Today it is important for investors to identify which stocks that will result in positive returns in order for the right decision to be made when trading on the stock market. For decades it has been an area of interest for academics, and it is still challenging due to many difficulties and problems. A large number of studies has been carried out in machine learning and stock trading,where many of the studies has resulted in promising results despite these challenges. The aim of this study was to develop and evaluate predictive models for identifying stocks that outperform the Swedish market index OMXSPI. The research utilized a dataset of historical stock data and applied three various machine learning algorithms, Support Vector Machine, Logistic Regression and Decision Trees to predict if excess performance was met. With the help of ten-fold cross-validation and hyperparameter tuning the results were an IT-artefact that produced satisfying results. The results showed that hyperparameter tuning techniques marginally improved the metrics focused-on, namely accuracy and precision. The support vector machine model achieved an accuracy of 58,52% and a precision of 57,51%. The logistic regression model achieved an accuracy of 55,75% and a precision of 54,81%. Finally, the decision tree model which was the best performer, achieved an accuracy of 64,84% and a precision of 65,00%.

No description available.

Die ASFiNAG unterhält mehr als 5.000 Brückenobjekte im Straßennetz. Die meisten von ihnen wurden in Stahlbeton- oder Spannbetonbauweise errichtet und insbesondere die Stützen sind hohen Belastungen durch Chlorid ausgesetzt. Dies führt langfristig zu einer Verringerung der Dauerhaftigkeit und Instandsetzungen sind daher erforderlich. Eine genaue Ermittlung der Depassivierung und ein Prognosemodell sind bei Chloridbelastung besonders wichtig für eine kosteneffiziente Erhaltung. Die neue Methode LA-ICP-MS (Laser Ablation Inductively Coupled Plasma Mass Spectrometry) ermöglicht exaktere Ergebnisse als die traditionelle Chloridbestimmung mittels Titration. Die Vorteile der beiden Methoden werden im Beitrag diskutiert und die Ergebnisse verglichen. Es wird gezeigt, dass die LA-ICP-MS-Methode zuverlässigere Ergebnisse liefert und durch die hohe Ortsauflösung der Chloridprofile sowie die genaue Bestimmung des Chloridgehalts noch nicht dagewesene Möglichkeiten für eine exakte Beurteilung und Prognose eröffnet.

Small businesses growth and profitability is a national problem in South Africa (SA) and the Emerging Contractors (ECs) in the construction industry in Gauteng Province are no exception. The SA government expect ECs to become the main players in economic transformation and growth, job creation and income generation. However, ECs fail to live up to the expectations as some are characterised by low performance levels and are at risk of business closure. The purpose of this research is to examine the perceptions of Small Business Executives (SBEs) on determinants of business performance so that a performance model that could guide ECs within the construction business is developed. The researcher used a mixed approach to examine the perceptions of SBEs on determinants of ECs’ performance by applying both qualitative and quantitative research approaches. The population from which the sample was drawn comprised of 1890 urban based ECs registered with CIDB in the construction industry in Gauteng. Systematic random sampling was utilized and a sample size of 501 ECs was used. A structured questionnaire and interview guide were administered as the data collection instruments. A pilot survey was conducted where 2 key informants (SBEs) and 15 respondents (SBEs) were interviewed as representatives of ECs. A survey was used to assess the four hypotheses in the study. Content analysis was used to identify themes that emerged from qualitative data. The analysis of qualitative data demonstrated the perceptions of SBEs that financial factors, manpower, materials, machinery and equipment, project implementation, quality of work, legal and environmental and strategic planning were the main determinants that should be addressed when setting up and running construction businesses to reduce the negative impact on business performance. Based on the perceptions of SBEs, in quantitative, the univariate and multivariate statistical methods were performed to measure the level of significance between and among the performance determinants were tested at 5% confidence interval. Multiple linear regressions carried out based on the perceptions of SBEs identified strategic planning, project implementation and project performance as the most statistically significant factors in predicting the performance of ECs in Gauteng Province. The results from both qualitative and quantitative methods were synthesised and analysed through data transformation, typological development and case study analysis and the perceptions of respondents indicated a strong agreement between the qualitative and quantitative results. Based on multivariate results which were a result of the perceptions of SBEs, a performance prediction model was developed which is the core contribution of the study. The structure of the model, how it is used, its advantages and disadvantages were presented. The model provides a platform upon which ECs could predict performance and this would benefit ECs, clients, community and policymakers. In conclusion, the study established three significant determinants that include strategic planning, project implementation and project performance as highly relevant variables to improve construction business performance. Based on the perceptions of the respondents, the study recommends ECs to acquire and develop financial management skills that would enable them to understand and manage financial requirements that stakeholders such as financial institutions need in the areas of budgeting and cash flow management, bookkeeping knowledge, financial policies and controls and project pricing strategy. On the shortage of skills, the study recommends SBEs to take an initiative role in training employees, motivating and retaining competent workforce. The government need to establish technical colleges for training workers (apprenticeship) in order to equip them with appropriate industry knowledge, skills and experience. Lack of collateral security was widely reported by most SBEs and the study recommends the formation of a sector bank that would support emerging contractor businesses. / Business Management / D.B.L.

