Exploitable results
Several control charts for individual observations are compared. Traditional ones are the well-known Shewhart individuals control charts based on moving ranges. Alternative ones are non-parametric control charts based on empirical quantiles, on kernel estimators, and on extreme-value theory. Their in-control and out-of-control performance are studied by simulation combined with computation. It turns out that the alternative control charts are not only quite robust against deviations from normality but also perform reasonably well under normality of the observations. The performance of the Empirical Quantile control chart is excellent for all distributions considered, if the Phase I sample is sufficiently large.
Joint Workshop on Statistical Datamining held 23rd & 24th April 2003 at EURANDOM, Eindhoven, The Netherlands (EURAN.ISG (CR2)). The workshop was funded jointly by Eurandom & Pro-ENBIS
EURAN.ISG (CR2).
In total there were 46 delegates (10 are members of Pro-ENBIS)
Pro-ENBIS member organisations who participated are EURAN.ISG (CR2), GREENF.CC (CR6), BUSTATS (MB7), UAM.IBIS (CR10), URJC.GECD (MB13), UEOT.DPTS (MB15), TUWRO.IPEA (MB33), KPA (MB32),
AHSDM (MB31), UPAV.EPMQ (MB36)
Background:
EURANDOM has had a long relationship to the ENBIS network and as work package leader is able to draw on its own experience in the area of statistical data mining and related areas. Indeed, EURANDOM had run two previous workshops in broadly this area and was able to draw on its experience to put on a successful workshop as a Pro-ENBIS partner. EURANDOM had, fortuitously, also made the area its leading area within the statistical programme of EURANDOM. For example there were parallel grant applications in the area. In particular, EURANDOM is a partner on the EU 6th Frame Network of Excellence grant application, PASCAL, for which contract negotiations are currently taking place.
Statistical data mining and statistical learning:
Another beneficial aspect of the involvement with data mining is the wider heading of computational and statistical learning. The subject is probably the fastest moving area of statistics. It is driven by the need to perform data analysis and modelling on very large data sets. The activity can range from relatively standard clustering methodologies to sophisticated use of kernel-based methods which are able in addition to more standard curve fitting styles are able to detect unforeseen patters in the data. A brief summary of a current review of the area is that data can be split broadly into model plus patterns plus noise.
The Meeting:
The meeting divided broadly into general theoretical and conceptual session on the first day and more practical and hands on talks on the second day, although, of course, theory and application was mixed throughout the workshop. Perhaps the most attractive feature of the workshop was the coming together of leading members of the statistical and the computer science community. Thus David Hand, Jerry Friedman and Alan Carr, each of whom described the challenges of the subject and why it was non-standard from a statistical point of view, represented the statistical community very well. Jacqueline Meulman gave an interesting talk from a more social science perspective and Petra Perner and Arno Siebes spoke more from a computer science/IT perspective.
It had always been the intention to include a practical session and Andrea Ahlemeyer-Stubbe and Elsa Jordaan represented this area well. The first in relation to marketing and the second in relationship to statistical process control. Elsa works for Dow Chemical and had just completed her PhD in the area. Other notable industrial participation was from Philips Research.
Conclusions:
This was a very well received workshop and the useful items can be listed as follows:
- Helping to establish data mining at a high technical level within the Pro-ENBIS framework.
- Establishing good links between Pro-ENBIS and the US particularly to the National Institute of Statistics (Director Alan Carr).
- Involvement of Industry.
- This was not enough but the contacts with Dow and Phillips are good and should be built upon enhancing of networking for Pro-ENBIS via EURANDOM (e.g. to Pascal).
ENBIS? What is it? Who made it? Where is it? Why is it? What does it do?
International competition is getting tougher, product development cycles shorter and manufacturing processes more complex; customers are expecting and demanding higher quality for products. Business and industrial statisticians have contributed to these changes. Fellows of the Royal Statistical Society have been to the fore, working within companies or as consultants, or through the business and industrial section and the quality improvement committee.
In spite of these efforts, there is much to be done. Indeed, there is a mutual frustration at the sight of so many companies who remain ignorant of what we might do for them. Their blind faith was displayed in a company director’s answer to my question: "What do you do about uncertainty and variation?" He replied: "They are not allowed."
There are many statisticians across Europe and beyond who know that our methods have improved business and industrial performances and can continue to do so. We know too that every opportunity must be seized and much effort must be applied to ensure that we continue to achieve continuing improvements. The Internet and easy travel within Europe provide the opportunities. These have been used to create a new society, ENBIS: The European Network for Business and Industrial Statistics. ENBIS is intended to be a forum for the dynamic exchange of ideas and to provide a networking mechanism for statistical practitioners. We aim to stimulate the application of statistical methods to enhance economic and technical development and to improve competitiveness of business and industry across the whole of Europe.
The need for networking arose from the realisation that many applied statisticians and statistical practitioners work in environments where they are isolated from interactions with, and stimulation by, likeminded professionals.
ENBIS was created by a small band of enthusiasts, led by Søren Bisgaard, and had its official launch at the first conference in December at the University of Amsterdam with more than 80 attending. The provisional executive committee had 14 members drawn from eight European countries. Everybody participated in one of six working groups: Industrial design of experiments; General statistical modelling; Data mining and data warehousing; Process modelling and control; Reliability and safety; Quality improvement.
