Financial Statistics/Marketing Seminar

Sponsored by SORA - the Southern Ontario Chapter of the American Statistical Association and the Southern Ontario Regional Association of the Statistical Society of Canada. This seminar is co-sponsored by SAS, the Bank of Nova Scotia, the Royal Bank and Trans Union of Canada and is the thirteenth in a series.

Everyone is welcome to attend. There is no fee for these seminars, but space is limited so pre-registration is essential.

Guest Speaker:

Dr. John MacGregor
Professor of Chemical Engineering, McMaster University

Latent Variable Methods and their use in the Analysis of Large Databases


Tuesday, February 11, 2003


2:00 - 4:00 p.m.


RBC Training Centre
20 King Street West, 10th Floor

King Subway

For those who drive, please don't park on any main streets (King, Yonge, Bay or Adelaide);
cars will be towed away from 3:00 to 6:00 p.m.


Latent Variable Methods and their use in the Analysis of Large Databases

Dr. John MacGregor
Professor of Chemical Engineering, McMaster University


The nature of much of the data collected routinely in industrial, financial and research settings has changed significantly in the past decade with the presence of on-line computer systems and with the collection of large data banks. These data sets are not only large, but typically are non-full rank and non-causal in nature. Traditional statistical analysis and design approaches are very ill-suited to the investigation of these problems.

In this presentation, we look at the latent variable model and latent variable estimation methods such as Principal Component Analysis (PCA) and Partial Least Squares (PLS) as a natural way of treating many of these problems. These methods have become the preferred data-mining approaches in the process industries where large amounts of data are collected in real time by process computers.

The success of these methods in three areas will be discussed and illustrated through actual industrial examples. The first involves the analysis of industrial databases for trouble-shooting process problems, and for establishing multivariate SPC schemes. The second looks at the problem of designing experiments (DOE) in such high dimensional systems for purposes such as drug discovery. The third considers the problem of extracting information from data intensive sensors such as multi-spectral digital images.

Please RSVP

If you plan to attend this seminar, please RSVP to Alison Burnham by e-mail.

Please contact Alison if you want more information about the SORA Financial Statistics/Marketing Seminars.

Alison Burnham
Chair, SORA Committee for Financial Statistics & Marketing
Exchange Solutions Inc.
(416) 646-7103




Co-sponsorship for these seminars is provided by SAS, Scotiabank, RBC Financial Group and Trans Union of Canada.

Return to the SORA home page