Leuven STATistics STATe of the Art Training Initiative 2016-2017.pdf · 2016-09-06 · 1 Leuven STATistics STATe of the Art Training Initiative 2016-2017 7977_BRO_LSTAT_2016-2017.qxp_7977_BRO_LSTAT_2016-2017

1

Leuven STATistics STATe of the Art Training Initiative2016-2017

7977_BRO_LSTAT_2016-2017.qxp_7977_BRO_LSTAT_2016-2017 31/08/16 10:51 Pagina 1


1

Preface

It is my pleasure to welcoming you to the Leuven STATistics STATe of the Art Training Initiative, a scientific and

educational project of the Leuven Statistics Research Centre (LStat), offering a wide range of short courses.

Statistics in Leuven is varied with roots in various disciplines. Statisticians are active throughout the university,

in mathematics, computer science, economy, psychology, education, bio-engineering, engineering, biology, chemistry,

medicine, pharmacy, physical education, psychology, social science, linguistics, etc. Many colleagues combine an

excellent international scientific reputation with highly effective teaching skills. At the same time, statistical consulting

for internal and external clients is a substantial component of LStat’s mission as well.

It is therefore not surprising that the short course programme has been highly successful and in great demand.

Building on the success of previous years, we have further expanded the programme with new courses on highly

relevant topics, many located at the heart of our faculty’s expertise. Due to increasing demand, some courses are

offered more than once per academic year.

A selected set of courses is offered in an open educational concept, in the sense that, for example, also contingents

of students of our highly successful MSc in Statistics partake in them. This ensures stimulating interaction.

Courses take place in one of the university’s campuses, dotted around the beautiful college town of Leuven.

Should your company or institute be looking for a tailor-made training initiative, perhaps on-site, then we will be

delighted to explore options and work towards an individualized proposal.

Professor Geert Verbeke

2015-2017 chair of LStat

Leuven Statistics Research Centre

Celestijnenlaan 200 B, bus 5307

3001 HEVERLEE

+ 32 16 32 88 75

[email protected]

www.lstat.kuleuven.be


2

Presenters

Jan Beirlant gobtained his university degree in Mathematics fromGhent University and his PhD in Statistics at KU Leuvenin 1984. Next he was assistant professor in theDepartment of Statistics of the University of Washingtonin Seattle. From 1986 he holds a professorship at theDepartment of Mathematics at KU Leuven university.

His main research topic is on the extreme valuemethodology, studying the occurrence of rare events withlarge impact. He published over 100 papers in statisticaland actuarial research journals and wrote a book onStatistics of Extremes: Theory and Applications jointly withY. Goegebeur, J. Segers and J.L. Teugels. A mono graph onReinsurance: Actuarial and Statistical Aspects, written withH. Albrecher and J.L. Teugels, is appearing early 2017.

Martine Beullensgraduated in Mathematics at KU Leuven. From 1990onwards she has been working at KU Leuven and theFederal Police on projects commissioned by the Belgiangovernment on the development and statisticalexploitation of federal databanks containing judicial orpolice information. Currently she is working at KU Leuvenat the Teaching and Learning Vision and Qualitydepartment in the Data management section.

Kris Bogaertsis project manager external consultancy at the LeuvenBiostatistics and Statistical Bioinformatics Centre (L-BioStat)of the KU Leuven in Belgium. He studied mathematics(1994) at the KU Leuven, received his Master inBiostatistics (1995) at Universiteit Hasselt and his PhD inScience (statistics) at KU Leuven in 2007. He has workedas a statistical consultant in a great variety of medicalapplications. He is co-author of more than 65 peerreviewed research articles. Dr. Bogaerts is.a member ofthe Belgian Statistical Society.

An Carbonezis professor at the Leuven Statistics Research Centre (LStat)of the KU Leuven. She received her PhD in Mathematicsat the same university in 1992. She is the coordinator ofthe Master in Statistics program of the KU Leuven. She isalso involved in statistical consulting projects and givestailor made statistical courses in internal and externaldivisions and companies.

Anne-Marie De Meyeris Professor emeritus at the KU Leuven in the Facultyof Science, Department of Mathematics. She receivedher PhD in Mathematics (Applied Probability) in 1979.For more than 25 years she has been involved in theshort course program for a variety of courses in appliedstatistics and she continues to be active in the statisticalconsulting of LStat.


3

Irène Gijbels is Full Professor at the Statistics Section, Department ofMathematics, at the Katholieke Universiteit Leuven inBelgium. She received a Licentiate in Mathematics fromthe Katholieke Universiteit Leuven, and a PhD degree inScience (Mathematical Statistics). She has been SeniorResearch Assistant at the National Science Foundation,Belgium, and Visiting Professor at the University of NorthCarolina, at Chapel Hill, USA (visit supported by aFullbright-Hays grant). She previously was affiliated withthe Université catholique de Louvain. She is a Fellow ofthe American Statistical Association as well as a Fellow ofthe Institute of Mathematical Statistics. She serves/served on editorial boards of several key internationalstatistics journals. Her research topics include: Generalstatistical methodology, semi-parametric and non-parametric statistics, flexible regression modelling, the studyof dependence structures, regularization techniques,variable selection in extended regression models.

Peter Goos is a full professor at the Department of Biosystems ofthe Faculty of Bioscience Engineering of the University ofLeuven and the Department of Environment, Technologyand Technology Management of the Faculty of AppliedEconomics of the University of Antwerp. He has been aguest professor at the Econometric Institute of theErasmus School of Economics (Erasmus University ofRotterdam), the Faculty of Business and Economics ofthe University of Leuven, the Antwerp ManagementSchool and the International School of Management inSaint-Petersburg. Peter Goos has received the ShewellAward and the Lloyd S. Nelson Award of the AmericanSociety for Quality, the Ziegel Award and the Statistics inChemistry Award of the American Statistical Association,and the Young Statistician Award of the EuropeanNetwork for Business and Industrial Statistics. In 2013,Peter Goos was ranked 7th in the top 40 of economistsin the Netherlands.

Kristel Hoydonckx works within the department “Facilities for Research” ofthe KU Leuven and is among others responsible for thesurvey service.

Marlies Lacante is since 1974 part of the Psychology Department of theKU Leuven. For more than 20 years she was and remainsinvolved in the statistical education within Psychology.She is currently teaching as an associate professor at theacademic teacher training cell, the Leuven StatisticsResearch Centre (LStat) and the Master in Psychology.She is also active in education-research, focusing onsurvey research and with special attention to the researchmethodology.

Emmanuel Lesaffre is Professor of Biostatistics at the KU Leuven and U Hasseltin Belgium. He also holds a honorary professorship at theuniversity of Erasmus, Rotterdam, the Netherlands. Hestudied mathematics at the Universiteit Antwerpen, andreceived his Doctorate of Science at KU Leuven in 1986.The statistical research of Dr. Lesaffre deals withhierarchical and clustered data with a focus onlongitudinal studies, interval censoring, missing dataproblems, Bayesian methods and in general statisticalmethods in clinical trials. He has worked in a great varietyof medical applications, with focus on research in oralhealth, cardiology and nursing studies. Dr, Lesaffre is thefounding chair of the Statistical Modeling Society, and isa past-president of the International Society of ClinicalBiostatistics. In addition he started up a bi-annualinternational conference on statistical methods in oralhealth.


44

Presenters

He is one of the three founding editors of StatisticalModeling and has been Associate Editor of Biometrics,and is currently Associate Editor of Biostatistics. He waselected Fellow of the American Statistical Associationand is an honorary member of the Society of ClinicalBiostatistics. Dr. Lesaffre started up the BiostatisticalCentre, a predecessor of the Leuven Institute forBiostatistics and statistical Bioinformatics. He was alsochair of the department of Biostatistics at Erasmus MCfrom 2007 to 2014.