The first conference was followed by a three-days course on design of experiments presented by Søren Bisgaard. This is a pattern that will be adopted at future conferences. For example, the second ENBIS conference in Oslo next September (17th and 18th) will be followed by a three day introductory course in Six Sigma.
Since that first meeting, solid foundations have been built: membership has grown to about 300 from more than 20 countries (mostly European but a few from the USA), a constitution has been written, financial management has been created, a secretariat established, and, most visibly, a website has been developed. Look at it: www.enbis.org. You will see reports of the working groups, information about future meetings and other activities, a report of the first conference, membership details and how to join, and membership of the executive committee of which the chairman for this year is a past president of the Royal Statistical Society, Henry Wynn.
The website has a "members only" section. You can access this with your username and password that will be sent to you when you join. A discussion page in this section enables members to post messages and to reply to other messages. There is also a network page, which displays all ENBIS members with their affiliations and email addresses. Members can also show further information about themselves such as their special interests and areas of expertise.
Each working group will also have its own discussion page to be used by group members to exchange ideas. All members can join in. Membership of ENBIS, and hence of its working groups, is open to statistical practitioners as well as to professional statisticians. By this, we mean people who use statistical methods in their work. I suspect that the discussions and information exchanges through the working group pages could develop into valuable resources for engineers, scientists and managers working in business and industry. They should be encouraged to join ENBIS.
The paper concerns the analysis and the design of the measurement process by which position tolerances on mechanical parts are checked by Coordinate Measuring Machines (CMM). This measurement process is widely used in industry and conditions the good functioning of millions of components, assemblies and systems. CMMs inspect parts by exploring their surface at a small number of points and return the point Cartesian coordinates. Then data are numerically elaborated to estimate the position error.
The analysis aims to evaluate measurement uncertainty as generated by two sources: the random error related to coordinate retrieval and the sampling error inherent to the way CMMs operate. By simulating random error via computer, the measurement process is fully recreated by a simulation model. Then extensive computer experimentation, combining Montecarlo simulation and DOE, is performed.
The study has revealed interesting statistical properties of the two-dimensional position error, which have useful practical implications and disprove a number of widely, used rules of thumb of engineers. Another contribution of the paper is the use of the uncertainty analysis to design an efficient measurement process, namely one that attains a good trade-off between cost and accuracy. For a given total number of measurement points, their optimal allocation on the different part surfaces is provided.
The Non-homogeneous Poisson Process (NHPP) is reviewed and some results regarding the Rate of Occurrence of Failures (ROCOF) are presented. The well-known ROCOFs are listed in conjunction with some not so well known ones to bring together the literature in line with non-repairable items. This review will also introduce some non-repairable distributions based on the ROCOFs. The theory is useful for modelling repairable systems reliability.
ISRU held a six-sigma event that was held at Dupont Powder Coatings in Darlington.
There were 17 delegates who attended this event. (Please see attached sheet). All were from different backgrounds such as Human Resources, Quality Management, Finance and Production.
Dave Stewardson was the first speaker he is the Director of ISRU and the president of ENBIS (European Network for Business and Industrial Statisticians). He talked of the importance of ENBIS and its role in industry. Dave also discussed how ISRU consultants could work with clients, via the use of specialised training programmes backed up by on-going business and technical support. Another support tool he talked about was the use of ISRU's distance learning modules that cover the major statistical tools involved in Six Sigma training programmes. This is a suitable option for companies who currently do not have the time or resources to embark on a major "Six-Sigma" programme but feel that their employees require an awareness and appropriate knowledge of "Six-Sigma" tools and applications.
Dr. Irena Ograjensek was the second speaker; she was from the University of Ljubljana (Slovenia) and was also representing ENBIS. Irena looked at the softer side of implementing Six-Sigma and titled her presentation "Opening up the statistical toolbox" which looked at why using statistical methods is very important in every department of a company and not just production.
Prof. Tony Greenfield was our last speaker from ENBIS. He looked at the importance of Design of Experiments and in his presentation introduced WINDEX a program for the design and analysis of experiments.
The feedback from this event was very encouraging. A company who attended by the name of Helena Bio-Sciences in Sunderland had delegates attending from their finance and quality departments, they said they found the second presentation very relevant to the job they do and now had a greater understanding of what Six-Sigma actually was.
Glaxo SmithKline found the third presentation most interesting and wanted to be contacted for an Industrial Event that ISRU organises.
Since the event Donnelleys in York expressed an interest in working with ISRU on a green-belt Six Sigma programme.
We have found these events to be extremely useful to both parties and intend to run a lot more in the very near future.
The network economy is more or less commonplace to all of us. This applies also in service business: the service production process combines several individual players, who together produce the service experience for the user. This paper presents concepts of total service offering (TSO), and service production network (SPN), pointing out the complexity of service quality.
Total service offering includes the whole service package offered to the customer. Besides the core service, it may contain several supporting and facilitating services (and goods) as well. The different parts of the Total Service Offering may be produced by various independent service providers (SP). SPs participating into the TSO production process form a service production network (SPN). When the production process of TSO is decentralised, the provider of the core service cannot directly control the service quality of TSO. In case of quality problem, the negative feedback is usually directed towards the brand owner, who usually is the core service provider as well. This means, that the task of building quality into the service production process is even more challenging and crucial task for service providers than it has been before.
