DEFENDED THESES IN 2017
Samedi HENG, UCLouvain, LouRIM/CEMIS
Impact of Unified User-Story-Based Modeling on Agile Methods: Aspects on Requirements, Design and Life Cycle Management
SUPERVISORS:
Manuel KOLP, UCLouvain
Yves WAUTELET, KULeuven
JURY :
Jean VANDERDONCKT, UCLouvain
Isabelle MIRBEL, Université Nice Sophia Antipolis
Vincent ENGLEBERT, UNamur
CHAIRMAN :
Per AGRELL, UCLouvain
https://cdn.uclouvain.be/groups/cms-editors-LouRIM/Samedi_HENG_Annonce_soutenance_publique_.pdf
ABSTRACT:
User stories (US) are the most used requirements artifacts within agile methods such as XP and Scrum. They are written in natural language using prose or following a specific template. In practice, many templates have been proposed with no semantic for each syntax used in these templates. Therefore, this thesis has studied US templates with the following structure: As [WHO], I want [WHAT], so that [WHY]. We have proposed a unified model for US templates and provided semantic of each syntax. We have also proposed a goal-based model called Rationale Tree (RT) aiming at analyzing US sets. It allows identifying dependencies between US, Epic US and Themes. In addition, the RT provides a global view of the systems which allows to improve the planning activities, and therefore the development life-cycle of agile methods. As an alternative to RT, we have also proposed a US model based on the Use-Case model that allows US to be used within non-agile projects and fill the gaps between US and Object-Oriented design.
Mehdi MADANI, UCLouvain, IMMAQ/CORE
Revisiting European day-ahead electricity market auctions: MIP models and Algorithms.
SUPERVISOR(s):
Mathieu VAN VYVE, UCLouvain
JURY:
François GLINEUR, UCLouvain
Anthony PAPAVASILIOU, UCLouvain
Carlos RUIZ, Universidad Carlos III de Madrid
Bertrand CORNELUSSE, Université de Liège
CHAIRMAN:
Per AGRELL, UCLouvain
ABSTRACT:
In Europe, orders are submitted to power exchanges to sell or buy substantially large amounts of electricity for the next day. The orders include so-called non-convexities rendering the underlying microeconomic optimization problem `non-convex’, departing from more classical assumptions in microeconomic theory. Most of the power exchanges are integrated at a European level under the Price Coupling of Region project. Uniform prices are computed, in the sense that every market participant of a same market segment (location and hour of the day) will pay or receive the same electricity price and no other transfers or payments are considered. This is done at the expense of having some bids “paradoxically rejected” in the market outcome, as for the computed market prices, some bids propose a price which is “good enough” but are yet rejected. It is also at the expense of welfare optimality, as most of the time, no welfare optimal solution can be supported by uniform prices such that no financial losses are incurred. The present thesis proposes mixed integer programming models and algorithms for such non-convex uniform price auctions. In particular, a new bidding product is proposed which generalizes both block orders used in the Central Western Europe Region (France, Germany, Belgium, the Netherlands, etc) or Nord Pool (coupling Northern countries), and, mutatis mutandis, complex orders with a minimum income condition as used in Spain and Portugal.
Mathieu ZEN, UCLouvain, LouRIM/PRISME
A Methodology for Assessing Aesthetics of a Graphical User Interface of an Information System. Visual Measures-based Automated Evaluation.