Geert Molenberghs is Professor of Biostatistics at the Universiteit Hasselt andKU Leuven in Belgium. He received the B.S. degree inmathematics (1988) and a PhD in biostatistics (1993)from the Universiteit Antwerpen. Dr Molenberghspublished methodological work on surrogate markers inclinical trials, categorical data, longitudinal data analysis,and on the analysis of non-response in clinical andepidemiological studies. He served as Joint Editor forApplied Statistics (2001-2004), Co-editor for Biometrics(2007–2009) and as President of the InternationalBiometric Society (2004-2005). He currently is Co-editorfor Biostatistics (2010–2013). He was elected Fellow ofthe American Statistical Association and received theGuy Medal in Bronze from the Royal Statistical Society.He has held visiting positions at the Harvard School ofPublic Health (Boston, MA). He is founding director of theCenter for Statistics at Hasselt University and currentlythe director of the Interuniversity Institute for Biostatisticsand statistical Bioinformatics, I-BioStat, a joint initiative ofthe Hasselt and Leuven universities.

José Nunez Ares is since 2014 a PhD candidate at KU Leuven under thesupervision of prof. Peter Goos. He holds a bachelor andmaster degrees on civil engineering (La Coruña, Spain,2001).

After working for several years in the private sector inSpain and in The Netherlands, in 2014 he completed amaster in Operations Research at Erasmus University.In his research he combines techniques frommathematical optimization and experimental design. Themain focus of his research lies on trend-robust responsesurface designs and the discovery of new efficientresponse surface designs. His first paper in aninternational peer reviewed journal was accepted forpublication in April 2016.

Tom Reynkens studied Mathematics at KU Leuven and got his masterdegree in 2013. He is now a PhD student in statisticsat KU Leuven under the supervision of Jan Beirlant.His research topic is applications of Extreme Value Theory(EVT) in finance and insurance. Currently, the focus lieson EVT for censored and/or truncated data. He has beenworking with R for over 7 years and wrote the R codesupplementing the book Reinsurance: Actuarial andStatistical Aspects.

Dominik Sznajder obtained a master degree in Economics and inMathematics at Warsaw University in 2007. Thereafter hecompleted a PhD programme at KU Leuven in thesection of Mathematical Statistics in 2011. Dominik wasa visiting lecturer at UGent in 2013 and an FWO post-doctoral researcher before transferring to business.Hecurrently works as an actuarial consultant with Millimansupporting insurance companies in the Beneluxregion.Dominik has over 10 years of experience in usingR, when he developed R based solutions both inBusiness and Academia.


55

Francis Tuerlinckxis Professor of Psychology at the KU Leuven in Belgium.He received the Master degree in psychology (1996) anda PhD in psychology (2000) from the KU Leuven. He wasa postdoc at the Department of Statistics of ColumbiaUniversity (New York). In general, Francis Tuerlinckx’research deals with the mathematical modeling of variousaspects of human behavior. More specifically, he workson item response theory, reaction time modeling, anddynamical systems data analysis for human emotions.

Katrijn Van Deunis assistant professor in Methodology and Statistics atTilburg University and a research fellow of the KU Leuven.She obtained a Master in psychology, a Master’s degreein statistics and a PhD in psychology. Her main area ofexpertise is scaling, clustering and component analysistechniques, which she applies in the fields of psychology,chemometrics and bioinformatics. She has variouspublications in both methodological and substantivejournals in psychometrics, chemometrics andbioinformatics. Katrijn is secretary of the Dutch/FlemishClassification society.

Geert Verbeke is Professor in Biostatistics at KU Leuven and UniversiteitHasselt. He received the B.S. degree in mathematics(1989) from the KU Leuven, the M.S. in biostatistics(1992) from Universiteit Hasselt, and earned a Ph.D. inbiostatistics (1995) from the KU Leuven. Geert Verbekehas published extensively on longitudinal data analyses.He has held visiting positions at the GerontologyResearch Center and the Johns Hopkins University(Baltimore, MD).

Geert Verbeke is Past President of the Belgian Region ofthe International Biometric Society, International ProgramChair for the International Biometric Conference inMontreal (2006). He was Board Member of the AmericanStatistical Association (2008-2010) and currently ismember of the Council of the Royal Statistical Society(2013-2016). He is past Joint Editor of the Journal of theRoyal Statistical Society, Series A (2005–2008) and editorof Biometrics (2010– 2013). He is the director of theLeuven Center for Bio-statistics and statisticalBioinformatics (L-BioStat), and vice-director of theInteruniversity Institute for Biostatistics and statisticalBioinformatics (I-BioStat), a joint initiative of the Hasseltand Leuven universities in Belgium.

Jan Wijffelsis the founder of www.bnosac.be - a consultancycompany specialised in statistical analysis and datamining. He holds a Master in Commercial Engineering, aMSc in Statistics and a Master in Artificial Intelligence andhas been using R for 10 years, developing and deployingR-based solutions for clients in the private sector. He hasdeveloped and co-developed the R packages ffbase,ETLUtils, RMOA and RMyrrix.

Tom Wilderjans is a post-doctoral researcher at the Fund for ScientificResearch (FWO-Flanders). He obtained a Master’s degree(2005) and a PhD (2009) in Mathematical Psychologyfrom the KU Leuven. Tom’s research deals with multi -variate data analysis (component analysis, clustering,and combinations thereof) and model selection.


6

Course outlineThis course gives an introduction to the use of the statis-tical software language R. R is a language for data analysisand graphics. This introduction course to R is aimed atbeginners. The course covers data handling, graphics,mathematical functions and some statistical techniques.R is for free and for more information you can visit the sitehttp://cran.r-project.org/

Target audienceEverybody who is interested in using the R programminglanguage. You will learn how to write and manage your R scripts.

PrerequisitesThere are no prerequisites.

PresenterDominik Sznajder

Course MaterialA .pdf file with the course material will be made available.

Dates27, 28, 29 and 30 September 2016, 18.00 – 21.00OR20, 21, 22 and 23 February 2017, 18.00 – 21.00

LanguageEnglish

Price• (PhD) students KU Leuven and Association

KU Leuven € 100• Staff KU Leuven and Association KU Leuven,

other (PhD) students € 160• Non profit/social sector € 250• Private sector € 600

Essential tools for R


77

Fundamentele statistischemethoden

BeschrijvingDeze basiscursus statistiek richt zich op het kiezenvan geschikte statistische methoden en het trekken vande correcte conclusies uit de verkregen resultaten.Wiskundige grondslagen van de gebruikte methodenkomen in deze cursus slechts beknopt ter sprake.De nadruk ligt op toepassing in de praktijk. Men krijgtinzicht in het adequaat gebruik van basis-statistieken:centrummaten, spreidings-maten, tabellen, box-plots,enz. Daarnaast worden betrouwbaarheidsintervallenopgesteld en krijgt men de grondslagen van toetsen vanhypothesen. Oefeningen komen in deze cursus niet aanbod. Daarvoor zijn afzonderlijke modules voorzien(toepassingen met SAS Eguide, SPSS of R) , aangepastaan de software dat de cursist gebruikt.

Inhoud van de cursus:

• Beschrijvende grootheden: grafische en numerischesamenvatting van de data

• Verdelingen: Binomiale, Poisson, Normale, T-verdeling• Steekproefverdeling van het gemiddelde• Betrouwbaarheidsintervallen• Hypothese testen omtrent een gemiddelde (één en

twee steekproeven)• Gepaarde t-test• Schatten en testen van proporties

DoelgroepIedereen die een opfrissing van fundamentele statistischetechnieken wenst.

VoorkennisEr wordt geen voorkennis ondersteld.

LesgeverProf. Marlies Lacante

CursusmateriaalCursusmateriaal wordt als .pdf file ter beschikking gesteld.

Datum 3, 4 en 5 oktober 2016, telkens van 9 tot 12 u.