Die Überlebenswahrscheinlichkeit der Unfallopfer nach Straßenverkehrsunfällen hängt von der Zeitspanne vom Unfall bis zur ersten suffizienten Hilfeleistung (therapiefreies Intervall) sowie von der Qualität der Erstversorgung ab. Das therapiefreie Intervall kann durch automatische Notrufsysteme (ACN) und suffiziente Laienhilfe verkürzt werden. In dieser Dissertation werden (1) Modelle zur Prognose der Verletzungsschwere von Fahrzeuginsassen für den Einsatz in erweiterten automatischen Notrufsystemen (AACN) entwickelt und (2) Anforderungen an Telemedizinsysteme zur Unterstützung von Laien bei der Durchführung von Erste-Hilfe-Maßnahmen untersucht. (1) Die Prognose der Verletzungsschwere ist für eine optimale Disposition der Rettungsmittel erforderlich. Auf der Basis der amerikanischen Unfalldatenbank NASS/ CDS und der deutschen Unfalldatenbank GIDAS werden Prognosemodelle mit multivariaten inferenzstatistischen Verfahren (decision tree, logistic regression) entwickelt, die Bedeutung einzelner Eingangsgrößen untersucht und die Leistungsfähigkeit der Modelle überprüft. Die Leistungsfähigkeit beider statistischen Verfahren stellt sich in ROC-Analysen als vergleichbar heraus. Mit den Prognosemodellen kann eine Dispositionsempfehlung erstellt werden, die heutigen Dispositionsentscheidungen auf der Grundlage eines Telefongesprächs mit dem Unfallmelder qualitativ entspricht. (2) Zur Entwicklung eines Telemedizinsystems für die Laienunterstützung wird untersucht, ob die telemedizinische Anleitung von Laien bei der Durchführung von Erste-Hilfe-Maßnahmen zu einer Erhöhung der Hilfsbereitschaft führt und welche Anforderungen ein Telemedizinsystem erfüllen muss. Dazu werden die am häufigsten indizierten Erste-Hilfe-Maßnahmen nach Verkehrsunfällen durch Auswertung der GIDAS-Datenbank identifiziert, der Stand der heutigen Laienhilfe analysiert, die erforderlichen Funktionen eines Telemedizinsystems abgeleitet und der Einfluss einer verbesserten Laienhilfe auf die Überlebenswahrscheinlichkeit bestimmt. / The probability of survival of motor vehicle accident victims depends on two factors: the time interval between the accident and the initial provision of appropriate first-aid (therapy-free interval), and the quality of the first-aid treatment. The therapy-free interval can be shortened significantly by automatic crash notification systems (ACN) and by adequate bystander first aid. In this thesis, (1) models are developed for predicting the severity of injury of vehicle occupants for use in advanced automatic crash notification systems (AACN) and (2) requirements of telemedical bystander support systems are examined. (1) The prediction of the severity of injury is needed for optimal disposition of rescue services. Therefore, prediction models are developed on the basis of the American accident database NASS/ CDS and its German counterpart, GIDAS, using multivariate inference statistical procedures (decision tree, logistic regression). In addition, the importance of individual parameters and the efficiency of these models are investigated. The performance of the two statistical procedures proves to be equal by using the ROC analyses. The prediction models allow for recommendations of rescue services that correspond in quality to contemporary dispatching decisions for rescue services on the basis of a telephone call between the individual reporting the accident and a dispatcher. (2) Concerning the development of a telemedical bystander support system, the paper examines whether the availability of telemedical guidance leads to an increase in bystander willingness to provide first aid. In order to evaluate the requirements of a telemedical system, the most frequently indicated first-aid measures after traffic accidents are identified by evaluating the GIDAS database. Furthermore, the status of assistance currently offered by bystanders is analyzed, the necessary functions of a telemedical system are scrutinized, and the influence of improved bystander assistance on the probability of accident victims’ survival is determined.

