Category Archives: Eigenvector News

The International Workshop on Multivariate Image Analysis was held September 28-29 in Valencia, Spain. We weren’t able to make it, but we were happy to sponsor the Best Poster prize, which was won by Carl Duchesne. Carl is an Assistant Professor at Université Laval, in Sainte-Foy, Quebec, CA, where he works with the Laboratoire d’observation et d’optimisation des procédés (LOOP).

With co-authors Ryan Gosselin and Denis Rodrigue, Prof. Duchesne presented “Hyperspectral Image Analysis and Applications in Polymer Processing.” The poster describes how a spectral imaging system combined with texture analysis can be used with multivariate models to predict the thermo-mechanical properties of polymers during their manufacture. The system can also be used to detect abnormal processing conditions, what we would call Multivariate Statistical Process Control (MSPC).

For his efforts Carl received a copy of our Solo+MIA software, which is our stand-alone version of PLS_Toolbox + MIA_Toolbox. We trust that Carl and his group at Laval will find it useful in their future MIA endeavors. Congratulations Carl!

BMW

Once again, the Federation of Analytical and Spectroscopy Societies conference, FACSS 2009, has drawn to a successful close. As usual, Eigenvector was there with a booth in the trade show, short courses, and talks.

We were very pleased at the popularity of our Chemometrics without Equations (CWE) and Advanced Chemometrics without Equations (ACWE) courses this year. CWE, led by our Bob Roginski, was at capacity with 20 students. I headed up ACWE, which was nearly full with 17. The courses were hands-on and gave the students a chance to test out the newest versions of our software, the soon to be released PLS_Toolbox 5.5 and MIA_Toolbox 2.0. Thanks to all who attended!

Our Jeremy Shaver gave two talks in the technical sessions. The first, “Soft versus Hard Orthogonalization Filters in Classification Modeling” provided an overview of numerous methods for developing filters. The bottom line is that all of the methods considered, including Orthogonal Signal Correction (OSC), Orthogonal PLS (O-PLS), Modified OSC (M-OSC), External Parameter Orthogonalization (EPO), and Generalized Least Squares weighting (GLS) perform very similarly.

Jeremy also presented “Analyzing and Visualizing Large Raman Images,” which was a joint work with Eunah Lee and Andrew Whitley of HORIBA Jobin Yvon, plus our own R. Scott Koch. This talk presented a method for doing PCA on multivariate images for situations where the entire data set cannot fit into computer memory. “Sequential PCA with updating” was shown to produce a very close approximation to conventional PCA. This is one potential strategy for dealing with the huge imaging data sets that are becoming increasingly common.

We’d like to thank the organizers of FACSS for another great meeting. We’re looking forward to the next one in Raleigh, NC, next year!

BMW

The the 17th International Conference on Secondary Ion Mass Spectrometry, SIMS XVII, is being held this week in Toronto, CA. I am here with our Willem Windig. I’ve had an interest in SIMS for about a dozen years now, after being introduced to it by Anna Belu. Anna invited me to do a workshop on chemometrics in SIMS, gave me some data sets to work with, and off I went. SIMS was especially interesting to me because it provides very rich data sets, and is also often done over surfaces, producing what we refer to as multivariate images.

The use of chemometric techniques in SIMS has expanded considerably since I first got involved. The primary tool is still Principal Components Analysis (PCA), though there are certainly applications of other methods, including Maximal Autocorrelation Factors (MAF). However, I still see lots of examples where authors present 10, 20 or even 30 (!) images at different masses that are obviously highly correlated. These images could be replaced with one or two scores images and corresponding loadings plots: there would be less to look at, the images would be clearer due to the noise reduction aspects of PCA, and more might be learned about the chemistry from the loadings. So we still have a way to go before we reach everybody!

Over the years I’ve learned quite a bit about the non-idealities in SIMS. It certainly isn’t like like NIR spectra that follow Beer’s Law and have very high signal-to-noise! But I’ve gotten an additional appreciation for some of the issues from some of the talks here. There are plenty of “problems” such as detector dead time, low ion count statistics, effects due to topography, etc., which must be considered while doing an analysis.

I’ve been known to say that the secret to EVRI is data preprocessing, i.e. what you do the data before it hits the analysis method, e.g. PCA or PLS. That is certainly true with SIMS, and it’s nice to see that this is well appreciated among this group. The challenge for us as software developers will be to integrate these “correction” methods into our packages in a way that makes using them convenient.

I think that SIMS is ripe for more applications of Multivariate Curve Resolution (MCR), and in fact we focused quite a bit on that in our course here, Chemometrics in SIMS. With proper preprocessing, many applications of curve resolution should be possible. The more chemically meaningful results should be a hit with this audience, I would think.