The complexity of this problem is illustrated with the case of developing smart card based electronic services. Individual actors of a network focus in restricted parts of the quality being incapable to manage the Total Quality of the TSO. As a result, a framework clarifying the roles of individual actors in developing the total quality of TSO is presented.
The web page was identified as an area of weakness in the First Year Progress Report. Maintenance of the site had fallen behind in some respects. Immediate steps to rectify this were initiated by the Coordinator. This included the appointment of a dedicated Web Master for Pro-ENBIS (John Logsdon) to remedy the immediate issues, to maintain and update the web page.
A consultation visit to UNEW (CO1) took place in November 2003. The meeting was very successful and a strategy for improving the website was developed. One of the major improvements planned and implemented was the transferring of tables containing details of all the deliverables (workshops, industrial visits, journal papers, articles, software and books) onto the member area of the pro-enbis website.
John Logsdon (JL) of Quantex Research Ltd agreed to help with maintenance, with particular emphasis on the Pro-ENBIS site. This was finally put into effect in April 2003, working closely with the project coordinators at ISRU.
The Pro-ENBIS web site is a sub-directory of the main ENBIS web site and was developed originally by Jeroen de Mast's with help from the developers at Taomi Internet Consultants. The site can be reached either from the ENBIS site at www.enbis.org/pro-enbis, from the sub-domain set up by TIC at pro.enbis.org or via www.pro-enbis.org. Note that www.pro.enbis.org will direct the user back to the ENBIS site.
Members are appointed so there is less administration but the site has rather more content that reflects the activity of the Pro-ENBIS function. There are a number of publicly available sections:
News contains chronological events, latest first, and is presented first to the viewer.
About Pro-ENBIS describes the mission of the thematic network. Contacts has links to the Co-ordinator, Contractors, Members and Steering Committee
Work Packages has links to the Work Packages.
Further sections - and links from the Work Packages - are available only to Members. These are:
- Forum where members can discuss issues
- A list of meetings both forthcoming and historical
- Actions is a list of open and closed actions.
Each Work Package has reports and some of these are available only to Members.
A new section has been added which contains:
- Management Reports etc,
- Pro-forma Templates,
- A standardised Questionnaire and
- Standard Cost statement template and links to the CORDIS site for information.
These are mainly required by Work Package Leaders.
Results
- Workshops,
- Visits,
- Books,
- Other publications and
- Software generated by the various work pages.
All these results are in a simple format.
Members' News
All issues that relate to Members that are not for public dissemination have been moved to a Members' News section. These include announcements of new information and results available and links. This page has been made retrospective so the old Members' information is no longer visible from the front page.
ENBIS
Some of the results for Pro-ENBIS have been linked to the ENBIS site so that ENBIS members can see what is available. This has been done by symbolic links on the server so that any update to the Pro-ENBIS page is automatically reflected on the ENBIS site.
Close collaboration has also been maintained both with the main ENBIS Webmaster (JdM) and the service supplier (Taomi). This has continued to ensure timely response in all these matters.
- Establishing a pool of European expertise:
The pool of European Statistical expertise collected for these deliverable uses the contributors for the state of the art report as the pool.
There are currently discussions ongoing with Wiley publishers looking at the possibility of expanding the state of the art report. This report is a deliverable in its own right. Deliverable 8.7 - Report on state-of-the-art in measurement, business and industrial statistics. The report (deliverable 8.7) has 16 sections each containing between 1 and 2 pages. Most authors are willing to expand their contributions up to full chapters for the proposed book.
- The list of contributors follows.
Shirley Coleman and Dave Stewardson UNEW.ISRU (CO1)
Wim Sendon and Alessandro Di Bucchianico EURAN.ISG (CR2)
Xavier Tort-Martorell and Lluis Marco UPC.SOR (CR3)
Oyvind Langsrud and Frøydis Bjerke NFRI (CR4)
Maria Ramalhoto IST.UMTE (CR5)
Tony Greenfield GREENF.CC (CR6)
Dimilar Vandev BUSTATS (MB7)
Marco S. Reis and Pedro M. Saraiva QUAL (MB8)
Anne De Frenne MATH (MB9)
Jeroen de Mast, Ronald Does UAM.IBIS (CR10)
Daniele Romano UCAG.DME (MB11)
David Rios Insua and Jorge Muruzábal URJC.GECD (MB13)
Bruce Sindahl SIN (MB14)
Andras Zempleni UEOT.DPTS (MB15)
John Tyssedal UTRON.NIT.IM (MB17)
Rainer Goeb UWUERZ.II.CS (MB18)
Rafaello Levi and Grazia Vicario PTRN.DSPEA (MB19)
John Shade GDLT (MB20)
Alberto Luceño and Jaime Puig-Pey UCANT.DMACC (MB21)
Poul Thyregod DTH.IMMOD (MB22)
Marco Busatto, Antonio Pievatolo and Fabrizio Ruggeri, CNR.AMI (MB23)
Bo Bergman and Per Johansson CUT.DTQM (MB24)
Jukka Salmikuukka and Susanna Kunttu VTT.A (MB25)
Alessandra Giovagnoli UBLG.DSS (MB26)
Oystein Evandt IPCONS (CR27)
Irena Ograjen?ek ULJUBL.FE.SI (CR29)
Pasquale Erto and Stefano Barone UNAP.DPA (MB30)
Andrea Ahlemeyer Shibbe AHSDM (MB31)
Ron Kenett KPA (MB32)
Adam Jednorog and kamil Torczewski TUWRO.IPEA (MB33)
Henry Wynn LSERS.STA (CR34)
Chris McCollin UTNOTT (MB35)
Paolo Giudici and Alberto Lombardo UPAV.EPMQ (MB36)
Petros Dellaportas and Stelios Psarakis RCAUEB (MB37)
Soren Bisgaard (Invited expert)
John Logsdon Peo-ENBIS webmaster
This deliverable may also be linked in with other completed databases within deliverable 5.2. Deliverable 5.2 established a list of industrial statistics groups. The database is structured so that lists under various headings can be accessed. The list headings are
- A collection of organisations:
These are specialist consultancies, institutes, companies etc - that concentrate on industrial statistics.
