Contents & media
European Network for Business and Industrial Statistics
You're NOT logged in. LOGIN Events Centre
You're NOT logged in. LOGIN Events Centre
Contents & media
European Network for Business and Industrial Statistics
Home > Activities > Past Events > Past Annual Conferences > ENBIS-20 Online Conference
ENBIS hosts an annual conference with presentations from a wide variety of sectors, ranging from manufacturing to service and from private to public. Statistical practitioners and academic statisticians from Europe and beyond meet, exchange ideas, and design new projects in this conference. The annual ENBIS conferences feature both theoretical and practical studies in the area of business and industrial statistics.
Due to the COVID-19 sanitary emergency, the face-to-face conference in Valencia, planned for the end of September, had to be postponed. But we wanted to give all ENBIS members the opportunity to also meet this year – although virtually – and attend interesting presentations. So, during the same days when the face-to-face conference was planned, ENBIS will offer an online conference with first class keynotes, on topics very relevant to the ENBIS community. Also, this year award recipients will deliver a presentation, and they will “virtually” receive their award (while we wait to “physically” give the awards in the next face-to-face ENBIS conference). The conference will include two active sessions, the Greenfield Challenge and a hands-on workshop using JMP.
The presentations took place from Monday 28 September 2020 to Wednesday 30 September 2020. The conference has only one track (no parallel sessions), so you will be able to attend all sessions. The conference will be streamed live at a time that is – we hope – the best possible for ENBIS members (afternoon in Europe, morning in America, evening-night in Asia). The online workshop will be offered on Thursday 1 October 2020 using JMP, with the title “All models are wrong – but which are useful?”
We invite you to check the ENBIS-20 Online Conference programme. You’ll be able to register very soon. Although registration will be required to attend the conference, the ENBIS-20 online conference is free for all ENBIS members.
“IMPROVING” PREDICTION OF HUMAN BEHAVIOR USING BEHAVIOR MODIFICATION
Galit Shmueli, PhD, Editor of INFORMS Journal on Data Science, and Tsing Hua Distinguished Professor at National Tsing Hua University, Taiwan
Berlin, Paris: 15:15 / London: 14:15 / New York: 9:15 am / Los Angeles: 6:15 am / Ciudad de México, Lima: 8:15 am / São Paulo: 10:15 am / Beijing: 21:15
Abstract:
The fields of machine learning and statistics have invested great efforts into designing algorithms, models, and approaches that better predict future observations. Larger and richer data have also been shown to improve predictive power. This is especially true in the world of human behavioral big data, as is evident from recent advances in behavioral prediction technology. Large internet platforms that collect behavioral big data predict user behavior for their internal commercial purposes as well as for third parties, such as advertisers, insurers, security forces, and political consulting firms, who utilize the predictions for user-level personalization, targeting and other decision-making. While machine learning algorithmic and data efforts are directed at improving predicted values, the internet platforms can minimize prediction error by “pushing” users’ actions towards their predicted values using behavior modification techniques. The better the internet platform is able to make users conform to their predicted outcomes, the more it can boast both its predictive accuracy and its ability to induce behavior change. Hence, internet platforms have a strong incentive to “make the prediction true”, that is, demonstrate small prediction error. This strategy is absent from the machine learning and statistics literature. Investigating the properties of this strategy requires incorporating causal terminology and notation into the correlation-based predictive environment. However, such an integration is currently lacking. To tackle this void, we integrate Pearl’s causal do(.) operator to represent and integrate intentional behavior modification into the correlation-based predictive framework. We then derive the expected prediction error given behavior modification, and identify the components impacting predictive power. Our formulation and derivation make transparent the impact and implications of such behavior modification to data scientists, internet platforms and their clients, and importantly, to the humans whose behavior is manipulated. Behavior modification can make users’ behavior not only more predictable but also more homogeneous; yet this apparent predictability is not guaranteed to generalize when the predictions are used by platform clients outside of the platform environment. Outcomes pushed towards their predictions can also be at odds with the client’s intention, and harmful to the manipulated users.
Biography:
Galit Shmueli is Tsing Hua Distinguished Professor at the Institute of Service Science, National Tsing Hua University, Taiwan. She is also Director of the Center for Service Innovation & Analytics at NTHU’s College of Technology Management. Dr. Shmueli’s research focuses on statistical and data mining methodology with applications in information systems and healthcare, and an emphasis on human behavior. She authors multiple books, including the popular textbook Data Mining for Business Analytics and over 100 publications in peer-reviewed journals and books. She has presented her work at multiple venues in the US and internationally. Dr. Shmueli is the inaugural editor-in-chief of the new INFORMS Journal on Data Science. She has chaired and served on many program committees of top conferences and workshops, and is an IMS Fellow and ISI elected member.