I hope to see more MCR by the time of the SIMS XVIII, which will be held in Riva del Garda, ITALY, September 18-23, 2011. I plan to keep working in this area, if only so I can go to that! Hope to see you there.

BMW

We got some good news last week when we found out we won an SBIR (Small Business Innovation Research) award to work with NIST (National Institute of Standards and Technology). The project will focus on continued advancement of the Temperature-Programmed Sensing (TPS) system developed at NIST by Steve Semancik and co-workers.

The TPS system is a multi-sensor array, typically with different materials on the individual sensing elements. But there is a twist: the TPS system allows for temperature control of the individual chemical sensor elements. TPS sensors are also known as “micro-hotplate sensors.” The ability to do temperature programming, i.e. vary the sensing element temperature as a function of time in a prescribed way, provides many opportunities for optimization. In addition, proper utilization of the sensor’s time/temperature response may lead to the realization of the “second order advantage” typically found in much larger and more complex analytical systems such as GS-MS.

We’re excited about working on this project and trust that our collective experience (100+ man-years of chemometrics consulting!) will lead to some improvements in the system performance.

I’ve included below the project title and abstract.

Chemometric Support for Temperature-Programmed Sensing System

The Temperature-Programmed Sensing (TPS) system developed at NIST presents many opportunities and unique challenges. The data output from the system can be quite complex and there are many opportunities to optimize the system for specific sensing scenarios. We propose a program aimed at characterizing the system so that potential problems (such as system drift) can be solved early so that the full potential of the system can be realized. The plan includes studies on the stability and theoretical functionality of the sensors. This will result in procedures for instrument standardization and data base-lining. After this is accomplished, advanced preprocessing methods will be considered, along with the use of multi-way (“second order”) data modeling methods for use in calibration and classification. Finally, procedures for optimizing the system for specific applications will be developed.

BMW

We had a number of people that just couldn’t make it to EigenU last month and wanted a Chemometrics course this summer. So we’re planning on doing a course here in Wenatchee, WA, July 13-16, 2009. We’re doing our “Basic Chemometrics” course on Monday-Wednesday, including:

Linear Algebra for Chemometricians
MATLAB for Chemometricians
Chemometrics I: Principal Components Analysis
Chemometrics II: Partial Least Squares and Regression

On the optional fourth day, Thursday, we’ll go over some special topics, including:

Instrument Standardization and Calibration Transfer
Variable Selection
Advanced Preprocessing

The cost of the course will be $1475/$650 (industrial/academic) for the first three days, and $475/$225 for the optional fourth day.

Please come with a laptop with either MATLAB + PLS_Toolbox or Solo installed. (The free demo versions are just fine for this.) Let me know if this is a problem and I’ll try to help you out.

For further information and to register, just contact me.

BMW

Last fall I had the pleasure, with Rasmus Bro, to teach a chemometrics course in Rome. The choice of this location was a result of Rasmus just wanting to go to Rome, and me making an email acquaintance of Prof. Giovanni Visco of the University of Rome, (La Sapienza). In 2008 Giovanni was organizing the second CMA4CH meeting, which is a rather un-obvious acronym for “Application of Multivariate Analysis and Chemometry to Cultural Heritage and Environment.” We gave Giovanni a copy of Solo for the Best Presentation Prize at CMA4CH 2008, and a friendship was born. So when we (Rasmus, along with my wife and daughters, had easily convinced me) wanted to do a course in Rome, I contacted Giovanni and he figured it all out for us. Between Giovanni and his colleague Federico Marini, we were very well taken care of during our stay in Rome!

Giovanni is now in the process of organizing CMA4CH 2010, which will be held on the island of Sicily on September 26-29. He was kind enough to ask me to be the Co-chair for Chemometrics, and I gladly agreed. While somewhat specific, this meeting considers in depth a rather important intersection between a scientific method and an application.

Of course, Italy is THE place for a meeting focused on cultural heritage; they have more of it than just about anybody. And there are so many potential applications of chemometric methods in this arena, (identification of artifacts, provenance of origin, fraud detection, effect of climate and pollution, restoration, etc.), that there should be plenty to discuss! We’re looking forward to it.

BMW

The other day I was updating my bio for a conference and was working on some sentences regarding our experience teaching chemometrics. It included a reference to teaching employees of Fortune 500 companies. So I decided to try to figure out how many of these companies had sent employees to our courses.