- University departments:
This also includes statistics departments which may not concentrate solely on industrial statistics applications.
- The national statistics bureaux:
These are generally government statistics agencies that may have useful information.
- Statistical societies:
A useful contact point.
This pool of European Statistical expertise can also include the information collected for deliverable 5.4 Established list of European industrial statistical expertise.
The collected data from all these locations constitutes a comprehensive list of European Statistical expertise.
A list of relevant journals was drawn up at month 12 (M5.1) with input from all members of the Network. The list is now available on the Pro-ENBIS web site at http://www.enbis.org/pro-enb
The list of all statistical (and related) journals is a very useful tool for the pro ENBIS members as it allows for quick reference in finding a suitable journal to submit papers to.
It is also a very useful tool to help in research and for finding potential partners for future collaboration.
The list gives a brief overview of each journal and lists the types of statistics mainly dealt with.
The journal list will be useful to any researcher/practitioner seeking to submit papers with a statistical slant. The journal list will also be very useful for those seeking to discover the most up to date applications of current techniques and also new concepts/ideas in related fields.
The list can be found in the section - WP5 – Discovering European resources and expertise. The list contains entries for 243 journals and is in a downloadable format and is available as a secured password protected site.
The journals are listed in alphabetical order and each entry gives an indication of the main subjects covered, the statistical focus of the journal and the publisher. There is also a link to the web pages available for most journals.
Data mining can be defined as the process of selection, exploration and modelling of large databases, in order to discover models and patterns. The increasing availability of data in the current information society has led to the need for valid tools for its modelling and analysis. Data mining and applied statistical methods are the appropriate tools to extract such knowledge from data. Applications occur in many different fields, including statistics, computer science, machine learning, economics, marketing and finance.
This book is the first to describe applied data mining methods in a consistent statistical framework, and then show how they can be applied in practice. All the methods described are either computational, or of a statistical modelling nature. Complex probabilistic models and mathematical tools are not used, so the book is accessible to a wide audience of students and industry professionals. The second half of the book consists of nine case studies, taken from the author's own work in industry, that demonstrate how the methods described can be applied to real problems.
- Provides a solid introduction to applied data mining methods in a consistent statistical framework.
- Includes coverage of classical, multivariate and Bayesian statistical methodology
- Includes many recent developments such as web mining, sequential Bayesian analysis and memory based reasoning.
- Each statistical method described is illustrated with real life applications features a number of detailed case studies based on applied projects within industry.
- Incorporates discussion on software used in data mining, with particular emphasis on SAS
- Supported by a website featuring data sets, software and additional material Includes an extensive bibliography and pointers to further reading within the text
- Author has many years experience teaching introductory and multivariate statistics and data mining, and working on applied projects within industry
- A valuable resource for advanced undergraduate and graduate students of applied statistics, data mining, computer science and economics, as well as for professionals working in industry on projects involving large volumes of data - such as in marketing or financial risk management.
Applying Statistical Methods in Business and Industry the state of the art
Mission Statement:
- To promote the wider understanding and application of contemporary and emerging statistical methods
For Students:
- For people working in business and industry worldwide.
- For the professional development of statistical practitioners (statistical practitioner is any person using statistical methods whether formally trained or not).
- To foster best practice for the benefit of business and industry.
Readership:
- All members of ENBIS.
- Industrial & Business statisticians in general.
- Quality managers.
Contents:
- Introduction to the book, to ENBIS and pro-ENBIS.
- Industrial statistics - History and background.
- Statistical Consultancy.
- Evaluating benefits from process improvement and Quality by Design.
- Management statistics.
- Service quality.
- Risk, Finance and Insurance.
- Applying Data mining methods to business and industry.
- Process Monitoring, Improvement and Control.
- Measurement system analysis.
- Design and analysis of experiments (DoE).
- Safety and reliability engineering.
- Multivariate and multiscale data analysis.
- Simulation.
- Communication.
- Summary and conclusion.
This paper reviews aspects of teaching of Quality to undergraduate Business students to a standard that combines the requirements of the Institute of Quality Assurance (IQA) Body of Knowledge (BOK) and Six Sigma. The paper reviews the aspects of Quality that can be taught University wide. The history of the Undergraduate BA Degree in Business and Quality Management at The Nottingham Trent University is presented showing how change within the University contributed to the change in the course. The change of the course is discussed with reference to the statistical elements of Quality.