Berlin, Paris: 16:15 / London: 15:15 / New York: 10:15 am / Los Angeles: 7:15 am / Ciudad de México, Lima: 9:15 am / São Paulo: 11:15 am / Beijing: 22:15
Abstract:
It is rightly pointed out that in the midst of a pandemic crisis of enormous proportions we needed high-quality statistics with extreme urgency, but that instead we are in danger of drowning in an ocean of data and information. Rarely has the lack of adequate statistics to make essential political decisions and to win popular support for their consequences been as visible and painful as it is now. Rarely have governments invested so much public money to combat the health, social and economic consequences of a crisis. The question is whether these monumental financial support programmes are associated with a direction or a mission, whether the investments are used for innovation in the sense and for the goals of “entrepreneurial states”. At this moment of confusion and in the search for orientation, it seems appropriate to take inspiration from previous initiatives in order to draw lessons for the current situation. More than 20 years ago in the United Kingdom, the report “Statistics – A Matter of Trust” laid the foundations for overcoming the previously spreading crisis of confidence through a soundly structured statistical system. This report is not alone in international comparison. Rather, it is one of a series of global, European and national measures and agreements which, since the fall of the Berlin Wall in 1989, have strengthened official statistics as the backbone of politics in democratic societies, with the European Statistics Code of Practice being an outstanding representative. Therefore, if we want to address our current difficulties, the following three questions should address precisely those points that have emerged as determining factors for the quality of statistics: What (statistical products, quality profile)? How (methods)? Who (institutions)? The aim must be to ensure that the statistical information is suitable to facilitate the resolution of conflicts by no longer arguing about the facts and only about the conclusions to be drawn from them.
Biography:
Walter J. Radermacher was the Director General of Eurostat and Chief Statistician of the European Union from 2008 to 2016. He worked at Destatis, the German Federal Statistical Office, for 30 years, ultimately as its President and Federal Returning Officer. He was the first Chair of the UN Committee of Experts on Environmental-Economic Accounting (UNCEEA) from 2005 to 2008. Since 2017 he has been a Researcher at the Department of Statistical Sciences, Sapienza University of Rome, and the President of FENStatS, the Federation of European National Statistical Societies.
Berlin, Paris: 16:15 / London: 15:15 / New York: 10:15 am / Los Angeles: 7:15 am / Ciudad de México, Lima: 9:15 am / São Paulo: 11:15 am / Beijing: 22:15
Abstract:
What do decision makers want from statisticians? What do I want from Stina or David, when I ask them for an analysis? We will have a glimpse into the dreams and headaches of managers, look at a few examples of statistical analyses from the manager’s side, and put them into a context of decision theory.
Biography:
Antje has been working in the pharmaceutical industry with Novo Nordisk for the past 23 years, first as a statistician, then in project management within optimisation of a broad range of production processes. This long-term work relationship included a period of product lifecycle management, and followed a short stint into opinion polling. Antje holds a PhD and MSc in mathematics from Ruhr University Bochum in Germany, a degree in product innovation from the Scandinavian International Management Institute in Copenhagen, Denmark, as well as a Six Sigma black belt certification from the American Society for Quality, and company internal certifications in Six Sigma and Scrum. She has been involved with ENBIS since its foundation in 2000, and received the ENBIS Best Manager Award in 2015.