SUPERVISOR(s):
Jean VANDERDONCKT, UCLouvain
JURY:
Francisco MONTERO, Universidad De Castilla-LaMancha
Ahmed Seffah LAPPEENRANTA, University of Technology
Suzanne KIEFFER, UX Strategist / VUB / UCLouvain
CHAIRMAN:
Per AGRELL, UCLouvain
ABSTRACT:
The Graphical User Interface (GUI) is probably the most frequently used interaction modality of all interactive systems. While the information contents of these GUIs are of high importance, their presentation is recognized as an essential factor that determines the GUI quality that is impacted by several determinants such as, but not limited to: visual design, aesthetics, pleasurability, arousal, and fun. Consequently, assessing the aesthetic quality of a GUI becomes a potential element to focus on in order to facilitate communication between the end user and her information system. “How to assess user interface aesthetics?” remains a question faced by several researchers and designers during the user interface development life cycle since aesthetics has been demonstrated to positively influence usability, user experience, pleasureability, and trust. Visual techniques borrowed from visual design suggest that the GUI layout could be assessed against a set of visual measures such as balance, symmetry, alignment, and simplicity, to name a few. For each visual technique, a visual measure could be computed based on a mathematical formula that captures the basic elements for each technique. Whereas different formulas exist for computing each visual measure, no consensus exists today on how to consistently evaluate these measures in a way that is aligned with human judgment, which is intrinsically subjective. In order to appropriately assess the aesthetic quality of a GUI of an information system, there is a need to align the machine objectivity based on computerized visual measures to the human subjectivity based on human perception.”
Stéphanie COSTER, UCLouvain, LouRIM/CRECIS
La « fabrique » de l’intégration professionnelle : entre contraintes institutionnelles et dynamiques organisationnelles
SUPERVISOR(s):
Evelyne LEONARD, UCLouvain
JURY:
Donatienne DESMETTE, UCLouvain
François PICHAULT, Université de Liège
Anne DIETRICH, IAE Lille I
CHAIRMAN:
Nathalie DELOBBE, UCLouvain
ABSTRACT:
La thèse examine l’influence des contraintes institutionnelles et des dynamiques organisationnelles au sein de la « fabrique » de l’intégration professionnelle. Dans le but de comprendre l’intrication entre les différents processus en jeu, cette recherche fait dialoguer les théories néo-institutionnelles et la théorie de la régulation sociale, mettant en avant les pressions institutionnelles cadrant les organisations et les choix managériaux, mais également les réactions des acteurs de l’organisation – gestionnaires et travailleurs – par rapport aux règles imposées. Au travers d’une enquête quantitative, d’une analyse documentaire et d’études de cas, cette thèse de doctorat démontre que l’intégration professionnelle est influencée par les contraintes institutionnelles et les processus organisationnels, mais dans une certaine mesure seulement. C’est finalement le jeu des acteurs en présence qui façonne l’intégration professionnelle.
Thi Thy Thao LE,UCLouvain, IACCHOS/CRIDIS
La formation professionnelle continue face au défi de la justice dans les organisations – Le cas des entreprises et universités vietnamiennes.
SUPERVISOR(s):
Matthieu de NANTEUIL, UCLouvain
JURY:
Evelyne LEONARD, UCLouvain
Hervé POURTOIS, UCLouvain
Philippe DEBROUX, University of Soka
Sandrine FERNEZ-WALCH, Université Toulouse 1 Capitole
CHAIRMAN:
Per AGRELL, UCLouvain
ABSTRACT:
Face au problème universel de l’inégalité d’accès des salariés actifs à la FPC, nous nous interrogeons sur les logiques d’action des acteurs. Les enjeux de justice en la matière se trouvent non seulement sur une proportion importante d’inégalité d’accès, mais la liberté d’action, la question du choix et les relations au travail constituent autant de questions qui touchent aux sphères de justice. Dans cette thèse, nous avons examiné les principes de justification qui ont légitimé l’engagement des acteurs dans cette pratique. Nous nous sommes intéressés également aux mécanismes institutionnels, socioéconomiques et culturels susceptibles d’influencer ces justifications. En partant d’une approche constructiviste et centrée sur les acteurs, nous avons étudié le sens de la justice en matière de FPC des salariés vietnamiens des secteurs manufacturier et universitaire. Les résultats de notre recherche montrent que l’inégalité d’accès en soi, ne constitue pas forcément un problème de justice en matière de FPC. Cette inégalité dissimule la pluralité des critères de justice des personnes concernées.
Thanh-Diane NGUYEN ,UCLouvain, LouRIM/PRISME
A cross-device pattern-based approach for designing multiple user interfaces to information systems.
