ENBIS: European Network for Business and Industrial Statistics

My abstracts

The following abstracts have been submitted for this event:

• An extendable Java-based tool for time series analysis
  
  Authors: G. Albeanu, H. Madsen, Fl. Popentiu-Vladicescu, Marius Dumitru
  Submitted at 28-May-10 20:58 by Grigore ALBEANU
  Accepted (view paper) Time series analysis is an important topic both for teaching statistics and research. Previous results in time series analysis methodologies [1, 2, 3], relevant applications and software development for time series analysis motivate us to promote an extendable Java-based tool for time series analysis, recently developed. Implemented methods, the extension methodology, obtained results in teaching time series and research will be detailed.
  
  References
  1. G. Albeanu, H. Madsen, B. Burtschy, Fl. Popentiu-Vladicescu, Manuela Ghica, Bootstrapping time series with application to risk management. R & RATA, Electronic Journal of International Group on Reliability, ISSN 1932-2321, Vol 1(3), pp. 84-93, 2008.
  2. Madsen, G. Albeanu, B. Burtschy, F. Popentiu-Vladicescu. Addressing Time Series Modelling, Analysis and Forecasting in e-learning Environments. In Marin Vlada, Grigore Albeanu, Dorin Mircea Popovici (eds.), Proceedings of the 1st International Conference on Virtual Learning, Bucharest, pp. 37-44. 2006.
  3. G Albeanu, L. Serbanescu, Fl. Popentiu-Vladicescu. On teaching data analysis and optimisation using software tools. In Grigore Albeanu, Dorin Mircea Popovici , Marin Vlada (eds.), Proceedings of the 2nd International Conference on Virtual Learning, Constata, 26-28 October, Bucharest, ISBN: 973-737-218-2 978-973-737-380-9, Romania. , I, pp. 255-260. 2007.
• Sequential choice designs to estimate the distribution of willingness-to-pay
  
  Authors: Vishva Danthurebandara; Jie Yu; Martina Vandebroek
  Affiliation: Katholieke Universiteit Leuven Submitted at 30-May-10 00:00 by Vishva Danthurebandara
  Accepted The concept of willingness-to-pay (WTP) has attracted the attention of marketeers because of its usefulness in many applications. Nowadays one aims at describing the market heterogeneity by estimating the distribution of WTP. However, this poses several problems that have been discussed repeatedly in the literature. Many authors report unrealistic, extreme or inaccurate individual-level WTP estimates.
  We propose to use an adaptive sequential approach to construct conjoint choice designs for estimating the distribution of WTP. It uses Bayesian methods to generate individually optimized choice sets. These choice sets are computed sequentially based on the prior information of each individual which is updated after each choice. The choices made by all respondents are then used to estimate the mixed logit model which yields individual-level utility coefficients and corresponding individual-level WTP estimates from which the distribution of WTP can be derived.
  This sequential approach is compared in a simulation study with two non-sequential designs: a semi-Bayesian D-optimal design for the conditional logit model and a nearly orthogonal design. The results show that the sequential design performs much better than the benchmark designs. It yields more accurate individual-level WTP estimates and produces a more accurate picture of the heterogeneity.
• The idea of an expert system to evaluate process state
  
  Authors: Agnieszka Kujawinska, Maria Pilacinska, Michal Rogalewicz
  Affiliation: Poznan University of Technology Submitted at 31-May-10 00:11 by Michal Rogalewicz
  Accepted The subject of the article is a diagnosis and assessment of process’ capability to meet quality requirements. In classical approach process quality evaluation is made by measurement of earlier specified features (mostly features critical to quality) of the product and computing so called capability indices (eg. Cp, Cpk, DPMO, ppm). A measurement can be made during the process or after it and information about the level of quality can have an influence on further runs of this process. In the first case it is called “quality control” because the process is controlled through a feedback loop, the second case is just an ordinary quality inspection.
  Described above approach lets one make an evaluation of the process taking into account only one feature. So for the sake of any quality feature one has to specify separate quality index or design separate quality control chart.
  More and more frequently scientists and practitioners try to work out and implement process state evaluation tools determined basing on many features simultaneously. An effect of these activities are statistical tools which enable to monitor process’ stability for some features in aggregated way. eg. multivariate quality control charts. These tools do not solve another problem. They do not make one possible to:
  1) specify an optimal value (or range of values) of input process parameters (eg. range of cutting speed) so that values of features were on established level
  2) specify an optimal values of set of input process parameters (machine, man, environment etc.) so that a value of feature (eg. membership to quality class) was on established level
  In the article an idea of expert system designed to evaluate a process state was shown. This assessment is understood as a capability index of a process and is computed on the basis of many process state measures eg. process parameters, diagnostic signals and events.
• Optimality Criterions For The Response Surface Designs
  
  Authors: Serpil Aktaş
  Affiliation: Hacettepe University Submitted at 31-May-10 09:49 by Serpil Aktas
  Accepted (view paper) Response Surface method uses a set of designed experiments to obtain an optimal response in the experimental design studies. The two most common designs generally used in response surface modeling are central composite designs and Box-Behnken designs. The optimal design is to select design points according to some criteria, some of those are D-optimality and the distance based optimality. D-optimality design minimizes the volume of the confidence ellipsoid of the regression estimates of the linear model parameters. The design spreads the design points uniformly over the design space in the distance based optimality. In this case, this method provides one solution for selecting the design points and the distance-based optimality algorithm selects design points from the candidate set, such that the points are spread evenly over the design space. Distance based-optimality selects the candidate point with the largest Euclidean distance from the origin or the point that is closest to a pure component as the starting point. The aim of the selection process is to reduce the number of experimental runs. In this study the optimality criterions are studied and which is the more efficient approach for achieving a robust design compared with the classical full factorial design approach was investigated both central composite and Box-Behnken designs.
• The challenges of a statistical consulting service inside the university: an example of university-industry partnership
  
  Authors: Celine Bugli Bernadette Govaerts
  Affiliation: SMCS, ISBA, Université Catholique de Louvain Submitted at 31-May-10 10:43 by Celine Bugli
  Accepted There are more and more universities that provides statistical advice and support to clients within and outside of the university, usually with the participation of graduate students. In the Université Catholique de Louvain, a technological platform named SMCS is offering Statistical Methodology and Computing Support to the UCLouvain community and external clients from industry.
  
  An effective partnership must benefit both participating parties. For industry, the benefits should have demonstrable financial impact, preferably in the short term. For universities, the benefits may be less tangible but no less real.
  
  In this talk, we will present examples of formal partnerships and informal cooperative arrangements between university and industry through the experience of the SMCS. We will also discuss factors that can affect the success of a university-industry partnership, working methods, expected benefits of such relationship, potential obstacles and their resolution.
• Calibration between internal and external quality assessment: a comparison of two non parametric approaches
  
  Authors: Paola Cerchiello Paolo Giudici Emanuela Raffinetti
  Affiliation: University of Pavia Submitted at 31-May-10 11:20 by Paola Cerchiello
  Accepted In this contribution we present two novel non parametric approaches useful to describe and rank quality evaluations expressed either by means of a qualitative or a quantitative scale. The two indexes are respectively based on the frequency quantiles of the given variable and on intensity measures such as the Lorenz concentration curve and the dual Lorenz curve. In the latter case, the definition of a k-variate concordance index is obtained by exploiting a ranks-based approach. The proposed methodology allows us to compare and match, in terms of concordance and discordance, quality evaluations expressed by “internal” and “external” assessors (for example with regards to academic and\or school performance). Thus the final aim is to calibrate and validate judgments expressed by different groups of evaluators.