From cast10-request@bevo.che.wisc.edu Mon Nov 15 09:15:34 1999 Received: (from slist@localhost) by bevo.che.wisc.edu (8.9.1/8.9.1) id JAA02239 for cast10-dist@bevo.che.wisc.edu; Mon, 15 Nov 1999 09:15:34 -0600 (CST) Resent-Date: Mon, 15 Nov 1999 09:15:34 -0600 (CST) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-ID: <14384.9109.518974.764414@bahaha.che.wisc.edu> Date: Mon, 15 Nov 1999 09:15:33 -0600 (CST) From: "John MacGregor" Reply-To: "John MacGregor" To: Subject: CAST: Posting for CAST10 X-CAST-Category: education course 2000-05 Resent-Message-ID: <"R3dg_cZB4EM.A.sFE.WOCM4"@bevo.che.wisc.edu> Resent-From: cast10@bevo.che.wisc.edu X-Mailing-List: X-Loop: cast10@bevo.che.wisc.edu Precedence: list Resent-Sender: cast10-request@bevo.che.wisc.edu Sender: cast10-request@bevo.che.wisc.edu ----------------------------------------------------------------- NOTE: Past postings on the CAST10 Email List are archived on the World Wide Web at http://www.che.wisc.edu/cast10 ----------------------------------------------------------------- SHORT COURSE ON PROCESS MODELING BY MULTIVARIATE STATISTICAL METHODS John MacGregor Svante Wold Nouna Kettaneh The Short Course (three days) on Multivariate Statistical Methods for Process Analysis, Monitoring, and Optimization is given again on Dec. 8,9,10, 1999, at the Tradewinds Resort, St. Petersburg Beach, Florida. Below is an outline of the course program, information on the course, and a registration form. COURSE OVERVIEW Process computers routinely collect massive amounts of process data, and yet little use is made of them. Multivariate projection methods (PCA and PLS) are capable of summarizing the information contained in these data as a few "scores" plus a deviation index. This summary then provides -- an easy interpretion of process behavior, -- SPC monitoring charts, -- tools for optimizing process performance. This intensive short course gives an overview of these multivariate methods, and shows their application to the analysis, monitoring, and optimization of processes. The course is illustrated by many industrial examples. A basic working knowledge of the methods is provided through a series of computer workshops with real industrial data. Applicants are also given the opportunity to analyze their own data in these workshops. Applicants should have a basic background in statistics and linear algebra. PROGRAM Day 1 8:00-9:00 Registration 9:00-12:00 Introduction, Principles of projection 2:30-5:30 pm Principal Components Analysis (PCA) 7:30-9:30 pm PCA workshop with Industrial data (Two students per PC) ----------------------------------------------------------------- Day 2 8:30- 12:00 PLS Modelling (Partial Least Squares Projection to Latent Structures). 2:30- 5:30 pm PLS continued 7:30-9:30 pm PLS Labs with industrial data ----------------------------------------------------------------- Day 3 8:30-10:00 Process and Quality improvement 10:30-12:00 Multivariate Statistical Process Control (MSPC) 1:00-2:30 pm MSPC continued 2:30-4.30 pm Multivariate workshop with industrial data Labs and Software Industrial data sets will be analyzed in the course using an interactive PC-based Windows software. Participants will work in pairs on PC's. Participants are encouraged to bring their own multivariate data for the workshops. The data should in be in tabular form on 3-1/2 diskettes in one of the following formats: ASCII flat file, spreadsheet DIF or XLS file or Matlab (.MAT). Lecturers John MacGregor is a professor in the chemical engineering department at McMaster University. He is a consultant to many companies in North America in the areas of advanced engineering process control, statistical process control, and polymer reaction engineering. He has taught many short courses to industrial audiences, and has published extensively on the topics in this course. Nouna Kettaneh-Wold is president of Umetrics Inc, Kinnelon, NJ, and part owner of Umetri AB, Umea, Sweden, both companies that consult with industry on Experimental design and multivariate modelling, and develop software for these tasks. Svante Wold is Professor of Chemometrics at Umea University, Sweden. He holds a Ph.D in Organic Chemistry from Umea University. Together with Herman Wold, he is one of the originators of PLS as it is used today in technical applications. He consults extensively for the chemical and pharmaceutical industries. GENERAL INFORMATION Registration Cost per person is $1450. This includes a comprehensive set of lecture notes. The Number of places is limited to 30. Acceptance is on a first come, first served, basis. We recommend that you register as early as possible since the course is usually oversubscibed. For further information call Umetrics at (973)-492-8355, or FAX: (973)492-8355 or e-mail to: svante.wold@umetrics.com or macgreg@mcmaster.ca _________________________________________________________________ REGISTRATION FORM Multivariate Process Modelling St. Petersburg Beach, Florida Dec. 8,9,10, 1999 Please register the following persons: Names: Company: Address: Tel: Fax: _________________________________________________________________ To register, complete this form and fax or e-mail it back to Umetrics. Cost per person is $1450 US. Your payment can be made by check payable to "Umetrics", with a PO number and billing address or by major credit card. Mailing Address: Umetrics, 17 Kiel Ave. Kinnelon, NJ 07405 Fax: (973)-492-8359 Phone: (973)-492-8355 E-mail: svante.wold@umetrics.com