Scanning the list, (and just from memory), I came up with 50+ companies we have either taught in-hours courses for or who have sent people to our open courses, including: 3M Company, Abbott Laboratories, Advanced Micro Devices, Agilent Technologies, Air Products and Chemicals, Alcoa, Amgen, Applied Materials, AT&T, Avery Dennison, Becton, Dickinson & Co., Boston Scientific, Boeing, Bristol-Myers Squibb, Chevron, Colgate-Palmolive, ConocoPhillips, Corning, Delphi, Dow Chemical, E. I. du Pont de Nemours & Co., Eastman Chemical, Eastman Kodak, Eli Lilly & Co., ExxonMobil, Ford, General Electric, General Motors, Goodrich, Goodyear Tire & Rubber, Hershey, Hewlett-Packard, Honeywell, Huntsman, Intel, IBM, International Paper, Johnson & Johnson, Kraft, Lockheed Martin, Lucent, Merck & Co., Micron, Owens Corning, Pfizer, Praxair, Procter & Gamble, Rohm & Haas, Schering-Plough, Sunoco, Texas Instruments, Weyerhaeuser and Wyeth.

First off, I’m pleased that personnel at all of these companies have thought enough of us to come to our courses, (such as EigenU). But beyond that, it shows that chemometrics is an important and widely applicable discipline. In these companies chemometric methods play a critical role in many aspects of their product life cycles, from basic research, through product development and scale-up to manufacturing. Multivariate methods improve efficiency and, therefore, are part of these companies competitive advantage.

BMW

The fourth edition of Eigenvector University came to a close last Friday, May 22. The week-long EigenU 2009 had 25 participants from a wide variety of industries and universities. This was somewhat smaller than 2008, but still a good showing given the current state of the economy. Of course, we think it’s the really smart companies that use the slow times to improve the skill sets of their employees!

As usual, the kind folks at the WAC took good care of us. Many thanks to Rick, Amanda, Wayne, Joe, Joshua, Timothy, Bernie, Randall, Eddie, Quentin and all the rest that kept us hydrated and well-fed.

Speaking of well-fed, we all enjoyed Thursday evening’s workshop dinner at Torchy’s. Here everyone contemplates the dinner choices while discussing the day’s courses, which included Chuck and Bob’s “Implementing Chemometrics in PAT,” Rasmus’ “Variable Selection,” and Willem’s “Chemometrics in Mass Spectrometry.”

Thanks to all the course participants for braving the swine flu panic to join us! Also, to the 8 “Eigenvectorian” instructors (Scott, Jeremy, Willem, Neal, Chuck, Bob, Rasmus and myself) for developing and leading the courses.

EigenU 2010 is tentatively scheduled for May 16-21, 2010. In the mean time, catch one of our courses at SIMS XVII, FACSS or EAS, or contact me to schedule in-house training!

BMW

This year’s Eastern Analytical Symposium Award for Achievements in Chemometrics goes to Romà Tauler. Romà is a Research Professor at CSIC, the Institute of Chemical and Environmental Research in Barcelona, Spain.

Romà continues to be a pioneer in Multivariate Curve Resolution, the collection of techniques used for decomposing spectral data into its physically meaningful underlying components. Romà has published an astounding number of papers concerning both the theoretical and practical aspects of MCR, in addition to many other papers in the general field of chemometrics.

Romà is also Editor in Chief of Chemometrics and Intelligent Laboratory Systems.

Professor Tauler joins the previous EAS Chemometrics Award winners (a distinguished group if I may say so myself) listed below:

1996-Steven D. Brown
1997-Tormod Næs
1998-Edmund R. Malinowski
1999-Harald Martens
2000-Svante Wold
2001-Barry M. Wise
2002-Paul Geladi
2003-Paul Gemperline
2004-Rasmus Bro
2005-David Haaland
2006-Age Smilde
2007-Philip Hopke
2008-John F. MacGregor

A special session honoring Romà’s achievements will be presented at this year’s EAS in November.

Eigenvector has been the sponsor of the Chemometrics Award since 2002, and we’re pleased to do it again this year. Congratulations Romà!

BMW

The fourth edition of Eigenvector University will be held in Seattle May 17-22. We’re excited about this edition in part because we have four new courses: Robust Methods, Correlation Spectroscopy, Common Mistakes in Chemometrics (and how not to make them), and Implementing Chemometrics in PAT.

But we’re also excited because EigenU is about the only time during the year where we get the whole Eigenvector staff together. This is a benefit for us–we like to see each other and exchange ideas on consulting projects, talk about software development, etc. But it’s also a benefit for the attendees–a chance to talk to all the Eigenvectorians and find whichever one of us has the most experience on your problem.

While we’ve been somewhat worried that the current economy would affect our attendance, I’m pleased to report that registrations are coming in and we’re up to 17 participants as of April 9. Apparently, there are some companies out there that realize that the best time to sharpen the saw is before all the orders for logs come in.

Early registration for EigenU ends on April 17. After that, prices go up, so get your training plan in order now!

BMW