La protection des documents fiduciaires et identitaires contre la fraude exige le développement d’outils de contrôle fondés sur des effets visuels sans cesse renouvelés, difficiles à contrefaire (même pour un expert ... de la contrefaçon !). Ce projet de recherche s’inscrit dans cette problématique et vise à apporter des solutions originales via l’impression de supports diffusants d’une part, et le développement de modèles de rendu visuel d’autre part. Les effets visuels recherchés sont des ajustements de couleurs entre les deux faces d’un imprimé lorsque celui-ci est observé par transparence devant une source lumineuse. Pour obtenir facilement des ajustements de couleurs quelles que soient les couleurs visées, il est capital d’avoir un modèle à disposition, permettant de calculer les quantités d’encre à déposer. Un modèle doit être capable de prédire les facteurs spectraux de réflexion et de transmission du support imprimé en décrivant les phénomènes de diffusion optique présents en pratique dans les couches d’encre et le support. Nous nous intéressons plus particulièrement aux imprimés translucides contenant des couleurs en demi-ton des deux côtés de la surface avec pour objectif de prédire le rendu visuel pour diverses configurations d’observation. Pour cela, nous proposons une nouvelle approche basée sur l’utilisation de matrices de transfert de flux pour prédire les facteurs spectraux de réflexion et de transmission des imprimés lorsqu’ils sont éclairés simultanément des deux côtés. En représentant le comportement optique des différents composants d’un imprimé par des matrices de transfert, la description des transferts de flux entre ces composantes s’en trouve simplifiée. Ce cadre mathématique mène à la construction de modèles de prédiction de couleurs imprimées en demi-ton sur des supports diffusants. Nous montrons par ailleurs que certains modèles existants, comme le modèle de Kubelka-Munk ou encore le modèle de Clapper-Yule, peuvent également être formulés en termes de matrices de transfert. Les résultats obtenus avec les modèles proposés dans ce travail mettent en évidence des qualités de prédiction équivalentes, voire supérieures, à celles qu’on retrouve dans l’état de l’art, tout en proposant une simplification de la formulation mathématique et de la description physique des échanges de flux. Cette simplification fait de ces modèles des outils de calcul qui s’utilisent très facilement, notamment pour la détermination des quantités d’encre à déposer sur les deux faces de l’imprimé afin d’obtenir des ajustements de couleurs / The protection of banknotes or identity documents against counterfeiting demands the development of control tools based on visual effects that are continuously renewed. These visual effects become thus difficult to counterfeit even by an expert forger ! This research tries to deal with that issue. Its objective is to bring new solutions using on the one side, the printing of diffusing materials, and on the other side the development of visual rendering models that can be observed. The visual effects that are sought-after are the color matching on both sides of a printed document when observed against thelight. To easily obtain a color matching, whatever the colors that are aimed for, it is essential to have a model that helps in calculating the quantity of ink to be left on the document. A model must be used to predict the spectral reflectance and the transmittance factors of the printed document by describing the phenomena of optical diffusion really present in the ink layers and in the document. We shall focus our interest especially on translucent printed documents that have halftone colors on both sides. Our goal here is to predict the visual rendering in different configurations of observation. To that end, we are offering a new approach based on the use of flux transfer matrices to predict the spectral reflectance and transmittance factors of prints when they are simultaneously lit up on both sides. By representing with transfer matrices the optical behavior of the different components present in a printed document, we see that the description of flux transfer between these elements is thus simplified. This mathematical framework leads to the construction of prediction models of halftone printed colors on diffusing materials. We also show that some existing models, such as the Kubelka-Munk or the Clapper-Yule models, can also be formulated in transfer matrices terms. The results that we get with the models used in this work make apparent identical prediction quality and in some cases even better ones to the ones found in the state of the art, while offering a simplification of the mathematical formulation and the physical description of the flux transfer. This simplification thus transforms these models into calculation tools that can easily be used especially for the choice of quantities of ink that must be left on both sides of the document in order to obtain color matching