TaalNederlands

Prijs• (PhD) studenten KU Leuven en Associatie

KU Leuven € 75• Personeel KU Leuven en Associatie KU Leuven,

andere (PhD) studenten: € 120• Non profit/sociale sector € 187,50• Private sector € 450


8

Course outlineThis is a web lecture with introductory and Q & A sessions.We first present linear mixed models for continuoushierarchical data. The focus lies on the modeler’sperspective and on applications. Emphasis will be onmodel formulation, parameter estimation, and hypothesistesting, as well as on the distinction between the random-effects (hierarchical) model and the implied marginalmodel. Apart from classical model building strategies,many of which have been implemented in standardstatistical software, a number of flexible extensions andadditional tools for model diagnosis will be indicated.Second, models for non-Gaussian data will be discussed,with a strong emphasis on generalized estimatingequations (GEE) and the generalized linear mixed model(GLMM). To usefully introduce this theme, a brief review ofthe classical generalized linear modeling framework will bepresented. Similarities and differences with the continuouscase will be discussed. The differences between marginalmodels, such as GEE, and random-effects models, suchas the GLMM, will be explained in detail. Third, it isoftentimes necessary to consider fully non-linear modelsfor longitudinal data. We will discuss such situations, andplace some emphasis on the non-linear mixed-effectsmodel. Fourth, non-linear mixed models will be discussed.Applications in the PK/PD world will be brought to thefront. Fifth, when analyzing hierarchical and longitudinaldata, one is often confronted with missing observations,i.e., scheduled measurements have not been made, dueto a variety of (known or unknown) reasons. It will beshown that, if no appropriate measures are taken, missingdata can cause seriously jeopardize results, andinterpretation difficulties are bound to occur. Methods toproperly analyze incomplete data, under flexibleassumptions, are presented. Key concepts of sensitivityanalysis are introduced. All developments will be illustratedwith worked examples using the SAS System. However,the course is conceived such that it will be of benefit toboth SAS users and users of other platforms.

Target audienceThe targeted audience includes methodological andapplied statisticians and researchers in industry, publichealth organizations, contract research organizations,and academia. Important: The course will also serve forthe MSc in Statistics students.

PrerequisitesThroughout the course, it will be assumed that theparticipants are familiar with basic statistical modelingconcepts, including linear models (regression and analysisof variance), as well as generalized linear models (logisticand Poisson regression) and basic knowledge of mixedand multilevel models. Moreover, pre-requisite knowledgeshould also include general estimation and testing theory(maximum likelihood, likelihood ratio). When registering forthis course, you have to mention the topics you havefollowed before and/or indicate where you becameacquainted with the requested material.

PresentersGeert Verbeke and Geert Molenberghs

Course MaterialA .pdf file with the course material will be made available.

Background reading• Verbeke, G. and Molenberghs, G. (2000) Linear Mixed

Models for Longitudinal Data. New York: Springer.• Molenberghs, G. and Kenward, M.G. (2007) Missing

Data in Clinical Studies. Chichester: John Wiley & Sons.• Molenberghs, G. and Verbeke, G. (2005) Models for

Repeated Discrete Data. New York: Springer.

Models for Longitudinal andIncomplete Data


9

Dates11 October 2016: 09.00 – 11.00 (Introductory)18 October 2016: 09.00 – 11.00 (Q & A)9 November 2016: 09.00 – 11.00 (Q & A non-Gaussian LDA)1 December 2016: 09.00 – 11.00 (Q & A missing data)

Web lectures have to be watched prior to the Q & A session(s)!

LanguageEnglish

Price• (PhD) students KU Leuven and Association KU Leuven € 250• Staff KU Leuven and Association KU Leuven, other (PhD) students € 400• Non profit/social sector € 625• Private sector € 1500


10

BeschrijvingDit is een inleidende cursus tot het gebruik van SPSS.Aan de hand van cases wordt geïllustreerd hoe men metSPSS tot exploratie van gegevens komt. Hierbij wordt denodige aandacht besteed aan het interpreteren van deverkregen output. Hypothesetesten voor onafhankelijkeen gepaarde groepen worden uitgevoerd en besproken.Er is tijd om zelf te werken met deze software.

DoelgroepIedereen die gegevens wenst te exploreren met SPSS.

VoorkennisDe technieken die aangeleerd werden bij FundamenteleStatistische Methoden.

LesgeverProf. Marlies Lacante


Datum18 oktober 2016, 9 u -12 u en 13 u -16 u.

TaalNederlands

Fundamentele statistische methoden,toegepast met SAS Eguide, SPSS en R

CURSUS 1: FUNDAMENTELE STATISTISCHE METHODEN,TOEGEPAST MET SPSS

BeschrijvingDit is een inleidende cursus tot het gebruik van SASEnterprise Guide (niet de SAS Language). Aan de handvan cases wordt geïllustreerd hoe men met de SASEguide tot exploratie van gegevens komt. Hierbij wordt denodige aandacht besteed aan het interpreteren van deverkregen output. Hypothese testen voor onafhankelijke engepaarde groepen worden uitgevoerd en besproken. Er istijd om zelf te werken met deze software.

DoelgroepIedereen die gegevens wenst te exploreren met SASEguide.

VoorkennisDe technieken die aangeleerd werden bij FundamenteleStatistische Methoden.

LesgeverMartine Beullens


Datum13 oktober 2016, 9 u -12 u en 13 u-16 u

TaalNederlands

CURSUS 2: FUNDAMENTELE STATISTISCHE METHODEN,TOEGEPAST MET SAS EGUIDE


11

Course outlineBy using cases, one explores data by using R. Attentionis paid to the interpretation of the output. Topics asexploring data, construction of confidence intervals andhypothesis testing is covered. This is a hands-on session.

Target audienceEverybody who wants to explore data by using R.

Prerequisites Fundamental Statistical Methods (distributions, confidenceintervals, hypothesis testing) and Introduction to R.

Presenter TBA

Course MaterialsA .pdf file with the course material will be made available.

Date13 October 2016, 09.00 – 12.00 and 13.00 – 16.00

Language English

COURSE 3: FUNDAMENTAL STATISTICAL RESEARCH METHODS,APPLICATIONS WITH R

• (PhD) students KU Leuven and Association KU Leuven € 50• Staff KU Leuven and Association KU Leuven, other (PhD) students € 80• Non profit/social sector € 125• Private sector € 300

PRICE PER COURSE


12

Course outlineThis course is a hands-on course covering the use ofstatistical machine learning methods available in R.

The following basic learning methods will be covered andused on common datasets.

• classification trees (rpart)• feed-forward neural networks and multinomial regression• random forests• boosting for classification and regression• bagging for classification and regression• penalized regression modelling (lasso/ridge regularized

generalized linear models)• model based recursive partitioning (trees with statistical

models at the nodes)• training and evaluation will be done through the use of

the caret and ROCR packages

The course will cover the techniques from a high-levelviewpoint, useful for day-to-day R users.

Target audienceThe course is for R users in industry/academics whoare interested in building predictive models in R whichhave some experience with regressions but have lessknowledge of machine learning and techniques of artificialintelligence.

Also persons interested in the statistical learning techniquesitself will find this course useful. Or people with a datascience background with less knowledge of R and whichare interested in machine learning in general.

PrerequisitesInitial experience in R ranging from a few weeks to severalyears (at least understanding of the course 'Essential Toolsfor R' is needed). Some practical experience in regressionmodelling.

PresenterJan Wijffels

Course materialsA .pdf file with the course material will be made available.

Date17 and 18 October 2016, 09.00 – 12.00 and13.00 – 16.00

LanguageEnglish




Statistical Machine Learning with R


13

Course outlineLinear statistical models are widely used today in manyapplications. Successfully applying these techniquesrequire a good understanding of the underlying theory andthe practical problems that you may encounter in real-lifesituations.

DAY 1: REGRESSION ANALYSIS

• Correlation • Simple linear regression

Ordinary least squares: estimating parameters, confidenceintervals and tests, diagnostics, prediction.

• Multiple regression Ordinary least squares: estimating parameters, confidenceintervals and tests, diagnostics, prediction.Variable selection techniques.