Berlin, Paris: 16:15 / London: 15:15 / New York: 10:15 am / Los Angeles: 7:15 am / Ciudad de México, Lima: 9:15 am / São Paulo: 11:15 am / Beijing: 22:15
Abstract:
Industrial statisticians played an important role in the success of the Industrial Revolution. The analytical methods developed in our field have been used to leverage data from machines and workers to improve processes, safety, and products for nearly 100 years. We are now in the midst of the Information Age, and the data revolution brings the challenges and opportunities of our time. Our success in the data revolution rests in our ability to leverage the new scale, scope, and type of data to improve processes, safety, products, health, and services. We have tremendous opportunities and tremendous challenges. An important opportunity for our field is to leverage data that emerges from sensors. Many low-cost, multifunction, wearable devices have been developed. The data from these devices have been used, for example, in recreational, health, productivity, and safety monitoring. This type of data is high frequency, noisy, and follows specific periodic patterns that depend on what is being monitored and the context of the monitoring. In this talk, we will explore the challenges associated with monitoring gait patterns using data from low-cost wearable Inertial Measurement Units (IMUs). The goal of the analysis is to understand how changes in gait patterns relate to fatigue with the end goal of automatically detecting fatigue in an industrial setting. Although many have used sensor data for gait analysis, most treat this as a classification problem, using either statistical or machine learning methods for binary classification. The classification approaches are generally supervised and require a large number of participants with labeled cases of non-fatigued and fatigued periods. Our research team approached this differently by developing personalized monitoring schemes for each individual. Several univariate, multivariate and profile statistical process monitoring methods were explored. The team developed personalized monitoring frameworks based on modifications of classical univariate and multivariate control chart methods and supplemented these frameworks with straightforward diagnostic information. Although it is intuitively appealing, using the classical methods on this new data brings new and unexpected challenges. With these challenges come many opportunities for improved methodology and research in process monitoring related to sensor data. This talk is based on joint work with Saeb Ragani Lamooki, Fadel M. Megahed, Jiyeon Kang, and Lora A. Cavuoto.
Biography:
Allison Jones-Farmer is the Van Andel Professor of Business Analytics at Miami University in Oxford, Ohio. Her research focuses on developing practical methods for analyzing data in industrial and business settings. She is on the editorial review board of Journal of Quality Technology and Quality Engineering. She is a current section editor for Journal of Quality Technology and a former Associate editor for Technometrics. In addition to her research in industrial analytics, Allison enjoys helping organizations improve their analytics capability, developing innovative curricula, and teaching data science. Prior to joining Miami University, Allison was a Professor of Statistics and Analytics at Auburn University where she held the C&E Smith chair. She received a B.S. in Mathematics from Birmingham-Southern College, an M.S. in Applied Statistics from the University of Alabama, and a Ph.D. in Applied Statistics from the University of Alabama.
Berlin, Paris: 15:00 / London: 14:00 / New York: 9:00 am / Los Angeles: 6:00 am / Ciudad de México, Lima: 8:00 am / São Paulo: 10:00 am / Beijing: 21:00
A short presentation on Wiley’s Statistics journals programme. The session will also cover practical ways for authors to maximise the impact of their articles.
The session will be presented by Stephen Raywood, senior journals publishing manager of physical sciences at Wiley in the UK.
Berlin, Paris: 17:00 / London: 16:00 / New York: 11:00 am / Los Angeles: 8:00 am / Ciudad de México, Lima: 10:00 am / São Paulo: 12:00 am / Beijing: 23:00
Abstract:
In this session, you will explore and discuss two fresh open problems. Two volunteers will briefly present an open case they are working on, you’ll ask questions and give advice, and Christian will facilitate the meeting to make sure that everybody contributes. It’s a session type we experimented with at a few ENBIS conferences which gives participants an opportunity to enter much more deeply into the subjects and we decided to try it out online this time. Be curious …
Practical information: We are looking for two volunteers to introduce projects they are working on and on which they would like to receive guidance and advice by the ENBIS audience. If you are currently working on a subject and aren’t sure how to deal with it or how to do it better, let Christian know and he will see if it could be a case for the ENBIS live session. If selected, you would be asked to briefly introduce the case during the session (7 minute maximum). You can either do it live or via a recorded mini-presentation. Then you should be ready to answer questions from the audience and to actively follow the discussions.
Biography:
Christian Ritter is an independent statistical consultant and part time. Over 30 years of professional experience, he contributed to hundreds of projects covering experimental designs, exploratory analysis, modeling, reporting and communication. His special interests include computational statistics and simulation applied to industrial statistics, statistical graphics, and of course, improvisation and interaction in teaching and consultancy. 25 years ago, he adapted the Wisconsin course on statistical consulting to the context of UCLouvain and has been teaching it since then. Five years ago, he created an intensive course called ‘Junior Data Analyst’ which gives quantitative-minded people with degrees in human and social sciences an opportunity to enter careers in data science. It now graduates more than 30 participants per year who fill positions in many different businesses and public services. The key to success in this training is the combination of communication and analytical skills. Facilitating communication to enhance analytics is also the objective of ENBIS Live.
