ENBIS-13 in Ankara

15 – 19 September 2013 Abstract submission: 5 February – 5 June 2013

My abstracts


The following abstracts have been accepted for this event:

  • Full Type 2 ... Full Type n Fuzzy System Models.

    Authors: Burhan Turksen (TOBB University of Economics and Technology)
    Primary area of focus / application: Other: Fuzzy
    Keywords: Fuzzy systems, Full type 2 fuzzy systems, Full type n fuzzy systems, Çelikyılmaz-Türkşen and Bezdek indices
    Submitted at 3-May-2013 13:49 by Burhan Turksen
    16-Sep-2013 15:05 Full Type 2 ... Full Type n Fuzzy System Models.
    We first present a brief review of the essentials fuzzy system models: Namely (1) Zadeh`s rulebase model, (2) Takagi and Sugeno`s model which is partly a rule base and partly a regression function and (3) Türkşen fuzzy regression functions where a fuzzy regression function correspond to each fuzzy rule. Next we review the well known FCM algorithm which lets one to extract Type 1 membership values from a given data set for the development of Type 1 fuzzy system models as a foundation for the development of Full Type 2 fuzzy system models. For this purpose, we provide an algorithm which lets one to generate Full Type 2 membership value distributions for a development of second order fuzzy system models with our proposed second order data analysis. If required one can generate Full Type 3,..., Full Type n fuzzy system models with an iterative execution of our algorithm. We present our results graphically for TD-Stockprice data with respect to two validity indeces, namely: 1) Çelikyılmaz-Türkşen and 2) Bezdek indices.
  • Comparing Performances of Clements’, Box-Cox, and Johnson Transformation Methods for Weibull Distributions

    Authors: Özlem Şenvar, Bahar Sennaroğlu
    Primary area of focus / application: Process
    Keywords: Pprocess capability, Process capability indices, Weighted variance method, Weibull distribution
    Submitted at 8-May-2013 10:46 by Özlem Şenvar
    Accepted (view paper)
    17-Sep-2013 10:50 Comparing Performances of Clements’, Box-Cox, and Johnson Transformation Methods for Weibull Distributions
    Weibull distributions are known to have significantly different tail behaviours, which greatly affect the process capability. In this study, Weibull distributions with different parameters are executed for examining the impact of non-normal data on the process performance index (PPI) Ppu. In the application study, PPI Ppu is utilized for process capability analysis (PCA) because Weibull data are generated without sub-groups having long term standard deviation estimate. This study is aiming to examine Clements’, Box-Cox, and Johnson transformation methods for Weibull distributions with different parameters and evaluate their estimation performances on the targeted PPI Ppu values of 1.0 and 1.5 for the issues of comparison. The performance comparisons of the three methods are performed through randomly generated data from Weibull distributions with different parameters using root-mean-square deviation (RMSD), which is used as a measure of error, and radar chart. The results of the application study show that the performance of a method depends on its capability to capture the tail behaviour of the Weibull distribution and targeted values of the PPI Ppu. It is observed that the effect of tail behavior is more significant when the process is more capable.
  • Optimum Factors for Adhesion Power

    Authors: Atakan Erdemgíl (Arçelik Elektronik İşletmesi)
    Primary area of focus / application: Six Sigma
    Keywords: DOE, Adhesion Power, VOC, SIPOC, Brainstorming, Gage R&R, DMAIC
    Submitted at 15-May-2013 21:45 by ATAKAN ERDEMGİL
    Having operations in durable consumer goods industry, Arçelik A.Ş. offers products produced in 14 production facilities with10 brands in 5 countries. In this study, a six sigma project conducted at the Arçelik television production facility is considered. Due to general safety rules, television cables should be arranged properly to avoid any problems. However, due to the number of cables used in televisions, this is a difficult task. Tapes are used to fasten the cables tightly. Nevertheless, defects caused by low adhesion power of the tapes was found to an important area for improvement. For improvement, quality tools such as Voice of Costumer, SIPOC, Priority Matrix, Brainstorming, Gage R&R and DOE are used. In the presentation, our quality improvement project will be discussed under the DMAIC framework and recommendations based on our experience will be provided for the practitioners.
  • Contact Management of Portuguese Banks through Business Intelligence Methods