SUPERVISOR(s):
Jean VANDERDONCKT, UCLouvain
JURY:
Oscar PASTOR, Universidad Politécnica de Valencia, Espagne
Ahmed SEFFAH LAPPEENRANTA, University of Technology, Finland
Suzanne KIEFFER, UX Strategist / VUB / UCLouvain
Tran Thi Ai Vi, ULB
CHAIRMAN:
Per AGRELL, UCLouvain
ABSTRACT:
Based on a comparative analysis of different user interface pattern catalogues, languages, and on the requirements elicited from it, this thesis defines a cross-device pattern as a pattern expecting to capture a general solution to a problem of designing multiple user interfaces of information systems for multiple contexts of use, particularly devices and platforms.A conceptual model of cross-device pattern is defined with its semantics expressed as a UML 2.0 class diagram and its syntax via User Interface Pattern Language Markup Language (UIPLML), a XML-compliant markup language for designing Multiple User Interfaces (MUIs) patterns for multiple contexts of use. To validate UIPLML, four pattern databases offering multi-faceted search have been created. A step-wise method for applying cross-device pattern is defined and applied on a running case study. Three frequently used design patterns, i.e., the Master-and-Details design pattern, the Population Interaction Unit and the Service Interaction Unit of the OO-Method, are then subject to a methodological instantiation. Various software prototypes (i.e., UsiMAD, PaDeRa) have been developed in order to investigate the impact of the method on designers who are responsible for designing multiple user interfaces based on cross-device patterns.
Saifeddine BEN HADJ, UCLouvain, IMMAQ/LFIN
Essays on Risk Management and Financial Stability
SUPERVISOR(s):
Isabelle PLATTEN, UCLouvain
Roland GILLET, Paris 1
JURY:
Jean-Paul LAURENT, Paris 1
Mikael PETITJEAN, UCLouvain
Frédéric LOBEZ, Lille 2
Hans DEWACHTER, KULeuven & National Bank of Belgium
CHAIRMAN:
Jean-Paul LAURENT, Paris 1
ABSTRACT:
“La thèse analyse la question de la stabilité du système financier international dans son ensemble et plus précisément comment améliorer sa résilience. Chaque chapitre se focalise sur un type d’acteur dans ce système complexe, à savoir les banques, les organismes de supervision et les régulateurs internationaux.
Le premier chapitre introduit de nouvelles techniques d’optimisation pour accélérer le calcul de mesure de risque dans les banques et les institutions financières. Il propose également une étude théorique pour valider les algorithmes d’optimisation proposés. Le second vise à quantifier l’externalité négative générée par les activités d’une banque ou d’une d’institution financière. Finalement, le dernier chapitre concerne la coopération entre régulateurs nationaux en présence de coûts de coordination en proposant une analyse qui s’appuie sur la théorie des jeux. ”
Franco STRAGIOTTI, UCLouvain, IMMAQ/LFIN
Applications for Macro Stress Testing in Times of Crisis
SUPERVISOR(s):
Leonardo IANIA, UCLouvain
JURY:
Michele LENZA, UCLouvain
Jan Willem van den END, De Netherlandsche Bank
Mikael PETITJEAN, UCLouvain
Giovanni PETRELLA, Università Cattolica del S. Cuore di Milano
CHAIRMAN:
Per AGRELL, UCLouvain
ABSTRACT:
This thesis discusses a number of potential applications for macro stress tests while focusing on the differences in the behaviours of banks between normal times and times of crisis. It comprises elements related to macro stress tests and profitability, the responses of banks to a monetary policy shock, the transmission channel of solvency risk on the funding of banks as well as the interaction between capital injections and liquidity measures that took place in the US in the aftermath of the subprime crisis. The work investigates these issues using innovative econometrics techniques, in particular dynamic factor analysis, panel data econometrics and quasi-natural experiments. The findings highlight that conditional forecasts is a valid approach to stress testing and that the dynamic factor model used to generate these forecasts captures a number of features specific to banks’ risk profiles and business models. Secondly, the results identifies how news related to solvency risk affect the funding profile of a bank, supporting the view that non-linear relationships may occur in times of crisis. At last, this research argues that the capital and liquidity measures implemented in the US in the aftermath of the subprime crisis created diverging incentives for banks, influencing their lending behaviour.
