to help a robotic soccer game referee judge a game in real time.

Observation and measurements of the production of prompt and non-prompt J/ψ mesons in association with a Z boson in pp collisions at s√ = 8 TeV with the ATLAS detector

The production of a Z boson in association with a J/ψ meson in proton--proton collisions probes the production mechanisms of quarkonium and heavy flavour in association with vector bosons, and allows studies of multiple parton scattering. Using 20.3fb−1 of data collected with the ATLAS experiment at the LHC, in pp collisions at s√=8 TeV, the first measurement of associated Z+J/ψ production is presented for both prompt and non-prompt J/ψ production, with both signatures having a significance in excess of 5σ. The inclusive production cross-sections for Z boson production (in μ+μ− or e+e− decay modes) in association with prompt and non-prompt J/ψ(→μ+μ−) are measured relative to the inclusive production rate of Z bosons in the same fiducial volume to be (88±16±6)×10−8 and (157±22±10)×10−8 respectively. Normalised differential production cross-sections are also determined as a function of the J/ψ transverse momentum. The fraction of signal events arising from single and double parton scattering is estimated, and a lower limit of 5.3 (3.7)mb at 68 (95) confidence level is placed on the effective cross-section regulating double parton interactions.

Assistive locomotion strategies for active lower limb devices

Dissertação de mestrado integrado em Engenharia Biomédica (área de especialização em Eletrónica Médica)

Metabolic profiling and classification of propolis samples from Southern Brazil: an NMR-based platform coupled with machine learning

The chemical composition of propolis is affected by environmental factors and harvest season, making it difficult to standardize its extracts for medicinal usage. By detecting a typical chemical profile associated with propolis from a specific production region or season, certain types of propolis may be used to obtain a specific pharmacological activity. In this study, propolis from three agroecological regions (plain, plateau, and highlands) from southern Brazil, collected over the four seasons of 2010, were investigated through a novel NMR-based metabolomics data analysis workflow. Chemometrics and machine learning algorithms (PLS-DA and RF), including methods to estimate variable importance in classification, were used in this study. The machine learning and feature selection methods permitted construction of models for propolis sample classification with high accuracy (>75%, reaching 90% in the best case), better discriminating samples regarding their collection seasons comparatively to the harvest regions. PLS-DA and RF allowed the identification of biomarkers for sample discrimination, expanding the set of discriminating features and adding relevant information for the identification of the class-determining metabolites. The NMR-based metabolomics analytical platform, coupled to bioinformatic tools, allowed characterization and classification of Brazilian propolis samples regarding the metabolite signature of important compounds, i.e., chemical fingerprint, harvest seasons, and production regions.

Sistema de controlo de processo e de organização numa empresa metalomecânica

Dissertação de mestrado em Engenharia Mecatrónica (área de especialização de Tecnologia de Manufatura)

Design of bio-based supramolecular structures through self-assembly of α-lactalbumin and lysozyme

Bovine α-lactalbumin (α-La) and lysozyme (Lys), two globular proteins with highly homologous tertiary structures and opposite isoelectric points, were used to produce bio-based supramolecular structures under various pH values (3, 7 and 11), temperatures (25, 50 and 75 °C) and times (15, 25 and 35 min) of heating. Isothermal titration calorimetry experiments showed protein interactions and demonstrated that structures were obtained from the mixture of α-La/Lys in molar ratio of 0.546. Structures were characterized in terms of morphology by transmission electron microscopy (TEM) and dynamic light scattering (DLS), conformational structure by circular dichroism and intrinsic fluorescence spectroscopy and stability by DLS. Results have shown that protein conformational structure and intermolecular interactions are controlled by the physicochemical conditions applied. The increase of heating temperature led to a significant decrease in size and polydispersity (PDI) of α-La–Lys supramolecular structures, while the increase of heating time, particularly at temperatures above 50 °C, promoted a significant increase in size and PDI. At pH 7 supramolecular structures were obtained at microscale – confirmed by optical microscopy – displaying also a high PDI (i.e. > 0.4). The minimum size and PDI (61 ± 2.3 nm and 0.14 ± 0.03, respectively) were produced at pH 11 for a heating treatment of 75 °C for 15 min, thus suggesting that these conditions could be considered as critical for supramolecular structure formation. Its size and morphology were confirmed by TEM showing a well-defined spherical form. Structures at these conditions showed to be stable at least for 30 or 90 days, when stored at 25 or 4 °C, respectively. Hence, α-La–Lys supramolecular structures showed properties that indicate that they are a promising delivery system for food and pharmaceutical applications.

Green's relations, regularity and abundancy for semigroups of quasi-onto transformations

Let V be an infinite-dimensional vector space and for every infinite cardinal n such that n≤dimV, let AE(V,n) denote the semigroup of all linear transformations of V whose defect is less than n. In 2009, Mendes-Gonçalves and Sullivan studied the ideal structure of AE(V,n). Here, we consider a similarly-defined semigroup AE(X,q) of transformations defined on an infinite set X. Quite surprisingly, the results obtained for sets differ substantially from the results obtained in the linear setting.

