| Author | Message | ||
     Lisa Patel |
I am new to this area. I am just wondering if anyone could me some suggestions on how to get familiar with chemometrics? Maybe suggest a list of books or articles to read. Thanks. | ||
| W. Fred McClure (Mcclure) |
Lisa, If you send me an email address I can give you my suggestions. Since these may be construed to be commercial in content, I do not think it proper to reply here. Contact me: [email protected] | ||
| Bruce H. Campbell (Campclan) |
Lisa, If you purchase only one book, it should be the second edition of The Handbook of Near Infrared Spectroscopy. Bruce | ||
| hlmark |
Lisa - there are also a number of short courses given, both in NIR, and in Chemometrics. I'm pretty sure there will be at least one of each at the EAS, which is coming up in a little over a month, so this is an ideal time to look into that. Howard \o/ /_\ | ||
| Don Burns |
Hello Lisa, At the risk of being accused of self-promotion, you might consider the Near-IR Workshop scheduled for Oct 14th at the FACSS meeting in Providence RI. | ||
| Michel Coene (Michel) |
Hi, welcome to the wonderfull world of Chemometrics. Why don't you tell us a little more? What will you be using chemometrics for? After having been in it for several years, I am still not certain if it is a science or an artform. The courses and books mentioned above are certainly a good start, but there is only one place where you will really learn it... Enjoy! (and remember: the impossible is often the untried.) | ||
| Lisa Patel |
Thank you all for quick reponses. I am interested more in classification methods, pattern recognition, and QSAR areas. | ||
| Lisa Patel |
Thank you all for quick reponses. I am interested more in classification methods, pattern recognition, and QSAR areas. | ||
| Tony Davies (Td) |
Dear Lisa, I have been holding back expecting many people to suggest a certain book but nothing has happened! So may I suggest: http://www.impublications.com/discus/userfriendly/index.html You might also be interested in a column in Spectroscopy Europe at: http://www.spectroscopyeurope.com/td_col.html Welcome to Chemometrics! Tony Davies | ||
| hlmark |
Lisa - both of Tony's recommendations are good, but perhaps not for the complete beginner. The columns are probably a bit too speciallized for you at this point, although in time you will come to appreciate them. The best explanation I know of for novices, covering many of the general concepts used in chemometrics is in "Chemometric Techniques for Quantitative Analysis" by Richard Kramer, Marcel Dekker (1998). On the other hand, as the title indicates, it is oriented toward quantiative analysis rather then qualitative. I don't know of any book with that clarity of explanation, that is oriented toward qualitative analysis. However, the one by K. Beebe, R. Pell and M. Seasholtz, "Chemometrics: A Practical Guide", Wiley, (1998) has a large section dealing with algorithms for qualitative analysis (identification, or as they call it "pattern recognition") and is reasonably well-written. Howard \o/ /_\ | ||
| shay ben-menachem |
Dear All, I am a UK student who completed his studies in Israel and is currently entering working at TEVA paharmaceutical compny, Israel, in development of NIR based quantitative analysis of drug active materials in granulates, blends and tablets. Looking back at past works (done others)I wonder about the reliability of sample selection method done with libraries (on which the system is based)(we do qualitative analysis first) with respect to outliers. Shouls those outliers be included in the library or ommited. Sould those be included in the system model development? Note that these were successfuly quantitatively identified by HPLC . So that these ouliers are spectrally rejected while chemicaly should be accepted. Any comment should be highly appreciated. Thanks for the time you took for looking into my work. Best regards, Shay I am willing to call anyone who could assist. | ||
| shay ben-menachem |
Dear All, I am a UK student who completed his studies in Israel and is currently entering working at TEVA paharmaceutical compny, Israel, in development of NIR based quantitative analysis of drug active materials in granulates, blends and tablets. Looking back at past works (done others)I wonder about the reliability of sample selection method done with libraries (on which the system is based)(we do qualitative analysis first) with respect to outliers. Shouls those outliers be included in the library or ommited. Sould those be included in the system model development? Note that these were successfuly quantitatively identified by HPLC . So that these ouliers are spectrally rejected while chemicaly should be accepted. Any comment should be highly appreciated. Thanks for the time you took for looking into my work. Best regards, Shay I am willing to call anyone who could assist. | ||
| Gabi Levin |
Shay, May I suggest the following: 1. Double check to verify that your reference method did not produce an error that leads you to think that you have an outlier. I have seen cases that a duplicate and triplicate served a great service to avert erroneous conclusions about the spectrum. 2. Example of reference method problem: In a stream of five constituents, four modeled very well, one yielded mediocre results. Duplicate and triplictaes showed the GC was correct. Since the same spectrum yielded four good regressions, I have requested that they check their GC procedure and verify that the eluant was a single compound. It turned out that it was a mixture of very close chemicals, (that is why the co-eluted and why they did not know that for a good number of years)and since the ratio between them was not a constant, the spectrum will be affected in a way that could not be correlated to the reference values, thus the poor regression. Questioning the reference method is a must, and it can pay very well in many cases. Even "proven" methods nust be questioned. 3. In parallel to questioning the reference method and if the analyzed sample truly represents the material from which spectrum was collected, dig for possible spectral errors. I can not speculate as to where they can come from. For this and other matters and if you are in Israel and wish to contact me, call me on my mobile between October 20th, and November 3rd, 051 733 911. If you are not in Israel, you can send me an e-mail [email protected] and I will call you back. Many thanks, Gabi Levin |