DAY 2: ANALYSIS OF VARIANCE

• One-way AnovaComparing meansThe Anova model: estimating parameters, hypothesistests, Anova tabel, F testMultiple comparisons, Contrasts

• Two-way AnovaThe two-way Anova ModelMain effects, interaction effectsMultiple comparisons

Target audienceThis course is important for persons involved withmodeling data.

PrerequisitesParticipants are familiar with basic statistical modelingconcepts (see topics described in Fundamental StatisticalMethods).

PresentersTBA


Dates26 and 27 October 2016 : 09.00 – 12.00 and13.00 – 16.00

LanguageEnglish




Regression and Analysis of variance


14

Course outlineThis course is a hands-on course covering the use of textmining tools for the purpose of data analysis. It coversbasic text handling, natural language engineering andstatistical modelling on top of textual data.

The following items are covered:

• Cleaning of text data, regular expressions• String distances• Graphical displays of text data• Natural language processing: stemming, parts-of-

speech tagging, tokenization, lemmatisation• Sentiment analysis• Statistical topic detection modelling (latent diriclet

allocation)• Automatic classification using predictive modelling

based on text data• Visualisation of correlations & topics• Word embeddings• Document similarities & Text alignment

Target audienceThe course is for R users in industry/academics who areinterested in practical natural language processing andstatistical learning on text data. People with a data sciencebackground with less knowledge of R and which areinterested in machine learning & text mining in general willfind this course, alongside the course on StatisticalMachine Learning with R very useful.

PrerequisitesInitial experience in R ranging from a few weeks to severalyears (at least understanding of the course 'Essential Toolsfor R' is needed, preferably also ‘Advanced R programmingtopics’). Some practical experience in regressionmodelling. In order to follow the parts on topic modellingand automatic classification of text data, you should haveknowledge of statistical modelling.



Date24 and 25 October 2016, from 09.00 to 12.00 and from13.00 to 16.00 (max. 15 participants)OR23 and 24 March 2017, from 09.00 to 12.00 and from13.00 to 16.00 (max. 15 participants)

LanguageEnglish




Text mining with R


15

Regression and Analysis of Variance,applications with SPSS, SAS and R

BeschrijvingDe technieken die aangeleerd werden bij Regressie-en variantieanalyse, worden hier toegepast met SPSS.Aan de hand van cases wordt geïllustreerd hoe men metSPSS tot het modelleren van gegevens komt. Hierbij wordtde nodige aandacht besteed aan het interpreteren van deverkregen output. Er is voldoende tijd om zelf te werkenmet deze software.

DoelgroepIedereen die gegevens wenst te modelleren via SPSS.

VoorkennisWe veronderstellen een basiskennis van SPSS. Cursistendienen eveneens vertrouwd te zijn met de methodiekaangebracht in Regressie- en variantieanalyse.

LesgeverAn Carbonez


Datum 7 november 2016 van 9 u -12 u en van 13 u -16 u.

TaalNederlands

CURSUS 1: REGRESSIE- EN VARIANTIEANALYSE, TOEGEPAST MET SPSS

CURSUS 2: REGRESSIE- EN VARIANTIEANALYSE,TOEGEPAST MET SAS EGUIDE

BeschrijvingDe technieken die aangeleerd werden bij Regressie- envariantieanalyse, worden hier toegepast met SAS Eguide.Aan de hand van cases wordt geïllustreerd hoe men metde SAS Eguide tot het modelleren van gegevens komt.Hierbij wordt de nodige aandacht besteed aan het inter-preteren van de verkregen output. Er is voldoende tijd omzelf te werken met deze software.

DoelgroepIedereen die gegevens wenst te modelleren via SAS Eguide.

VoorkennisWe veronderstellen een basiskennis van SAS Eguide. Cursisten dienen eveneens vertrouwd te zijn met de methodiek aangebracht in Regressie en variantie analyse.

LesgeverMartine Beullens


Datum 14 november 2016 van 9 u -12 u en van 13 u -16 u.

TaalNederlands


16

Course outlineThe linear models, provided by the course ‘Regressionand Analysis of Variance’, are applied on examples. In thiscourse, the R package is used. By means of cases, weillustrate how to model your data in R and how to interpretthe corresponding output. There is a hands-on session totrain you with the functionality of R.

Target audienceEverybody who wants to model data with R.

Prerequisites Everybody should be familiar with the techniques coveredin ‘Regression and Analysis of Variance’ and have a basicknowledge of working with R.

Presenter TBA


Date15 and 17 November 2016, 09.00 – 12.00 and13.00 – 16.00

Language English


PRICE FOR COURSE 1 AND COURSE 2 (PRICE PER COURSE)


PRICE FOR COURSE 3

COURSE 3: REGRESSION AND ANALYSIS OF VARIANCE:APPLICATIONS WITH R


17

Optimization & Numerical Methodsin Statistics

Course outlineThis is a web lecture with Q & A sessions

Numerical problems are frequently encountered bystatisticians. Prominently, the estimation of the parametersof a statistical model requires the solution of anoptimization problem. In a few simple cases, closed-formsolutions exist but for many probability models the optimalparameter estimates have to be determined by means ofan iterative algorithm. The goal of this course is threefold.First, we want to offer the readers an overview of somefrequently used optimization algorithms in (applied)statistics. Second, we want to provide a framework forunderstanding the connections among several optimizationalgorithms as well as between optimization and aspectsof statistical inference. Third, although very common,optimization is not the only numerical problem andtherefore some important related topics such as numericaldifferentiation and integration will be covered.

Target audienceThe intended target audience includes PhD students andresearchers in a variety of fields, including biostatistics,psychometrics, educational measurement, public health,sociology. We aim at readers who apply and possiblydevelop statistical models and who wish to learn moreabout the basic concepts of numerical techniques, withan emphasis on optimization problems, and their use instatistics.

Prerequisites Participants should have a basic knowledge of theprinciples of statistical inference. This includes somefamiliarity with the concept of a likelihood function andlikelihood-based inference for linear, binomial, multinomial,and logistic regression models. Readers should also havea basic understanding of matrix algebra. A workingknowledge of the basic elements of univariate calculus isalso a prerequisite, including (the concepts of continuityof a function, derivative and integration).

PresentersGeert Molenberghs, Francis Tuerlinckx, Katrijn Van Deun,Tom Wilderjans


Background reading• Everitt, B.S. (1987). Introduction to Optimization

Methods and Their Application in Statistics. London:Chapman & Hall.

• Lange, K. (1999). Numerical Analysis for Statisticians.New York: Springer.

• Lange, K. (2004). Optimization. New York: Springer.

DatesThe course is organized in a dual format:

• First, web lectures need to be followed.• Second, a two-day contact session is organized on

November 17 (10.30 – 12.15 and 13.15 – 17.15) andNovember 18 (08.45 – 12.45 and 13.45 – 15.45).

The goal is:

• Questions and answers• Exercises to further deepen knowledge

Note, these are not formal lectures; it is important that theweb lectures are watched prior to the contact session.

LanguageEnglish





18

Uitbreiding bij Regressie- envariantieanalyse

BeschrijvingDe resultaten van een lineaire regressieanalyse zijn sterkbeïnvloedbaar door speciale datapunten. Het detecterenvan uitschieters en invloedrijke waarnemingen wordt indeze cursus bestudeerd. Daarnaast wordt geïllustreerdhoe men via robuuste regressie dit probleem kanopvangen. De praktijk leert ook dat de resultaten van eenlineaire regressie ook sterk beïnvloed worden doorassociaties tussen verklarende variabelen. Dit probleemvan multi collineariteit wordt besproken en geïllustreerd.Verder is er een uitbreiding van variantieanalyse naarspecifieke deelhypothesen en covariantieanalyse. Er wordttelkens geïllustreerd hoe de analyses met SAS Eguide enSPSS kunnen uitgevoerd worden.