Encompassing both aspects of statistical thinking and business skill is of prime importance. Knowledge transfer partnerships sponsored jointly by companies and the DTI attempt to bridge the gap between study and practice by employing graduates and providing academic guidance. Placing students in industry as part of their undergraduate studies is also effective. The reverse situation is currently supported by a large grant from the European Social Fund so that a university-based group can train employees of small to medium enterprises (SMEs) in statistical thinking skills at the university as part of the professional development of the employees.
The Industrial Statistics Research Unit of University of Newcastle-upon-Tyne held an Industrial workshop at Nissan, Tyne and Wear. The idea of this was to bring Manufacturing companies who wished to improve their products and processes through techniques of planned measurement and analysis. For this particular event we titled our workshop "How to implement Six Sigma without becoming Sick Sigma as a parrot".
Six-Sigma itself is a structured methodology that allows companies to dramatically improve their bottom line by designing and monitoring business processes in ways that minimise waste and resources whilst improving customer satisfaction.
It is a continuous improvement process focusing on the following within an organisation:
- Customer requirements;
- Process improvement;
- Cross-structural co-operation;
- Improving their bottom line;
- The benefits of statistical techniques and analysis.
Reliability, Safety & Quality Improvement:
Organisational background ECMI stands for European Consortium for Mathematics in Industry (www.ecmi.dk). ECMI was founded in 1986 by mathematicians from ten European universities. The aims of ECMI are to promote the use of mathematical models in industry and to educate industrial mathematicians to meet the growing demand for such experts. ECMI activities include a two-yearly major congress and an educational programme with modelling weeks, courses etc. The goals of ECMI and ENBIS are quite similar.
The major difference is that ENBIS concentrates on statistical methods, while ECMI has its roots in mathematical analysis (modelling with differential equations) and numerical analysis. In other words, ECMI concentrates on deterministic models, which ENBIS concentrates on stochastic models. The ECMI board felt that a concentration on deterministic models only was artificial and would like to be open a broad view on mathematics, which also encompasses statistics. As a first step ECMI has invited ENBIS and pro-ENBIS to be co-organizers of the ECMI 2004 congress (www.ecmi2004.tue.nl). Two (pro-) ENBIS statisticians (Bisgaard and Ruggieri) were therefore invited as plenary speakers at ECMI 2004.
In summary, the ECMI 2004 congress offers an opportunity for two major European organisations to expand their scope and explore possible ways of cooperation. Such efforts will certainly strengthen European industry.
Audience for congress
The conference is intended for mathematicians, statisticians, scientists and engineers, both from industry and academia. ECMI conferences have a long-standing tradition of bringing together researchers from various disciplines, who work on often only seemingly different disciplines. The transversally of mathematics makes it a versatile tool in a large variety of applications in particular when using it in computational modelling.
"Everything you wanted to know about Six-Sigma but were afraid to ask"
There were around 90 delegates who attended this event. (Please see attached sheet). All from different backgrounds such as Manufacturing, Maintenance, Human Resources, Quality Management, purchasing, Finance and Production and from a variety of companies mostly based in the North east of England. Some delegates came from the Regional Funding Agency.
Dave Stewardson spoke for pro-ENBIS. The presentation was especially put together including material and contributions for this event from Ronald Does (Netherlands), Soren Bisgaard (Denmark), Ron Kennet (Israel), Oystein Evandt (Norway), Bo Bergman (Sweden) and Xavier Tort-Martorelli (Spain).
We outlined the basic six-sigma strategy then showed how this worked in practice. We suggested some improvements to the usual programme and discussed variants that had been used around Europe. The reasons that these project focussed programmes tend to work was explained. ISO quality standards were criticised. The integration of various tolls and methods were demonstrated. A case study was shown that linked the full six-sigma strategy and a re-cap of the main points was shown. Delegates seemed enthused by the presentation.
Dave later mentioned how ISRU consultants can work with clients, via the use of specialised training programmes backed up by on-going business and technical support. Another support tool he talked about was the use of ISRU’s distance learning modules that cover the major statistical tools involved in Six Sigma training programmes. This is a suitable option for companies who currently do not have the time or resources to embark on a major ‘Six-Sigma’ programme but feel that their employees require an awareness and appropriate knowledge of "Six-Sigma" tools and applications. Following this 4 companies expressed a wish to engage in Six-Sigma, Cookson-Fukuda of Shiremoor, Jackel International of Cramlington, Rite-Vent of Washington, Apollo Plastics of Hull and these are all being contacted directly.
The Six Sigma philosophy is currently being introduced in European industries following the successful implementation in companies like Sony, General Electric and Dupont. Now, the Danish biotech company Novozymes is assessing the Six Sigma concept by formulating 12 improvement projects in the production and using the Six Sigma approach in these projects. The Danish Society for Applied Statistics (local ENBIS network estimates that currently 10 to 15 larger Danish companies have implemented the 6 sigma strategy.
We focused our research on the economic optimality for control charts.
This is an important question, as traditional control charts do not take important questions, such as cost of sampling, expected cost of false alarm and erroneous production. If we know at least the approximate probability and distribution of possible shift in the process, then our method can design the cost-optimal x-chart to this situation.