Sonja Kuhnt, PhD, Professor of Mathematical Statistics at FH Dortmund University of Applied Sciences and Arts, Germany |
Shirley Coleman, PhD, Technical Director at Industrial Statistics Research Unit, Newcastle University, UK |
Berlin, Paris: 17:00 / London: 16:00 / New York: 11:00 am / Los Angeles: 8:00 am / Ciudad de México, Lima: 10:00 am / São Paulo: 12:00 am / Beijing: 23:00
About the session:
Are you interested in case studies and real-world problems for active learning of statistics? Then come and join us in this interactive session organised by the SIG Statistics in Practice. A famous project for students to apply the acquired knowledge of design of experiments is Box’s paper helicopter. Although being quite simple and cheap to build, it covers various aspects of DoE. Beyond this, what other possible DoE projects are realistic in a teaching environment? What are your experiences in using them? Can we think of new ones? There are lots of ideas we could explore, involving more complex scenarios like time series dependents with cross overs, functional data analysis, as well as mixture experiments. We want to share projects, discuss pitfalls and successes and search our mind for new ideas. Come and join us for this session. You may just listen, enjoy and hopefully contribute to the discussion or even share a project idea. Please send an email to Sonja (sonja.kuhnt@fh-dortmund.de) or Shirley (shirley.coleman@newcastle.ac.uk) if you have a hands-on project and are willing to share it with us. The project should be doable within 2-3 teaching lessons and affordable material. If selected, we ask you to introduce the case during the session briefly (5 minute maximum). You can do it either live or via a recorded mini-presentation.
Biography:
Sonja Kuhnt is a professor of mathematical statistics in the department of computer science at FH Dortmund University of Applied Sciences and Arts. She received her doctoral degree and habilitation in Statistics from TU Dortmund University. Her research interests include the design and analysis of real and computer experiments, categorical data analysis and robust statistics. She serves the statistical community in her function as director of ENBIS and by being an associate editor of CSDA and Statistical Papers.
Shirley Coleman is a Technical Director in the School of Mathematics, Statistics and Physics at Newcastle University. She received her doctoral degree in computer graphics, statistics and morphometry from Newcastle University. Her research interests include the design and analysis of experiments to tackle real challenges and find workable solutions in marine and healthcare sectors, process and manufacturing industry and digital marketing. She serves the statistical community in her function as an Honorary Officer of the Royal Statistical Society, a past president of ENBIS and contributor to academic journals and popular press.
Berlin, Paris: 15:00 / London: 14:00 / New York: 9:00 am / Los Angeles: 6:00 am / Ciudad de México, Lima: 8:00 am / São Paulo: 10:00 am / Beijing: 21:00
Volker Kraft (SAS Institute GmbH – JMP Division, Germany)
You have a business or research question, you’ve collected or found appropriate data, and you are ready to analyze. But which analytical methods should you try? And how will you choose a final – hopefully the most useful – model? In this seminar, we will look at several data scenarios and discuss modeling options and a framework for comparison. We will look at how different questions or goals affect the modeling choices we make (Predict? Explain? Find associations?). Models covered will include traditional methods like (penalized) regression, structural equation modeling or tree-based models, as well as unsupervised and supervised machine-learning tools like clustering, neural networks or support vector machines. Comparison techniques will include residual analysis, comparing fit statistics and cross-validation.
In order to support interworking with other software, we will also cover (briefly) how to import data into JMP and how to export model scoring code (e.g. as C, Python or SAS code). We will also show options to share results and visualizations from the model building and assessment phases.
The format of this course will be a mix of conceptual presentations, live demos and interactive communications. All content will be shared with the participants.
Regarding hands-on, attendees without access to JMP Pro may request a JMP Pro 15 trial version (for Windows or Mac) to explore the workshop examples at their own pace. JMP users may also join the Early Adopter 16 program to try some new features for advanced modelling which will be demonstrated as well. No prior familiarity with JMP is required.
Keywords: Predictive modeling, machine learning, exploratory modeling, model evaluation, statistical thinking.
| Volker Kraft is Senior Academic Ambassador for JMP in Europe, fostering and supporting the use of JMP in teaching and research. He holds a Ph.D. in Electrical Engineering (University Bochum, Germany), and used statistical methods extensively in his research in psychoacoustics and speech communication. In later positions at Vodafone and Voxeo he developed a passion for customer advocacy, developing and supporting programmes to foster the effective use of software in business. Drawing on this experience and the success of JMP’s Academic Program globally, Volker is charged with enabling teachers, faculty and students in Europe who are interested in JMP to get the most from the software. |