DAG 1: UITBREIDING VAN REGRESSIE

• Speciale datapunten: detectie van uitschieters en invloed -rijke waarnemingen

• Inleiding tot robuuste regressie• Multicollineariteit

DAG 2: (HALVE DAG)

• Covariantie analyse

Doelgroep Deze cursus is bedoeld voor personen die regelmatig lineaire regressieanalyse wensen te gebruiken.

VoorkennisCursisten dienen vertrouwd te zijn met de methodiek aangebracht in ‘Regressie- en variantieanalyse’.

LesgeverMarlies Lacante, An Carbonez


Datum 22 november 2016, van 9 u -12 u en 13 u -16 u en 23 november 2016, van 9 u -12 u.

TaalNederlands





Introduction to JMP: Data management and basic statistics

Course outlineBasic two-days hands-on course to give a broadintroduction on JMP, with special focus on graph buildingand interactive features of JMP.

DAY 1: DATA MANAGEMENT AND DESCRIPTIVE STATISTICSWITH JMP

• Basic static and dynamic graphs, • Maps, • Descriptive statistics, • Managing data (filtering, coding, merging, …), • Hypothesis testing, • Power calculations

DAY 2: BASIC REGRESSION AND ANOVA WITH JMP

• Simple and multiple linear regression, • Model building (incl. stepwise regression), • Visualization of regression modelling results (e.g., quadratic

effects, interaction effects, …), • One-way and multi-way analysis of variance (ANOVA)• Logistic regression

Target audienceEveryone who wants to use JMP for data visualization, descriptive statistics and model fitting.

PrerequisitesParticipants are familiar with basic statistical modeling,regression and analysis of variance concepts (see topicsdescribed in Fundamental Statistical Methods andRegression and Analysis of variance). If you have basicknowledge in statistics and you want to know how to useJMP, then this is a course for you. If you miss this basicknowledge, then you can first follow ‘FundamenteleStatistische methoden’ or ‘Regression and Analysis ofVariance’.

PresenterJosé Nunez Ares

Course materialsA .pdf file with the course material will be made available.Textbook (not compulsory): Statistics with JMP: Graphs,Descriptive Statistics and Probability (Peter Goos, DavidMeintrup)

Date24 and 25 November 2016, 09.00 – 12.00 and13.00 – 16.00

LanguageEnglish




19


20

Logistic Regression Modelswith SAS, SPSS and R

DescriptionThe focus is on the statistical model with a categoricaloutcome or response variable.

A categorical response variable can be a binary variable,an ordinal variable or a nominal variable and each typerequires a different model to describe its relationship withthe predictor variables.

We will define, interpret and illustrate the models for eachtype of outcome and place the models in the frameworkof the Generalized Linear Model

SAS ( through SAS Enterprise Guide), SPSS or R will beused in the applications.

Outline

• Introduction• Binary Logistic Regression• Multinomial Logistic Regression for nominal outcome

variables• Proportional Odds Model - Ordinal Logistic Regression• Logistic regression in the framework of the Generalized

Linear Model

Target audienceData analysts in all disciplines

Prerequisites‘Fundamental Statistical Methods’ and ‘Introduction to the analysis of contingency tables’.

Knowledge of the standard regression model

PresenterAnne-Marie De Meyer


Dates 30 November and 2 December 2016, from 09.00 to 12.00

Language English




SPSS


BeschrijvingDeze cursus behandelt een aantal statistische technieken- analoog aan parametrische statistiek (bv. t-test, variantie -analyse) - waarbij de klassieke onderstellingen uit deparametrische statistiek niet hoeven gemaakt te worden(distributievrije technieken), technieken gebaseerd op‘ordeningen’ of ‘rankings’, alsook technieken specifiekgeschikt voor nominale gegevens.


• Chi- kwadraat goodness of fit testen• Testen mbt verschil tussen twee onafhankelijke

steekproeven• Testen mbt verschil tussen twee afhankelijke

steekproeven • Testen mbt verschil tussen meerdere onafhankelijke

steekproeven• Testen mbt verschil tussen meerdere afhankelijke

steekproeven• Kengetallen mbt de samenhang tussen variabelen

DoelgroepGebruikers van basis statistische technieken (t-test -variantie-analyse)

VoorkennisCursisten dienen vertrouwd te zijn met de methodiek aangebracht in ‘Fundamentele Statistische technieken’ en variantie analyse.

LesgeverMarlies Lacante


Datum 6 december 2016, van 9 u tot 12 u.

TaalNederlands




Niet-parametrische statistiek

21


22

Course outlineThis course is useful for data scientists and data analystswhich work frequently with data with a spatial component(data with latitude/longitude information). It gives anintroduction to the numerous spatial facilities of R andsome standard spatial statistical models.

The following items are covered during the course.

• The sp package to handle spatial data (spatial points,lines, polygons, spatial data frames)

• Importing spatial data and setting the spatial projection• Plotting spatial data on static and interactive maps • Adding graphical components to spatial maps• Manipulation of geospatial data, geocoding, distances, …• Density estimation, kriging and spatial point pattern

analysis• Spatial regression

Target audienceThe course is for R users in industry/academics who workfrequently with spatial data and R and who want to eitherplot results on spatial graphs or used spatial modellingtechniques on their data.

PrerequisitesInitial experience in R ranging from a few weeks to severalyears (at least understanding of the course 'Essential Toolsfor R' is needed, preferably also ‘Advanced R programmingtopics’). Some practical experience in regressionmodelling, handling of correlations and using R graphics.



Date 8 and 9 December 2016, 09.00 – 12.00 and13.00 – 16.00

LanguageEnglish




Applied Spatial modelling with R


23

Course outlineThis course discusses the design of factorial experiments.Initially, the focus is on completely randomized experimentaldesigns. Next, the focus shifts to experimental designsinvolving a restricted randomization. First, the concept ofblocking is discussed. Next, split-plot and strip-plotdesigns are studied.

The emphasis in the course is on the optimal design ofexperiments. In optimal design of experiments, theexperimental design is tailored to the problem at hand(unlike classical experimental design, where standarddesigns from catalogs are chosen).

The course builds on concepts from regression andanalysis of variance, such as fixed and random effects,power calculations, variance inflation factors, multi collinearity,confidence intervals, prediction and lack-of-fit tests.

Every topic in the course is introduced and illustrated bymeans of a case study from industry. The case studies arerealistic in the sense that they involve quantitative andqualitative experimental factors, experimenters have todeal with limited budgets and difficulties to randomize,and forbidden combinations of factor levels. In each ofthe case studies, the goal is to enhance to performanceof a process or a product.

The statistical software package used is JMP.

Target audienceThe target audience for the course is master students orPhD students in statistics, engineers and engineeringstudents planning to perform experiments. The courseis ideal for Six Sigma black belts, quality managers andpeople working in environments where collecting data isexpensive.

PrerequisitesPrerequisites for the course are knowledge of basicstatistics, regression analysis (least squares, multicollinearity),and matrix algebra (matrix products, inverse matrices,determinants).

PresenterPeter Goos

Course MaterialsTextbook:Optimal Design ofExperiments: A Case Study Approach(Peter Goos & Bradley Jones)

Dates 14 and 21 February 2017, 09.00 – 11.307, 14, 21 and 28 March 2017, 09.00 – 11.3018 and 25 April 2017, 09.00 – 11.302 and 9 May 2017, 09.00 -11.30

LanguageEnglish




Experimental Design


24

Nonparametric smoothingtechniques and applications

Course outlineNonparametric smoothing techniques are an importantclass of tools for identifying the true signal hidden in noisydata. These tools are widely used in statistical analysis ina variety of application areas. This course will provide thestudents with a thorough overview of the most importantsmoothing techniques (such as kernel smoothing,local polynomial fitting, spline smoothing, waveletdecomposition, regularization techniques, …). The coursewill address theoretical and computational aspects.We will discuss how to use these techniques in differentsettings (e.g. in a univariate or multivariate regressionsetting, in case of incomplete data, …). The courseincludes illustrations with data examples and the use ofthe R software.

Target audiencePhD students or practitioners/researchers with a goodbackground knowledge of statistics and statisticalinference.