We have investigated a similar - but somewhat simpler question for the CUSUM-chart, which is known to possess good properties in shift-detection.
The optimal chart was found by a Markov-chain approach, which has wide applications in process control.
The paper is about to be published in a refereed journal and so it will be available for the target audience (quality managers, engineers throughout the business sector). As we have shown the strength of our methods by an example taken from a Paper Mill, we believe that the methods will be found especially interesting in this sector.
The main aim of the workshop was to let ENBIS and Pro-ENBIS and the possible use of modern statistical methodology in the industry be known for the local practitioners. To give an overview of the activities of Pro ENBIS as well as the most important papers of the 2nd Conference of ENBIS, held in Rimini in 2002. It also served as a preparation for a 3-day workshop, planned in Budapest for 2004.
During the workshop we focused on the economic criteria for control charts, used extensively in our joint work within Pro-ENBIS. The second part was devoted to the ENBIS as a recent organisation that is quickly expanding and which can promote the use of advanced statistical thinking within the business sector of Europe. Unfortunately there are not many Hungarian members at the moment, so kind of an advertisement was also needed.
Some works presented at the 2nd conference of ENBIS were analysed, out of which the evaluation of statistical consulting - Kennet et-al was in the focus of the discussion.
The participants present will do the dissemination indirectly. As there were teachers of the higher education as well as leading quality engineers of some Hungarian companies, we can hope that the suggested approach will be used at certain areas of the business sector as well as in the teaching at universities.
The experience of the workshop will be used when preparing the main event to be held in Budapest in February 2004: where a 3-day workshop will provide more detailed practical knowledge to an even wider audience.
Nowadays, electronic products tend to be economically outdated before their technical end-of-life has been reached. The ability to analyse and predict the (remaining) technical life of a product would make it possible either to re-use sub-assemblies in the manufacture process of new products, or to design products for which the technical and economical life match. This requires models to predict and monitor performance degradation profiles. In this paper we report on designed experiments to obtain such models. We show how wavelet analysis can be used to extract features from electrical signals. These features are analysed using the Analysis of Variance in order to establish relations between these features and performance degradation.
Six Sigma approach has in the past been predominantly used to improve manufacturing processes. However, Six Sigma is now increasingly applied to a wide variety of non-manufacturing operations also. This is an important development-there are potentially more benefits to be achieved in those areas than in traditional manufacturing, where decades of good work have already paid off.
Pro-ENBIS Workshop for Business and Management "Designed experiments within a problem solving framework with examples in management and the service sector"
Workshop held at Is Molas Hotel & Golf Club Conference Centre, Santa Margherita of Pula – Cagliari, Sardinia, Italy. Organised by the Pro-ENBIS working group on DOE, Warwick University/LSE LSER.STA (CR34).
The workshop was attended by 25 delegates.
Background
Pro-ENBIS was fortunate in being able to attach its Workshop on Design of Experiments to the Conference DEINDE, which took place during the previous week. This meant that people were able to stay on for the Pro-ENBIS Workshop. Thanks should go particularly to Daniele Romano and his team who had the task of being the local organisers of both events. The superb location in Sardinia added to what was a very pleasant occasion.
DEINDE
Although DEINDE was not formally part of the Pro-ENBIS activity a few notes should be made on it. DEINDE is a conference that has been taking place in northern Italy over the last several years. It was set up deliberately to cover the design and analysis of industrial experiments. DEINDE 2003 was particularly innovative in that it covered the application of experimental design to the service sector. There were a number of excellent papers covering the improvements in business process via the use of experimental design techniques and modelling. Because of the newness of the area it could be said that practically every paper represented an innovation and the willingness to in a way experiment with experimental design in the service sector was impressive.
The Pro-ENBIS Workshop
The talks at the workshop were given largely by Pro-ENBIS members with the audience being from academia and industry. The three main talks by Grazia Vicario, Raffaelo Levi, Henry Wynn, Stephano Barone and Alberto Lombardo all covered interesting physical industrial processes cutting, mixing, sticking and so on. The talks showed a high level of sophistication in the use of experimental design and the also the relationship with the underlying science and engineering. Tony Greenfield demonstrated the latest version of his teaching process simulator that was extremely well received.
Conclusions
The double event was, as mentioned, very enjoyable and productive. Participation by industry was quite good but could have been better for the Workshop but the 25 attendees in fact exceeded expectations. The DEINDE Workshop has had a catalytic effect on the use of industrial experimentation in Italy and is something of a role model, which could be used elsewhere in Europe.
Recent research on developing a structural approach to analysis of reliability data has been published. This paper brings together these approaches and introduces some other aspects of reliability analysis, i.e. data manipulation prior to analysis and analysis of multivariate and covariate structures.
Methodology/Approach:
A methodology is proposed which incorporates this structure to aid the determination of the solution to root causes of problems. The methodology uses statistical hypothesis testing principles and well-known quality improvement and reliability techniques incorporated into a Shewhart PDSA cycle. Case studies have been presented in the fields of reliability analysis, steel design and process control.
Findings / Practical implications:
It can be seen from application of the proposed procedure to case study material that each analysis requires the aid of engineering and statistical judgment and knowledge to help identify structure.