Anthony BELLOFATTO, UCLouvain, IMMAQ/LFIN
Essays on Retail Investors’ Behavior
SUPERVISOR(s):
Catherine D’HONDT, UCLouvain
JURY:
Olivier CORNEILLE, UCLouvain
Werner DE BONDT DEPAUL, University, USA
Rudy DE WINNE, UCLouvain
Patrick ROGER, University of Strasbourg
CHAIRMAN:
Ingrid PONCIN, UCLouvain
ABSTRACT:
Cette thèse s’inscrit en finance comportementale, champ de recherche interdisciplinaire qui a mis à jour l’existence de nombreux biais comportementaux chez les individus les amenant parfois à poser des choix non rationnels. Dans cette thèse, nous nous intéressons au comportement des investisseurs particuliers sur les marchés financiers. Plus précisément, nous analysons dans quelle mesure certains aspects du comportement sont affectés par la perception que l’investisseur a de ses connaissances financières (dénommée « littératie financière subjective » dans la littérature). Nos résultats montrent que l’auto-évaluation des connaissances financières permet d’expliquer les différences observées entre investisseurs en termes d’activité de trading, de diversification et de performance. Ils révèlent également un niveau de cohérence global entre la perception que l’investisseur a de ses connaissances financières et son comportement sur les marchés financiers.
Thomas LECLERCQ,UCLouvain, LouRIM/CERMA
Essays on the engagement mechanisms in the value co-creation process: The case of gamification
SUPERVISOR(s):
Ingrid PONCIN,UCLouvain
Wafa HAMMEDI, UNamur
JURY:
Pierre VOLLE, Université Paris Dauphine
Vera BLAZEVIC, Radboud University
Hubert GATIGNON, Business School INSEAD
Pietro ZIDDA, UNamur
CHAIRMAN:
Catherine D’HONDT, UCLouvain
ABSTRACT:
Au travers de quatre essais, cette thèse de doctorat explore les mécanismes de gamification – l’utilisation de mécanismes de jeu dans des contextes de non-jeu – et analyse comment ces mécanismes peuvent être mobilisés pour générer, coordonner et maintenir l’engagement du consommateur à travers le temps. Combinant des méthodes qualitatives et quantitatives, nous mettons en évidence plusieurs dynamiques d’engagement et mesurons l’impact que peut avoir la gamification tout au long de l’expérience des participants. Nous expliquons comment la gamification agit sur l’expérience des consommateurs pour générer de l’engagement. Nos résultats montrent également que, dans certains cas, la gamification peut détériorer l’expérience des participants et donc diminuer leur niveau d’engagement. Dès lors, alors que la gamification est décrite dans la littérature comme un moyen efficace pour engager les consommateurs, nous identifions les limites de ces mécanismes et fournissons aux managers des pistes de réflexion pour mieux les gérer à travers le temps.
Thomas PONGO,UCLouvain, IACCHOS/CIRTES
Le secteur agro-alimentaire : Des controverses aux mobilisations. Analyses d’engagements militants organisés au sein d’activités de production et de distribution.