    Authors: Ali Aydın KOÇ (Department of Statistics, Hacettepe University), Mustafa Murat Arat (Department of Statistics, Hacettepe University), M. Özgür Yeniay (Department of Statistics, Hacettepe University)
    Primary area of focus / application: Business
    Keywords: Artificial Neural Networks, Support Vector Machines, Logistic Regression, Classification, Marketing
    Submitted at 21-May-2013 12:58 by Ali Aydın KOÇ
    16-Sep-2013 11:20 Contact Management of Portuguese Banks through Business Intelligence Methods
    This study was aimed to compare the classification effectivities of artificial neural network, support vector machines and logistic regression analysis. Comparison of artificial neural network, support vector machines and logistic regression analysis techniques was carried out according to the marketing campaigns result‘s classification ratios.
    There are many classification algorithms in the literature, though. We briefly introduce the techniques and discuss the advantages and disadvantages of these three methods through an application with real-world data set related with direct marketing campaigns of a Portuguese banking institution. The classification goal is to predict if the client will subscribe a term deposit or not after campaigns.
  • The Garden Sprinkler: An Interactive Web-based Application for Teaching Response Surface Methodology

    Authors: Koen Rutten (KU Leuven), Karl Siebertz (Ford Forschungszentrum Aachen GmbHKU Leuven), David van Bebber (Ford Forschungszentrum Aachen GmbH), Thomas Hochkirchen (Ford Werke GmbH), Bart De Ketelaere (KU Leuven)
    Primary area of focus / application: Design and analysis of experiments
    Keywords: Teaching statistics, Design of Experiments, Interactive teaching, Simulation, Education
    Submitted at 23-May-2013 14:14 by Koen Rutten
    18-Sep-2013 09:40 The Garden Sprinkler: An Interactive Web-Based Application For Teaching Response Surface Methodology
    Most exercises and textbooks for teaching students and professionals Response Surface Methodology deal with setting up the experiment itself or the analysis of the data after the fact. As such, the whole experience of “learning-by-doing” and drawing intermediate conclusions that lead to follow up experimentation is lost. For this reason an interactive web-based application was written, based on a real-life engineering problem that is described in Siebertz et al (2010). The application deals with the multi-objective optimization of a garden sprinkler having 8 design parameters and 3 responses. Participants interact with the tool to get the response data for their own designs. Depending on the choice of the considered response(s), different design parameters are active and should be screened for. By inserting a (screening) design into the application, the corresponding responses are generated and can be used to analyze the data and plan follow up experiments (e.g. limited and full response surface experimentation) to find the optimum. This tool has been successfully used both in company training and academic courses. Any statistical software can be used to generate and analyze the designs and does not impact the usability of the application.
  • Graphical Models for Fault Detection and Diagnosis of Complex Systems, with an Application to Hybrid Vehicles

    Authors: Diana Flaccadoro (University of Genova (DIMA) / National Research Council of Italy (ISSIA)), Cristiano Cervellera (National Research Council of Italy (ISSIA)), Giorgio Bosia (Eutecne Srl), Eva Riccomagno (University of Genova (DIMA))
    Primary area of focus / application: Reliability
    Keywords: Conditional independence, Chain graph, D-separation, Markov properties, Graphical modelling, Elicitation
    Submitted at 23-May-2013 15:22 by Diana Flaccadoro
    Accepted (view paper)
    17-Sep-2013 16:35 Graphical Models for Fault Detection and Diagnosis of Complex Systems, with an Application to Hybrid Vehicles
    Graphical models are graphs in which nodes represent random variables and arcs (undirect or direct) represent probabilistic relationship of conditional independence. In particular, graphical causality models are a special kind of graphical models in which the direction of arcs stands for a cause-effect relationship. This makes them suited to represent the behavior of a complex system that is difficult to model through mathematical equations. In this work we exploit this possibility in a context of diagnostics and fault detection. Specifically, we reduce the fault detection problem to the evaluation of a conditional distribution. We show how to derive this conditional distribution, using available data, from the analysis of a suitable graphical model taking advantage of the Markov properties.
    As a case study we consider diagnostics of a hybrid bus, characterized by the combined use of a Diesel and an electrical engine. The aim of the study is to verify the correct operation of the electrical system. This procedure underlines the potentiality of graphical models, by which the study of a complex system can be reduced to an easier analysis of the topology structure of a graph.