PrerequisitesParticipants should have a good background in statistics,in particular in statistical inference.

PresenterIrène Gijbels


Background reading• Fan, J. and Gijbels, I. (1996). Local Polynomial Mode-

ling and Its Applications. Chapman and Hall, New York. • Hastie, T., Tibshirani, R. and Friedman, J. (2001).

The Elements of Statistical Learning. Springer, NewYork.

• Simonoff, J.S. (1996). Smoothing Methods in Statistics.Springer, New York.

Dates 14, 21 and 28 February 2017, 09.30 – 12.30 and14.00 – 16.007 and 14 March 2017, 09.30 – 12.30 and 14.00 – 16.00

Language English




NEW


25

Course outlineThe goal of the course isThe goal of the course is to teachstudents how to perform multivariate sensor calibration.Students will become familiar with the use of statisticalconcepts in chemometric applications. Most attention willbe given to the ideas underlying the different methods andthe application of these methods to realistic examples.Theoretical considerations and equations will be limited towhat is needed to have sufficient insight to properly usethe methods. Most examples will be related tospectroscopy and analytical chemistry, but the scope isbroader. By using a combination of lectures, computersessions and take home assignments the students willreally learn how to apply the chemometric methods.

The following aspects of chemometrics will be handled inthis course:

• Classical modelling concepts for quantitative calibration:Classical Least Squares (CLS), Inverse Least Squares(ILS), Multivariate Linear Regression (MLR), PrincipleComponent Regression (PCR) and Partial Least Squares(PLS).

• Necessary steps for the creation and successfuldeployment of calibrations; selection of calibrationstandards and assessment of the reliability of themodels: (Test set validation vs. Cross-validation, modelstatistics). Special attention will be given to the methodsfor the selection of the number of principle componentsor latent variables in the projection methods.

• Methods for data pre-processing with special attentionfor the phenomena of light scattering and instrument driftand the methods to deal with these phenomena:derivatives, standard normal variate (SNV), multiplicativesignal correction (MSC) and extended multiplicativesignal correction (EMSC).

• Variable selection in a chemometric context and somecommonly used methods for this.

• Qualitative analysis in a chemometric context:discrimination and classification

• New trends in chemometrics such as functional dataanalysis and augmented classical least squares (ACLS).

Target audienceThe intended target audience includes PhD students andresearchers in a variety of fields, including statistics,chemistry, biosciences and engineering. We aim atreaders who wish to learn more about multivariatecalibration of sensor systems and the use of statisticalconcepts in chemometric applications.

Prerequisites Knowledge of basic concepts of statistics and linearalgebra is required. Some notions of analytical chemistry,sensor technology and multivariate statistics are a plus.

PresenterWouter Saeys

Course MaterialsSlides from the lecturesPapers discussed in the lecturesSoftware manualDatasets for the take homeassignmentsAdditional material (suggested)

• A user-friendly guide toMultivariate Calibration andClassifi cation by Naes,Isaksson, Fearn and Davies,NIR Publications 2004

• Multivariate Calibration byMartens and Naes, 1989

Dates15 and 22 February 2017, 09.00 – 12.00 8, 15 and 22 March 2017, 09.00 – 12.00

LanguageEnglish



other (PhD) students € 200• Non profit/social sector € 312,50• Private sector € 750

Chemometrics


26

Course outlineR is the lingua franca of statistical research and dataanalysis. But in order to get you up and running with R,and to get over the steep learning curve, you need toknow how to use it efficiently.

This course is a hands-on course covering the basic toolkityou need to have in order to use R efficiently for dataanalysis tasks.

It is an intermediate course aimed at users who have theknowledge from the course ‘Essential tools for R’ and whowant to go further to improve and speed up their dataanalysis tasks.

The following topics will be covered in detail:

• The apply family of functions and basic parallelprogramming for these, vectorisation, regularexpressions, string manipulation functions andcommonly used functions from the base package.Useful other packages for data manipulation.

• Making a basic reproducible report using Sweave andknitr including tables, graphs and literate programming

• If you want to build your own R package to distributeyour work, you need to understand S3 and S4methods, you need the basics of how generics workas well as R environments, what are namespaces andhow are they useful. This will be covered to help youstart up and build an R package.

• Basic tips on how to organise and develop R code andtest it.

Target audiencePeople who have had their initial use of R and want to goone step further.

This covers people using R for a few months already toseveral years. And more specifically users who want toextend their data manipulation techniques to speed uptheir day-to-day data analysis tasks.

Researchers from the university interested in makingreproducible research reports or users who want to use Ras a report generating tool.

R users interested in getting the fundamentals you needto know before you can create your own R package.

Business users who want to learn how to get themaximum out of R by speeding up their code, learnvectorisation, execute the basics of parallel programmingand want to learn how to build methods and code whichis reproducible in production environments.

PrerequisitesInitial experience in R ranging from a few weeks to severalyears.



Date21 and 22 February 2017, 09.00 – 12.00 and13.00 – 16.00

LanguageEnglish




Advanced R programming


27

Course outlineMultivariate data consist of observations on two or morevariables for each individual or unit.

The variables will be generally correlated, and a variety oftechniques are available to analyse these data.

The objective of cluster analysis is to form groups ofobservations such that each group is as homogeneous aspossible with respect to certain characteristics. The groupsare as different as possible.

Principal component analysis is one of the popular toolsto summarize quantitative multivariate data.

During this course, PCA and exploratory factor analysis,will be introduced and the relation between themexamined.

The emphasis of the course will be on the interpretationof the example data and on the results through the Biplot.Mathematical details are kept to a minimum.

For the exercises, participants can choose to use thestatistics package SAS( through Enterprise Guide), SPSSor R

Course content:

• Hierarchical cluster analysis• Nonhierarchical cluster analysis• Linear combination of variables• Eigenvalues and eigenvectors • PCA scores and Factor scores• What is a loading or the factor pattern?• Screenplot• How many components or factors to retain? • Communalities• The biplot• The Varimax rotation

Taget audienceData analysts and scientists involved in analysing multi -variate data.

PrerequisitesA practical knowledge of basic statistics will be assumed,such as standard deviations and correlations.

PresentersAnne-Marie De Meyer and An Carbonez

Course materialsA .pdf file with course material will be made available

Dates 7 March (9.00 – 13.00), 9 and 10 March 2017,09.00 – 12.00

Language English




Cluster analysis, principal componentanalysis and exploratory factor analysiswith SAS Eguide, SPSS and R


28

NEW

Course outlineThis course is a hands-on course covering the use ofimage analysis. It covers basic image manipulation, featureengineering techniques and finding patterns in images

The following items are covered:

• image manipulation & adjustments• finding blobs, corners, gradients, edges & lines• feature & object detection• applying filters• deep learning for image analysis• image segmentation

Target audienceThe course is for R/Python users in industry/academicswho are interested in image analysis and have a computerscience or data science background.

PrerequisitesYou need some experience in R & Python for following thecourse. For the part on the modelling, knowledge ofstatistical learning techniques is preferred.



Date9 and 10 March 2017, 09.00 – 12.00 and 13.00 – 16.00

LanguageEnglish




Computer Vision with R and Python


29

Course outlineThis basic course in statistics emphasizes on selectingthe appropriate statistical method and drawing the rightconclusions from the obtained results. Mathematicaldetails will be kept to a minimum. The emphasis will beon understanding the concepts and on practicalapplications. The adequate use of basis statisticalsummaries (measures of central tendency, measures ofdispersion, box-plots,...) will be illustrated. The foundationsof confidence intervals and of testing hypotheses will bedealt with.

Course content:

• Descriptive statistics: graphical and numerical summariesof the data

• Distributions: Binomial, Poisson, Exponential, Normaland t-distribution

• Distribution of the sample mean• Confidence intervals• Hypothesis tests for a population mean (one and two

samples)• Paired t-test• Estimating and testing for proportions

Target audienceAnyone who wishes to understand basic statisticaltechniques more thoroughly.