Screening for depression and anxiety disorders from pregnancy to postpartum with the EPDS and STAI

The Edinburgh Postnatal Depression Scale (EPDS) and the State Anxiety Inventory (STAI-S) are widely used self-report measures that still need to be further validated for the perinatal period. The aim of this study was to examine the screening performance of the EPDS and the STAI-S in detecting depressive and anxiety disorders at pregnancy and postpartum. Women screening positive on EPDS (EPDS ≥ 9) or STAI-S (STAI-S ≥ 45) during pregnancy (n = 90), as well as matched controls (n = 58) were selected from a larger study. At 3 months postpartum, 99 of these women were reassessed. At a second stage, women were administered a clinical interview to establish a DSM-IV-TR diagnosis. Receiver operator characteristics (ROC) analysis yielded areas under the curve higher than .80 and .70 for EPDS and STAI-S, respectively. EPDS and STAI-S optimal cut-offs were found to be lower at postpartum (EDPS = 7; STAI-S = 34) than during pregnancy (EPDS = 9; STAI-S = 40). EPDS and STAI-S are reasonably valid screening tools during pregnancy and the postpartum.

Anxiety and depression symptoms in women and men from early pregnancy to 3-months postpartum: parity differences and effects

This study aimed to investigate both anxiety and depression symptoms from early pregnancy to 3-months postpartum, comparing women and men and first and second-time parents. Methods: A sample of 260 Portuguese couples (N=520), first or second-time parents, recruited in an Obstetrics Out-patients Unit, filled in the State-Anxiety Inventory (STAI-S) and the Edinburgh Post-Natal Depression Scale (EPDS) at the 1st, 2nd and 3rd pregnancy trimesters, childbirth, and 3-months postpartum. Results: A decrease in anxiety and depression symptoms from early pregnancy to 3-months postpartum was found in both women and men, as well as in first and second-time parents. Men presented less anxiety and depression symptoms than women, but the same pattern of symptoms over time. Second-time parents showed more anxiety and depression symptoms than first-time parents and a different pattern of symptoms over time: an increase in anxiety and depression symptoms from the 3rd trimester to childbirth was observed in first-time parents versus a decrease in second-time parents. Limitations: The voluntary nature of the participation may have lead to a selection bias; women and men who agreed to participate could be those who presented fewer anxiety and depression symptoms. Moreover, the use of self-report symptom measures does not give us the level of possible disorder in participants. Conclusions: Anxiety and depression symptoms diminish from pregnancy to the postpartum period in all parents. Patterns of anxiety and depression symptoms from early pregnancy to 3-months postpartum are similar in women and men, but somewhat different in first and second time parents. Second-time parents should also be considered while studying and intervening during pregnancy and the postpartum.

Mother's anxiety and depression during the third pregnancy trimester and neonate's mother versus stranger's face/voice visual preference

Background: Neonates show visual preference for their mother's face/voice and shift their attention from their mother to a stranger's face/voice after habituation. Aim: To assess neonate's mother versus stranger's face/voice visual preference, namely mother's anxiety and depression during the third pregnancy trimester and neonate's: 1) visual preference for the mother versus the stranger's face/voice (pretest visual preference), 2) habituation to the mother's face/voice and 3) visual preference for the stranger versus the mother's face/voice (posttest visual preference). Method: Mothers (N=100) filled out the Edinburgh Postnatal Depression Scale (EPDS) and the State Anxiety Inventory (STAI) both at the third pregnancy trimester and childbirth, and the “preference and habituation to the mother's face/voice versus stranger” paradigm was administered to their newborn 1 to 5 days after childbirth. Results: Neonates of anxious/depressed mothers during the third pregnancy trimester contrarily to neonates of non-anxious/non-depressed mothers did not look 1) longer at their mother's than at the stranger's face/voice at the pretest visual preference (showing no visual preference for the mother), nor 2) longer at the stranger's face/voice in the posttest than in the pretest visual preference (not improving their attention to the stranger's after habituation). Conclusion: Infants exposed to mother's anxiety/depression at the third gestational trimester exhibit less perceptual/social competencies at birth.

Partner relationships during pregnancy in anxious and depressed women and men

To assess anxiety, depression and relationship satisfaction in both women and men during pregnancy, the State Anxiety Inventory (STAI), The Center for Epidemiological Studies-Depression Scale (CES-D) and The Relationship Questionnaire (RQ) were administered during the second trimester to a sample of 59 pregnant women and their partners. Anxious pregnant women rated their relationships as less positive. Depressed pregnant women also rated their relationships as less positive. The women’s anxiety scores were predictive of their positive and negative relationship scores. The women and their partners’ negative relationship scores were also predictive of each others’ negative relationship scores. These results highlight the importance of targeting anxiety as well as depression, and pregnant women as well as their partners in prenatal intervention programs.