Originality/value of paper:
The incorporation of this new approach within reliability and safety case methodology may attempt to aid the solution of ongoing problems and may be used during a project to aid a project manager in developing a closed loop solution to engineering problems.
The aim of the paper is to show how the information, concerning the order in which the pages of a web site are visited, can be profitably used to predict the visit behaviour at the site. Usually every click corresponds to the visualization of a web page. Thus, a web click-stream defines the sequence of the web pages requested by a user. Such a sequence identifies a user session.
Typically, a web mining analysis only concentrates on the part of each user session concerning the access at one specific site. The set of the pages seen in a user session, on a determinate site, is usually referred to with the term server session or, more simply, visit.
Our objective here is to show how web click-stream data can be used to understand the most likely paths of navigation in a web site, with the aim of predicting, possibly on-line, which pages will be seen, having seen a specific path of pages in the past. Such analysis can be very useful to understand, for instance, what is the probability of seeing a page of interest (such as the buying page in an e-commerce site) coming from a specified page. Or what is the probability of entering or (exiting) the web site from any particular page.
The exposure of banks to operational risk is increased in the recent years. The Basel Committee on Banking Supervision (known as Basel II) has established a capital charge to cover operational risk other than credit and market risk. According to the advanced methods defined in "The New Basel Capital Accord" to quantify the capital charge, in this paper we shall present an Advanced Measurement Approach based on a Bayesian network model that estimates an internal measure of risk of the bank. One of the main problems to face to measure the operational risk is the scarcity of loss data.
The methodology proposed solves this critical point because it allows a coherent integration, via Bayes' theorem, of different sources of information, such as internal and external data, and opinion of 'experts'(process owners) about the frequency and the severity of each loss event. Furthermore, the model corrects the losses distribution considering the eventual relations between different nodes of the network that represent the losses of each combination of business line/event type/bank/process and the effectiveness of the correspondent internal and external controls. The operational risk capital charge is quantified by multiplying the VaR per event, a percentile of the losses distribution determined, and an estimate of the number of losses that may occur in a given period. Furthermore, it becomes possible to monitor the effectiveness of the internal and external system controls, in place at the bank.
The methodology we shall present in this document has been experimented, as a pilot project, in one of the most important Italian banking group, Monte dei Paschi di Siena (MPS).
Operational risks in banking have increased in the recent years. These risks spring from globalisation of the financial markets, growth of IT, and the spread of complicated financial products. The new Basel capital accord requires banks to put aside a minimum capital sum to insure against credit risk, market risk and operational risk. Twelve percent of this sum is specifically assigned to cover of operational risks. A press release with more details can be found at www.bis.org/press/p040115.htm The Basel agreement is still being revised and is expected to become compulsory for all banks from 1st January 2007.
The increasing availability of data in the current information society has led to the need for valid tools for its modelling and analysis. Data mining and applied statistical methods are the appropriate tools to extract knowledge from such data. Applied Data Mining: Statistical Methods for Business and Industry provides an accessible introduction to data mining methods in a consistent and application oriented statistical framework. It describes six case studies, taken from real industry projects, highlighting the current applications of data mining methods.
The Links to other industrial statistics groups will be useful to all proENBIS members and through their connections with ENBIS to a wider network of statisticians across Europe. These links will be invaluable for future collaborative work and also for help and advice on a wide range of statistical areas.
A list of links to other industrial statistics groups was drawn up at month 18 (Milestone 5.2) in consultation with the other members of the Network. The list is now available on the Pro-ENBIS web site http://www.enbis.org/pro-enbis/
The list can be found in the section - WP5 - Discovering European resources and expertise.
The results are available to Pro-Enbis members only as an Excel Spreadsheet or as 4 separate pages:
A collection of organisations - These are specialist consultancies, institutes, companies etc - that concentrate on industrial statistics. The database lists the organisations by country which makes searching for expertise in a particular country very easy.
University departments - These statistics departments may not concentrate on industrial statistics applications. The database lists the organisations by country. The list also entries for non-European countries as well such as Australia, Brazil, Canada, China, India, Mexico, New Zealand, South Africa, South Korea, Taiwan, Thailand and United States.
The national statistics bureaux - These are generally European government statistics agencies that may have useful information.
Statistical societies - A useful contact point. Listed by country
The list is available in a download able format and is secured as a password protected site.
Six Sigma is: the scientific method to tackle problems, made operational to work well in business and industry. It gives a methodological framework to tackle quality problems, but it also offers an organisational structure to make possible the required change in a company. The program can be characterised by its emphasis on data, and by its focus on financial results.
Originally developed at Motorola in 1987, the approach created a furore in the USA, with important multinationals such as American Express, Boeing, Citibank, Dow, Ford, and General Electric all obtaining billion dollars' savings from the program. In recent years the approach is implemented in Europe as well at, e.g., DAF Trucks, Nokia, and Philips.
Six Sigma projects are carried out by so-called Black Belts, who are recruited from middle management and are trained in statistical techniques for problem solving. Typically, a Black Belt training consists of four modules of four days each, which is taught in a four months' period. Efficient problem solving methodology, based on a stepwise approach named DMAIC (abbreviation of Define, Measure, Analyse, Improve, and Control) is the main topic of this training.