SUPERVISOR(s):
Matthieu de NANTEUIL, UCLouvain
Marthe NYSSENS, UCLouvain
JURY:
Jean NIZET, UNamur & UCLouvain
Olivier GERMAIN, UQAM
CHAIRMAN:
Valérie SWAEN, UCLouvain
https://uclouvain.be/fr/chercher/cirtes/evenements/defense-de-these-0-1-2-3.html
ABSTRACT:
Cette thèse a pour objectif d’interroger les formes de militantisme agricole. Les dimensions politiques de l’activité de consommation ayant déjà reçu l’attention des chercheurs (Pleyers, 2011), cette recherche se concentre sur les engagements qui répondent aux limites des modes de production et de distribution dominants. Pour ce faire, quatre études distinctes ont été menées. La première enquête interroge la pluralité des initiatives de circuit-court de fruits et légumes en Région Wallonne. L’étude exploratoire des dimensions organisationnelles (Defourny et Nyssens, 2012) de treize organisations permet de construire une typologie originale des distributeurs alternatifs. Les deuxièmes et troisièmes enquêtes, quant à elles, questionnent l’activité décisionnelle d’un collectif de producteurs luttant pour la reconnaissance des spécificités et de la désidérabilité de modes de production alternatifs (locaux, biologiques et paysans). L’étude approfondie de cette décision autorise le chercheur à mettre au jour non seulement l’indésidérabilité des opérations critiques au sein de ce contexte (Boltanski et Thévenot, 1991) mais aussi les causes d’échecs argumentatifs. De nature plus réflexive, la quatrième recherche permet de souligner des apprentissages méthodologiques et paradigmatiques du chercheur. Ces enquêtes rendent possible une discussion concernant l’influence des modes de gouvernance sur le degré de militantisme des entreprises sociales (Borzaga et Depedri, 2015; Fazzi, 2012; Leviten-Reid et Fairbairn, 2011), la place des outils de gestion au sein de ces mêmes entreprises et, enfin, le potentiel expressif du dispositif sociocratique étudié. A l’issue de cette recherche, nous tentons de mettre au jour les enjeux et pistes d’action possibles pour la construction de dispositifs de certification alternatifs subversifs.
Félix SOMMER, UCLouvain, LouRIM
Social Network Analysis, A multi-Label Graph Approach
SUPERVISOR:
François FOUSS, UCLouvain
JURY:
Sophie BÉREAU, UCLouvain
Michael EKSTRAND, Boise State University, USA
Benoît FRÉNAY, UNamur
Marco SAERENS, UCLouvain
CHAIRMAN:
Per Joakim AGRELL, UCLouvain
ABSTRACT:
“Notre travail de thèse analyse l’implication des réseaux sociaux dans une approche data mining. Dans ce contexte, différentes applications sont proposées notamment dans les domaines des graphes (graph learning), de la segmentation (clustering) et des systèmes de recommandation.
Tout d’abord, nous avons développé un nouvel algorithme à étiquettes multiples (multi-label graph learning algorithm) tenant compte de l’entièreté des données présentes dans un graphe.
Dans une seconde partie du travail, nous proposons une nouvelle méthode de segmentation qui intègre différentes distances entre les nœuds de graphes et qui permet, entre autres, la détection de communautés dans les réseaux sociaux. Cette approche se base sur une technique de kernel k-means pondéré ainsi que sur l’utilisation du critère de modularité qui donnent plus d’autonomie à l’algorithme.
La troisième contribution à notre recherche est déclinée dans le domaine des systèmes de recommandation et appliquée dans le secteur de la grande distribution. Une approche de factorisation matricielle est utilisée pour étudier les comportements d’interactions répétés lors des achats des clients en grande distribution.
Finalement, nous avons construit un algorithme de recommandation multi-niveaux que nous avons intégré dans la plateforme LensKit (LensKit recommender toolkit). Ce modèle fonctionne à deux niveaux en tenant compte, d’une part, de la nature complémentaire des biens dans leur catégorie et, d’autre part, du caractère substituable de ceux-ci.”