The aim of this thesis was to develop a touchscreen-adapted version of the well established Keystroke-level model, which is a user performance model designed to be a quick and easy way to evaluate user interfaces prior to creating prototypes. A quantitative research method was chosen to measure general execution times for common touchscreen gestures on mobile devices. Over 1000 data points were collected from participants who partook in a observational experiment using a prototype specifically programmed for this study. The results of the thesis involves the analyzed measurements acquired from the experiment, as well as a four important aspect to consider when performing similar experiments, namely: (1) The individual speed of the participants; (2) The participants’ level of expertise; (3) The participants’ methods of performing the gestures; (4) Designing the experiment for medium difficulty. / Målet med denna uppsats var att utveckla en pekskärmsanpassad version av den välkända modellen Keystroke-level model, vilket är en modell som kan användas för att snabbt och enkelt utvärdera gränssnitt utan att behöva utveckla fungerande prototyper. En kvantitativ forskningsmetod användes för att mäta hur lång tid det tar för användare generellt att utföra de vanligaste fingergesterna på mobila pekskärmar. Över 1000 datapunkter samlades in från ett experiment där deltagarna fick använda en prototyp som automatiskt mätte den tid det tog för användarna att utföra de olika fingergesterna. Resultaten av studien består av de mätvärden som räknats ut samt fyra aspekter som är viktiga att ha i beaktning när liknande experiment ska utformas. Dessa är: (1) Varje deltagares individuella hastighet; (2) Deltagarnas pekskärmsvana; (3) Sättet de olika deltagarna utför fingergesterna; (4) Utforma experimentet för att representera en medium svårighetsgrad. / Le but de ce projet fut de créer une version du Keystroke-level model adaptée aux écrans tactiles. Ce modèle de performance utilisateur est une méthode d’évaluation simple et rapide d’interfaces utilisateur avant la création de prototypes. Une étude quantitative fut choisie pour mesurer le temps d’exécution général de mouvements spécifiques effectués sur des écrans tactiles de téléphones. Plus de 1000 points de données furent collectés auprès de participants ayant pris part à une étude observationnelle utilisant un prototype spécialement programmé pour l’occasion. Les résultats de ce projet incluent l’analyse des mesures acquises lors de cette étude ainsi que quatre aspects importants à considérer lors de la réalisation d’études similaires, à savoir: (1) La vitesse individuelle des participants; (2) Le niveau d’habitude d’utilisation d’écrans tactiles des participants; (3) Les méthodes utilisées par les participants lors de la réalisation des mouvements; (4) La conception d’étude pour des mouvements de difficulté dite moyenne.

Small businesses growth and profitability is a national problem in South Africa (SA) and the Emerging Contractors (ECs) in the construction industry in Gauteng Province are no exception. The SA government expect ECs to become the main players in economic transformation and growth, job creation and income generation. However, ECs fail to live up to the expectations as some are characterised by low performance levels and are at risk of business closure. The purpose of this research is to examine the perceptions of Small Business Executives (SBEs) on determinants of business performance so that a performance model that could guide ECs within the construction business is developed. The researcher used a mixed approach to examine the perceptions of SBEs on determinants of ECs’ performance by applying both qualitative and quantitative research approaches. The population from which the sample was drawn comprised of 1890 urban based ECs registered with CIDB in the construction industry in Gauteng. Systematic random sampling was utilized and a sample size of 501 ECs was used. A structured questionnaire and interview guide were administered as the data collection instruments. A pilot survey was conducted where 2 key informants (SBEs) and 15 respondents (SBEs) were interviewed as representatives of ECs. A survey was used to assess the four hypotheses in the study. Content analysis was used to identify themes that emerged from qualitative data. The analysis of qualitative data demonstrated the perceptions of SBEs that financial factors, manpower, materials, machinery and equipment, project implementation, quality of work, legal and environmental and strategic planning were the main determinants that should be addressed when setting up and running construction businesses to reduce the negative impact on business performance. Based on the perceptions of SBEs, in quantitative, the univariate and multivariate statistical methods were performed to measure the level of significance between and among the performance determinants were tested at 5% confidence interval. Multiple linear regressions carried out based on the perceptions of SBEs identified strategic planning, project implementation and project performance as the most statistically significant factors in predicting the performance of ECs in Gauteng Province. The results from both qualitative and quantitative methods were synthesised and analysed through data transformation, typological development and case study analysis and the perceptions of respondents indicated a strong agreement between the qualitative and quantitative results. Based on multivariate results which were a result of the perceptions of SBEs, a performance prediction model was developed which is the core contribution of the study. The structure of the model, how it is used, its advantages and disadvantages were presented. The model provides a platform upon which ECs could predict performance and this would benefit ECs, clients, community and policymakers. In conclusion, the study established three significant determinants that include strategic planning, project implementation and project performance as highly relevant variables to improve construction business performance. Based on the perceptions of the respondents, the study recommends ECs to acquire and develop financial management skills that would enable them to understand and manage financial requirements that stakeholders such as financial institutions need in the areas of budgeting and cash flow management, bookkeeping knowledge, financial policies and controls and project pricing strategy. On the shortage of skills, the study recommends SBEs to take an initiative role in training employees, motivating and retaining competent workforce. The government need to establish technical colleges for training workers (apprenticeship) in order to equip them with appropriate industry knowledge, skills and experience. Lack of collateral security was widely reported by most SBEs and the study recommends the formation of a sector bank that would support emerging contractor businesses. / Business Management / D.B.L.