PrerequisitesThere are no prerequisites

PresenterMarlies Lacante

Course materialA .pdf file with the course material will be made available.

Dates 14, 15 and 16 March 2017, 09.00 – 12.00

LanguageEnglish




Fundamental Statistical Methods


30

Concepts of Multilevel,Longitudinal and Mixed models

Course outlineStarting from ANOVA models with random factor levels,the concepts of mixed models are introduced and thebasics about inference in random-effects models will beexplained. Afterwards, the mixed ANOVA model isextended to general linear mixed models for continuousdata. Finally, extensions to models for binary or count datawill be briefly discussed. Omitting all theoretical details,sufficient background will be given such that practisingstatisticians can apply mixed models in a variety of contexts,know how to use up-to-date software, and are able tocorrectly interpret generated outputs. Many applications,taken from various disciplines, will be discussed.

Target audienceThe targeted audience includes methodological andapplied statisticians and researchers in industry, publichealth organizations, contract research organizations andacademia. Important: this course will also serve the MScin Statistics students.

Prerequisites The student knows the basics of statistical inference, andstatistical modeling (regression, Anova and generalizedlinear models).

PresenterGeert Verbeke

Course MaterialsA .pdf file with the copies of the transparencies used inthe course will be made available.

Dates 14, 21 and 28 March 2017, 13.00 – 16.0018 and 25 April 2017, 13.00 – 16.00

Language English





31

Introduction to the analysis ofcontingency tables

Course outlineIn this course, chi-square tests and association measuresare used to identify if there is significant association incontingency tables and to determine how strong thisassociation is. By use of examples, it is illustrated thatexact tests are necessary in certain situations. There isenough time to practice with SAS Eguide, SPSS and R.

Content:

• Construction of a contingency table• Tests for independence: chi square tests• Association measures• Analysis of a 2x2 table: relative risk, odds ratio• Exact test

Target audience Everyone who wants to analyse contingency tables.

PrerequisitesParticipants are familiar with basic statistical concepts(which are e.g. introduced in the course ‘Fundamental Statistical Methods’).

PresenterAn Carbonez


Dates 28 March 2017, 09.00 – 12.00 and 13.00 – 16.0030 March 2017, 09.00 – 12.00

Language English





32

NEW

Course outlineStatistical analysis of extreme values is receiving a growingattention in many different fields of applications.

For instance:• In hydrology and meteorology the impact of extreme

rainfall and storms is receiving quite some attention inthe context of climate change,

• In finance and insurance accurate modelling of largenegative returns and large claims is requested by theregulators,

• In geophysics the estimation of the largest earthquakemagnitude in a given region is of great importance,

while applications can be found in sports, lifetime analysisand risk assessment in general.

In this course we present the basic principles behind thestatistical modelling and estimation of extreme valueparameters, return levels and return periods. We overviewmethods in a univariate, multivariate and regressioncontext.

The course is illustrated with applications from differentfields and hands-on implementation in R is provided.

Target audiencePotential users of extreme value methods from industryand academics and from different fields such as finance,insurance, hydrology, climatology, geophysics, ...

PrerequisitesIt will be assumed that participants are familiar with thebasic concepts of statistics and probability

PresenterJan Beirlant and Tom Reynkens


Date25 and 26 April 2017, 09.00 – 12.00 and 13.00 – 16.0027 April 2017, 09.00 – 12.00

LanguageEnglish




Extremes


33

Course outlineThis is a web lecture with face-to-face and Q & Asessions.

Different methods for selecting a (survey) sample from anexisting population will be considered. Problems arising inthe sampling designs will be discussed. The focus will beon the concepts rather than on the formulas. Nevertheless,attention will be paid at the estimation of the populationparameters of interests.

Target audienceEveryone with an interest in sampling theory, from an appliedand/or methodological point of view.

PrerequisitesParticipants should have an intimate knowledge of basicconcepts of descriptive and inductive statistics.

PresenterGeert Molenberghs

Course materialA .pdf file with the course material will be made available

Dates 25 April 2017, 16.30 – 20.00 (face-to-face lecture)

The other lectures will be accessible via the web.

9 May 2017, 18.00 – 20.00 (Face- to-face Q & Asession 1)18 May 2017, 18.00 – 20.00 (Face-to-face Q & Asession 2)

Web lectures need to be watched before the Q & Asessions!

Language English




Survey Sampling Methods


34

BeschrijvingVia survey onderzoek wil men informatie verzamelen omtrentmensen, ideeën, opinies, houdingen, plannen, gezondheid,sociale - educatieve - of familiale achtergrond. Zulk soortonderzoek gebeurt vaak bij sociologische vraagstellingen, inpsychologie, bij marktonderzoek, enz… Informatie overdergelijke onderwerpen kan men moeilijk op ‘experimentelewijze’ verzamelen. Daarom moet men de personen inkwestie bevragen. Dit kan via een interview, een vragenlijst,een telefonische enquête, enz…. Dit soort bevragingen kenteen eigen methodologie en eigen onderzoeksregels diemoeten gerespecteerd worden. In deze cursus wordtachtereenvolgens ingegaan op de verschillende stappen indit onderzoeksproces.

Tevens zal een half dagdeel besteed worden aan deenquêteservice aan de KU Leuven, die gebaseerd is op deopen-source software “Limesurvey”. Deze software laatgebruikers toe om snel zeer krachtige online enquêtes teontwikkelen.


• Analyse van de onderzoeksvraag: wat wil men te wetenkomen?

• Verzamelen van de gevraagde informatie • Welke regels moet men in acht nemen bij het formuleren

van de vragen? (invloed van de vraagstelling op het antwoord, betrouwbaarheid en validiteit)

• Methoden van steekproeftrekkingen • Verwerken van de gegevens • Rapportering • Hoe werkt de enquêteservice van de KU Leuven• Algemene instellingen voor de enquête• Beschikbare vraagtypes en hun mogelijkheden • Werken met tokens• Uitnodigen van de respondenten en opvolgen van de

responses• Exporteren van de resultaten naar statistische pakketten

DoelgroepGebruikers van vragenlijstonderzoek

VoorkennisCursisten dienen vertrouwd te zijn met de methodiek aangebracht in ‘Fundamentele Statistische technieken’ ende cursus ‘Regressie- en variantie analyse’.

LesgeversMarlies Lacante en Kristel Hoydonckx


Datum 9 en 10 mei 2017, van 09.00 tot 12.0011 mei 2017, van 09.00 tot 12.00 en van 13.00 tot 16.00

TaalNederlands



andere (PhD) studenten: € 160• Non profit/sociale sector € 250• Private sector € 600

Inleiding tot enquêtering


35

NEW

Course outlineThere is a growing acknowledgement of the value ofBayesian methods for modelling complex data structuresin basically every application area. This course willintroduce the essentials of Bayesian ideas, emphasizingthe practical application using simulation-based SASsoftware, such as the procedures BCHOICE, FMM,GENMOD, MI, MIXED LIFEREG, PHREG and MCMC.Examples will include the use of Bayesian methods inhierarchical models, measurement error models, longitudinalstudies, missing data problems, survival models, etc.The course is primarily based on the Wiley book ofLesaffre and Lawson (2012): Bayesian Biostatistics.

DAY 1:

• Morning: Introduction to Bayesian concepts.• Afternoon: Sampling with SAS.

DAY 2:

• Morning: Gibbs sampling to obtain the posteriorsummary measures.

• Afternoon: Bayesian analysis of generalized linearmodels using PROC GENMOD. Parametric Bayesiansurvival models using PROC LIFEREG and PROCPHREG.

DAY 3:

• Morning: Metropolis(-Hastings) sampling to obtain theposterior summary measures.

• Afternoon: Bayesian analysis of complex datastructures using PROC MCMC.

Target audienceSAS users who wish to understand Bayesian analysismore thoroughly.

PrerequisitesBasic mathematical skills such as derivatives andintegration are needed. Also an intermediate level ofstatistical knowledge is required, such as knowledge onregression models is mandatory. This course has anintermediate to advanced level.