Jonas TEUSCH, UCLouvain, IMMAQ/CORE
Horizontal integration under yardstick competition
SUPERVISOR(s):
Per AGRELL, UCLouvain
Axel GAUTIER, ULIEGE
JURY:
Paul BELLEFLAMME, UCLouvain
Mette BJÖRNDAL, Norwegian School of Economics, Norway
Antonio ESTACHE, ULB
CHAIRMAN:
Sergio PERELMAN, ULIEGE
ABSTRACT:
“Horizontal integration through mergers and cross-ownership is common in industries regulated by yardstick competition. However, even though yardstick competition is applied to sectors of vital economic importance, such as energy and water networks, the welfare effects of horizontal integration in these settings have largely escaped scientific scrutiny. The dissertation therefore studies the welfare implications of horizontal integration under yardstick competition and identifies potential trade-offs. Specifically, whereas mergers and cross-ownership may bring about efficiency gains, there is a possibility that horizontal integration simultaneously decreases the effectiveness of regulation, benefitting certain firms at the expense of consumers and competitors. Key conditions for such strategic effects to occur and to dominate efficiency effects are derived in a principal-agent model. The thesis also demonstrates that the issue is not merely of academic interest, but has affected the practical application of yardstick competition in the electricity distribution industries of both Belgium and Norway.
Chapter 1 analyses how horizontal mergers, joint ventures and regulatory decentralisation complicate the application of yardstick competition in Belgian electricity distribution. Chapter 2 develops a theoretical model of mergers under yardstick competition – formalising the welfare trade-offs at the heart of this dissertation. Chapter 3 leverages Data Envelopment Analysis to quantify both efficiency and strategic effects of horizontal restructuring in Norwegian electricity distribution and points to a recent merger that appears to have been driven by strategic motivations. Chapter 4 uses several econometric identification strategies to establish that between 2007 and 2015, cross-ownership increased firm performance in Norwegian electricity distribution – which suggests that the net welfare effects were positive in this application. ”
Matias SCHUSTER,UCLouvain, IMMAQ/CORE
Supply chain network design under demand uncertainty
SUPERVISOR(s):
Jean-Sébastien TANCREZ, UCLouvain
JURY:
El-Houssaine AGHEZZAF, Universiteit Gent
Mathieu VAN VYVE, UCLouvain
Stefan MINNER, Technische Universität München
Philippe CHEVALIER, UCLouvain
CHAIRMAN:
Per AGRELL, UCLouvain
ABSTRACT:
Due to the increasing volatility and complexity of global markets, the design and management of supply chains has become crucial in the success of a company. Supply chain network design decisions (i.e. number and location of facilities), have a significant impact on the performance of the company in the long term, and are directly affected by demand uncertainty. Mathematical models are an essential tool for companies to manage their supply chain efficiently, and thereby to reduce costs (e.g. facility, inventory, transportation) and to improve customer service. The purpose of this thesis is to provide different supply chain design models, integrating location and inventory decisions, that address the impact of demand uncertainty from different points of view. The ultimately goal is to provide managerial insights that may support the decision-making process in the supply chain design phase, and may allow to mitigate the impact of demand uncertainty.
Virginie BRUNEAU, UCLouvain, LouRIM/CERMA
Assessing the effectiveness of loyalty programs in building engagement and loyalty
SUPERVISOR(s):
Valérie SWAEN, UCLouvain
Pietro ZIDDA, UNamur
JURY:
Alain DECROP, UNamur
Nathalie DEMOULIN, IESEG School of Management
Yuping LIU-THOMPKINS, Old Dominion University
CHAIRMAN:
Marie-Paule KESTEMONT, UCLouvain
ABSTRACT:
“There is no doubt that loyalty programs are still popular. But are they still effective?
Marketers continue to put more and more time and money into their programs, but are they getting the kind of payoffs they expect in customer engagement and loyalty? Taking this issue as a starting point, this doctoral dissertation assesses the effectiveness of loyalty programs in building customer engagement and loyalty across three essays. Specifically, in the first essay a valid scale for customer engagement with loyalty programs is developed and its link with company engagement is assessed. The second essay identifies what reward types and communication features should be used to make members more engaged with LPs. The third essay identifies the behavioural patterns and types of online communication that tend to make customers sustainably loyal towards a company rather than short-term program loyal. ”
Joseph GILLAIN, UNamur
Essays on Requirements Optimization Problems in the Core Ontology for Requirements Engineering
SUPERVISOR(s):
Stéphane FAULKNER, UNamur
Monique SNOECK, KULeuven
JURY:
Ivan JURETA – UNamur
Yves WAUTELET – KULeuven
Nauman QURESHI – King Faisal University
CHAIRMAN:
Annick CASTIAUX, UNamur
Sous-domaine Sciences économiques et de gestion
http://agenda.unamur.be/upevent.2017-06-06.6750280472/view
ABSTRACT:
In contemporary societies, Information Systems have become ubiquitous. We use them, for instance, in accounting, marketing, human resource and also beyond the business world in health care, education or even daily life activities. It has become common to say that such systems should be of high quality, that is, they should ensure the satisfaction of the needs they were intended for.