La dégénérescence maculaire liée à l’âge (DMLA) est la première cause de cécité dans les pays industrialisés. C’est une maladie complexe et multifactorielle ayant des conséquences majeures sur la qualité de vie des personnes atteintes. De nombreux facteurs de risque, génétiques et non génétiques, jouent un rôle important dans la pathogénèse des stades avancés de la DMLA. Les modèles de prédiction développés à ce jour reposent sur un nombre limité de ces facteurs, et sont encore peu utilisés dans la pratique clinique.Ce travail de thèse avait pour premier objectif d’identifier de nouveaux biomarqueurs circulants du risque de DMLA. Ainsi, à partir d’une étude post-mortem basée sur une approche de lipidomique, nous avons identifié les composés lipidiques sanguins les plus prédictifs des concentrations rétiniennes en acides gras polyinsaturés omégas 3 (AGPI w-3). Nous avons développé un modèle de prédiction basé sur 7 espèces de lipides des esters de cholestérol. Ce modèle, obtenu en combinant pénalisation et réduction de la dimension, a ensuite été validé dans des études cas-témoins de DMLA et dans un essai clinique randomisé de supplémentation en AGPI w-3. Ces biomarqueurs pourraient être utiles pour l’identification des personnes à haut risque de DMLA, qui pourraient ainsi bénéficier d’une supplémentation en AGPI w-3.Le deuxième objectif de cette thèse était de développer un modèle de prédiction du risque de progression vers une DMLA avancée à partir de facteurs de risque génétiques, phénotypiques et environnementaux. Une originalité de notre travail a été d’utiliser une méthode de régression pénalisée – un algorithme d’apprentissage automatique – dans un cadre de survie afin de tenir compte de la multicollinéarité entre les facteurs de risque. Nous avons également pris en compte la censure par intervalle et le risque compétitif du décès via un modèle à 3 états sain-malade-mort. Nous avons ensuite validé ce modèle sur une étude indépendante en population générale.Il serait intéressant de valider ce modèle de prédiction dans d’autres études indépendantes en y incluant les biomarqueurs circulants identifiés à partir de l’étude de lipidomique effectuée dans le cadre de cette thèse. Le but final serait d’intégrer cet outil prédictif dans la pratique clinique afin de rendre la médecine de précision une réalité pour les patients atteints de DMLA dans le futur proche. / Age-related macular degeneration (AMD) is the leading cause of blindness in industrialized countries. AMD is a complex and multifactorial disease with major consequences on the quality of life. Numerous genetic and non-genetic risk factors play an important role in the pathogenesis of the advanced stages of AMD. Existing prediction models rely on a restricted set of risk factors and are still not widely used in the clinical routine.The first objective of this work was to identify new circulating biomarkers of AMD’s risk using a lipidomics approach. Based on a post-mortem study, we identified the most predictive circulating lipids of retinal content in omega-3 polyunsaturated fatty acids (w-3 PUFAs). We combined penalization and dimension reduction to establish a prediction model based on plasma concentration of 7 cholesteryl ester species. We further validated this model on case-control and interventional studies. These biomarkers could help identify individuals at high risk of AMD who could be supplemented with w-3 PUFAs.The second objective of this thesis was to develop a prediction model for advanced AMD. This model incorporated a wide set of phenotypic, genotypic and lifestyle risk factors. An originality of our work was to use a penalized regression method – a machine learning algorithm – in a survival framework to handle multicollinearities among the risk factors. We also accounted for interval censoring and the competing risk of death by using an illness-death model. Our model was then validated on an independent population-based cohort.It would be interesting to integrate the circulating biomarkers identified in the lipidomics study to our prediction model and to further validate it on other external cohorts. This prediction model can be used for patient selection in clinical trials to increase their efficiency and paves the way towards making precision medicine for AMD patients a reality in the near future.

No description available.