PresenterEmmanuel Lesaffre and Kris Bogaerts


Date29, 30 and 31 May 2017, from 09.00 to 12.00 and from13.00 to 16.00

LanguageEnglish




Bayesian data analysis with SAS


36

Consulting was our historical embryo and remains a core business.

The statistical consulting service center acts as the main pivot to determine the ideal combination between the customer

and the most appropriate university entity.

We recommend that you contact us in an early stage of your project and write a short description of your problem

and send it to [email protected].

The LStat Statistical consulting Service covers

• Statistical support for researchers within the university. The LStat provides statistical help for university research

groups and for the central administration of the university. We help you with advice and we offer support with the design

of your study and with the statistical analysis of your data whether elementary or sophisticated. The first hour of

first-line consulting is provided free of charge.

• Statistical service and execution of projects for government and industry, in service or in partnership.

The LStat uses the administrative help of Leuven Research and Development (LRD) in the negotiation of the contracts

with industries and the private sector.

We have experience with major financial companies, inter national institutions, manufacturers, medical organizations,

marketing companies, FMCG as well as small and medium-sized enterprises.

Our solutions range from basic regression, multivariate techniques, analysis of variance, mixed models, data mining,

process control, to risk theory, categorical data analysis, longitudinal data analysis and tailor-made simulations

and calculations.

If you have any questions, do not hesitate to contact us at: [email protected] and take a look at our website:

http://lstat.kuleuven.be/consulting/

Statistical consulting Service


37

Registration costs The indicated prices correspond to training for oneperson.

There are several fee categories:

• (PhD) students KU Leuven and AssociationKU Leuven € 50/full day

• Staff KU Leuven and Association KU Leuven,other (PhD) students € 80/full day

• Non-profit sector, Social sector € 125/full day • Private sector € 300/full day

Prices include all course material.

DiscountWhen you register for several courses, you can get adiscount of 10% if the total number of full training daysequals or exceeds 5 days per person and a discount of20% is attributed if you follow courses for at least 10 fulldays.

How to register• If you are a KU Leuven staff member: via KU Loket

(Personnel – Staff training – Research)

• For all others: complete the online registration formavailable on our website:https://lstat.kuleuven.be/forms/Registration2016-2017

You can register for several courses at the same time,or complete different forms throughout the year for one ormore courses.

Confirmation and payment You will receive a confirmation mail with payment detailsand additional information shortly after receipt of yourregistration. Please contact us in case you do not receiveany confirmation mail.

Payment needs to be settled before the start of thecourse.

Cancellation• If you are unable to attend a course for which you have

registered, a colleague can take your place. Kindly informus timely of any change.

• Cancellation without charges is possible up to 2 weeksbefore the start of the course. After that, the full coursefee will be due and invoiced.

• We reserve ourselves the right to cancel a course if theminimum number of participants is not reached or inexceptional circumstances (i.e. illness presenter,extreme weather conditions etc). Registeredcandidates will be informed timely of any cancellation.In such case, registration fees will not be due andreimbursed if already paid.

Practical Matters

InformationFor questions on registration and extra informationcontact:

tel. + 32 16 32 88 75 [email protected]

Direct link to short courses:https://lstat.kuleuven.be/training


38

Course timetable 2016-2017

DATE TITLE PRESENTERS LEVEL AND MORE LANGUAGE ON PAGE

September 2016 27, 28, 29, 30 September Essential Tools for R N. Basic (English) 6

October 2016 3, 4, 5 October Fundamentele statistische Dominik Sznajder Basis (Nederlands) 7methoden

11, 18 October Models for Longitudinal Geert Molenberghs Advanced (English) 89 November and Incomplete Data Geert Verbeke1 December (Web lecture with introductory

and Q&A sessions)

13 October Fundamentele statistische Martine Beullens Basis (Nederlands) 10methoden, toegepast metSAS Eguide

13 October Fundamental Statistical N. Basic (English) 11Methods, applications with R

17, 18 October Statistical Machine Learning Jan Wijffels Advanced (English) 12with R

18 October Fundamentele statistische Marlies Lacante Basis (Nederlands) 10methoden, toegepast met SPSS

24, 25 October Text mining with R Jan Wijffels Intermediate (English) 14

26, 27 October Regression and Analysis N. Basic (English) 13of variance

November 2016 7 November Regressie-en variantieanalyse An Carbonez Basis (Nederlands) 15toegepast met SPSS

14 November Regressie-en variantieanalyse Martine Beullens Basis (Nederlands) 15toegepast met SAS Eguide

15, 17 November Regression and Analysis of N. Basic (English) 16Variance, applications with R

17, 18 November Optimization and Numerical Geert Molenberghs Advanced (English) 17Methods in Statistics Francis Tuerlinckx(Web lecture with Katrijn Van DeunQ&A sessions) Tom Wilderjans

22, 23 November Uitbreiding bij regressie- An Carbonez Verdiepend (Nederlands) 18en variantieanalyse Marlies Lacante

24, 25 November Introduction to JMP José Nunez Ares Basic (English) 19

30 November Logistic regression models Anne-Marie De Meyer Intermediate (English) 202 December with SAS, SPSS and R

December 2016 6 December Niet-parametrische statistiek Marlies Lacante Basis (Nederlands) 21

8, 9 December Applied spatial modelling Jan Wijffels Intermediate (English) 22with R


DATE TITLE PRESENTERS LEVEL AND MORE LANGUAGE ON PAGE

February 2017 14, 21 February Experimental Design Peter Goos Intermediate (English) 237, 14, 21 and 28 March18, 25 April2, 9 May

NEW 14, 21, 28 February Nonparametric smoothing Irène Gijbels Advanced (English) 247, 14 March techniques and applications

15, 22 February Chemometrics Wouter Saeys Advanced (English) 258, 15, 22 March

20, 21, 22, 23 February Essential Tools for R Dominik Sznajder Basic (English) 6

21, 22 February Advanced R programming Jan Wijffels Intermediate (English) 26

March 2017 7, 9, 10 March Cluster analysis, principal Anne-Marie De Meyer Intermediate (English) 27component analysis and An Carbonezexploratory factor analysiswith SAS Eguide, SPSS and R

NEW 9, 10 March Computer Vision with R Jan Wijffels Intermediate (English) 28and Python

14, 15, 16 March Fundamental statistical Marlies Lacante Basic (English) 29methods

14, 21, 28 March Concepts of multilevel, Geert Verbeke Advanced (English) 3018, 25 April longitudinal and mixed models

23, 24 March Text Mining with R Jan Wijffels Intermediate (English) 14

28, 30 March Introduction to the analysis An Carbonez Basic (English) 31of contingency tables

NEW 25, 26, 27 April Extremes Jan Beirlant Advanced (English) 32Tom Reynkens

25 April Survey Sampling Methods Geert Molenberghs Intermediate (English) 339, 18 May (web lecture with face-to-face

and Q&A sessions)

May 2017 9, 10, 11 May Inleiding tot enquêtering Marlies Lacante Basis (Nederlands) 34Kristel Hoydonckx

NEW 29, 30, 31 May Bayesian data analysis Emmanuel Lesaffre Advanced (English) 35with SAS Kris Bogaerts

39


40



LEUVEN STATISTICSRESEARCH CENTRE

Celestijnenlaan 200 B, room 02.083001 HEVERLEE, België

tel. + 32 16 32 88 [email protected]

v.u.

: P

rof.

M.

Laca

nte,

Leu

ven

Sta

tistic

s R

esea

rch

Cen

tre,

Cel

estij

nenl

aan

200

B,

3001

HE

VE

RLE

E,

Bel

gië


Documents

Leuven STATistics STATe of the Art Training Initiative 2016-2017.pdf · 2016-09-06 · 1 Leuven STATistics STATe of the Art Training Initiative 2016-2017 7977_BRO_LSTAT_2016-2017.qxp_7977_BRO_LSTAT_2016-2017