Requirements Engineering is the field dedicated to investigate those requirements as well as the conditions for their satisfaction. However, previous and current research has mainly concentrated its effort on methods and concepts able to prove that a particular system could fulfill the requirements. In contrast, few effort had the purpose of discussing about how we should compare alternative solutions and select the optimal one.
This thesis suggests a new formulation of the Requirement Problem which sets the process of optimization in the center of concerns. It shows how different criteria can be used to design optimization models and provides tools to resolve them. Through those models, we describe how the fields of Requirement Engineering and Mathematical Optimization can be connected.
Sarah BOURAGA, UNamur
The Design of an Online Social Network and its Knowledge-Based Recommendation System
SUPERVISOR(s):
Stéphane FAULKNER, UNamur
Ivan JURETA, UNamur
JURY:
Michael PETIT, UNamur
Manuel KOLP, UCLouvain
Anna PERINI, Sofware Engineering – FBK – CIT
Haris MOURATIDIS, University of Brighton
CHAIRMAN:
Annick CASTIAUX, UNamur
http://agenda.unamur.be/upevent.2017-03-17.5763988381/view
ABSTRACT:
Over the last decade, Online Social Networks (OSNs) have been growing quickly to become some of the largest systems available. Their users are sharing more and more content, and in turn have access to vast amounts of information from and about others. This increases the risk of information overload for every user.
Information overload usually refers to the difficulty to make decisions when there is too much information. In such situations, the decision-maker is confronted to so much information that it may become unclear what the exact decision problem is, what the alternatives are, how to compare them in order to single one out.
We are interested in how to mitigate information overload for users of OSNs. Specifically, we are interested in how to design an OSN and its Recommendation Systems (RSs), which filter out and prioritize content for users, and thereby take over some of the decision-making effort from the user. The overall idea is that there can be an overload of information available to users of an OSN, and we want to design an OSN with an integrated RS which processes the information, so as to recommend to the user only those which, according to some specific criteria are the most relevant to that user.
Marco GERACI, UNamur
Essays on Complexity in the Financial System
SUPERVISOR(s):
Jean-Yves GNABO, UNamur
Davy PAINDAVEINE, ULB
JURY:
Christaine BAUMEISTER, University of Notredam, Etas Unis
Sophie BEREAU, UNamur
Catherine DEHON, ULB
David VEREDAS, Vlerick Business School
CHAIRMAN:
Annick Castiaux, UNamur
http://agenda.unamur.be/upevent.2017-09-05.2989055221
ABSTRACT:
The goal of this thesis is to study the two key aspects of complexity of the financial system: interconnectedness and nonlinear relationships.
In Chapter 1, I contribute to the literature that models the nonlinear relationship between variables at the extremes of their distribution. In particular, I study the nonlinear relationship between stock prices and short selling.
In Chapter 2, I propose a measure of interconnectedness that takes into account the time-varying nature of connections between financial institutions. Here, the parameters underlying comovement are allowed to evolve continually over time through permanent shifts at every period. The result is an extremely flexible measure of interconnectedness, which uncovers new dynamics of the US financial system and can be used to monitor financial stability for regulatory purposes.
In Chapter 3, rather than assuming interconnectedness as an exogenous process that has to be inferred, as is done in Chapter 2, I model interconnectedness as an endogenous function of market dynamics. Here, I take interconnectedness as the realized correlation of asset returns. I seek to understand how short selling can induce higher interconnectedness by increasing the negative price pressure on pairs of stocks.
—————– 