The workshop gives an overview of the Six Sigma program and philosophy. The background and main concepts are taught, such as the sigma metric and the strategy that it employs to arrive at process improvements. Moreover, the statistical and non-statistical tools which are employed are briefly introduced.
Subjects
The following topics will be treated:
- History of Six Sigma
- Why improvement?
- What is Six Sigma?
- Six Sigma project selection
- DMAIC methodology
- Case studies
- Design for Six Sigma
Target audience
People who get involved with the Six Sigma program, such as managers, executives, engineers and senior staff.
Learning objective:
You will learn what the Six Sigma program comprises. Especially, you will have an overview of the roadmap that it follows, as well as the various tools and techniques that it employs.
Teachers
The teachers are experienced consultants in international industry, but are also well grounded in statistical science. They have implemented Six Sigma at, among others, Achmea Pensions, General Electric Plastics, Getronics, Perlos, Red Cross Hospital and Samsung.
- Soren Bisgaard is Eugene M. Isenberg Professor of Technology Management at the University of Massachusetts-Amherst.
- Ronald J.M.M. Does is professor in industrial statistics at the University of Amsterdam, and general manager of the related Institute for Business and Industrial Statistics IBIS UvA.
- Jeroen de Mast is senior consultant at IBIS UvA.
Clinical Trials at Work
A special hands on workshop with computer simulations designed and lead by Ron Kenett (KPA Ltd. and University of Torino) and David Steinberg (Tel Aviv University and KPA).
Clinical trials are the gold standard for assessing the value of new medical treatments and interventions. The key to a successful clinical trial is careful planning. What is the primary endpoint? What is the comparator for the new treatment? What rule(s) will be used to assign treatments? How many patients are needed? How many centers? What data will be collected? How many follow-up visits are essential? How will blinding be applied? The answers to these questions will dictate how the study data can be analyzed and interpreted. A common thread is that all the questions have statistical components and good statistical thinking can help direct you to effective answers. Our workshop will explore these issues in the context of a mock clinical trial, with actual simulated patient flow and data collection. We will see how good trial design provides the basis for clear conclusions from the trial.
The hands on workshop will allow participants to run METAGEN, a Simulation based Clinical trial, analyze its outcome using basic Statistical techniques, review its design and rerun it with modifications. The objective of the workshop is to provide a basic introduction to statistical aspects of trial design and statistical analysis.
Specifically the METAGEN simulation exercise will discuss:
- Responses to treatments.
- Influence of other factors such as age, gender or life style.
- Number of patients.
- How patients are selected for the trial.
- How patients are allocated to treatments.
- Type of trial: parallel or crossover.
- Compliance of patients to treatment.
- Show how data are recorded, analyzed and interpreted.
For more on the METAGEN simulation package look at:
http://www.greenfieldresearch.co.uk/pgk.html or www.kpa.co.il
Ron S. Kenett, is CEO and Senior Partner of KPA Ltd., an international management consulting firm with expertise in biostatistics and clinical trials design and Professor at the University of Torino, Torino, Italy. His 4 books and 130 papers cover topics such as Industrial Statistics, Experimental Design, Change Point Detection, Multivariate Quality Control, Survey Design, Quality Management and Biostatistics. Professor Kenett¿s Ph.D. is in Mathematics from the Weizmann Institute of Science. Ron is the 2006/2007 President of ENBIS, the European Network for Business and Industrial Statistics and Editor in Chief of the Wiley Encyclopedia of Statistics in Quality ad Reliability and the Journal of Quality Technology and Quantitative Management (QTQM).
David M. Steinberg is Professor of Statistics at Tel Aviv University and a Senior Statistical Consultant at KPA Ltd. He has a Ph. D. in statistics from the University of Wisconsin-Madison. Professor Steinberg¿s primary research area is experimental design. He has served as statistical consultant for a large number of clinical research projects, including both study design and data analysis. David is the editor of Technometrics and Associate Editor of several international publications.
Following the Second Conference on Business and Industrial Statistics (which took place in Rimini on September 23 and 24, 2002), a workshop on Statistical Methods for Evaluating the Effectiveness of Advertising was given at the Department of Statistics of the Universita' di Bologna (the Rimini branch) in co-operation with ENBIS. The workshop was organised in view of the recognition that measurement of consumers' response to advertising should play a leading role in advertising strategy planning by both companies and media centres.
App. 20 workshop participants from different European countries listened to the topics covered by four lecturers (Elisa Montaguti, Sergio Brasini and Giorgio Tassinari of the Universita’ di Bologna as well as Vittorio Bonori of Optimedia). Their lectures were an interesting mixture of theory and practice as shown in the following overview of topics:
- Relationship between advertising and market structures.
- The role of advertising as a relevant component in the marketing mix strategies.
- Advertising, brand equity and brand awareness.
- Advertising and alternative strategies of media planning.
- Reasons for measuring the effectiveness of advertising expenditures.
- Identification of the most suitable statistical data for the measurement process.
- The role of research in advertising.
- Measures and models for evaluating memorial and behavioural effectiveness of advertising.
- Case histories: presentation and discussion.
The workshop was a very useful introduction to the field, since it covered a broad variety of topics. Follow-up workshops and courses will elaborate on specific issues in detail - that the demand exists, was proven by the response of participants.