Author |
Message |
Howard Mark (hlmark)
Senior Member Username: hlmark
Post Number: 190 Registered: 9-2001
| Posted on Tuesday, May 13, 2008 - 9:02 am: | |
Hmmmm .... this is a tricky one - the e-mail message below is a commercial message, but it contains several references that are probably useful for this thread. I compromised by deleting all the information not pertinent to this discussion. I hope Scott (and maybe some others) can find it useful \o/ /_\ ----- Original Message ----- From: Springer To: [email protected] Sent: Tuesday, May 13, 2008 7:03 AM Subject: Updates in Statistics and Special Offers SpringerAlerts on Statistics, May 2008 Visit us at springer.com Book News Data Manipulation with R By P. Spector This book presents a wide array of methods applicable for reading data into R, and efficiently manipulating that data. In addition to the built-in functions, a number of readily available packages from CRAN (the Comprehensive R Archive Network) are also covered. Book Series: Use R! This series of inexpensive books on R publishes shorter books aimed at practitioners. Books discuss the use of R in a particular subject area or as it relates to statistical topics. In most cases, books will combine LaTeX and R so that the code for figures and tables can be put on a website. Titles in this series Forthcoming Use R! Books Morphometrics with R Wavelet Methods in Statistics with R Thank you for your continued interest in Springer's publications. Kind Regards, Margit Dann Product Manager Springer In this Newsletter: Statistics Book Club Free Journal of the Month Book News Series Use R! Forthcoming Books This email has been sent to [email protected]. You are receiving this email because as a registered user with SpringerAlerts you have opted to receive information from Springer. If you want to review or update your profile, please go to springer.com/alertprofile If you would not like to receive further information on Springer's publishing program in your interest field, please use this link: http://springer.p.delivery.net/m/u/klw/k.asp?e=hlmark%40prodigy.net&mid=1263899771&cid=12793351518 For all enquiries, problems or suggestions regarding this service, please contact [email protected]. Springer respects your privacy and does not disclose, sell or rent your personal information to any nonaffiliated third parties without your consent. Please visit our Springer Privacy Statement. Springer-Verlag GmbH, Tiergartenstrasse 17, 69121 Heidelberg, Germany, phone: +49 6221 487 0, fax: +49 6221 487 8366 � Springer 2008, springer.com ECS 1781 |
Bernard North (bnorth)
New member Username: bnorth
Post Number: 3 Registered: 5-2007
| Posted on Friday, May 09, 2008 - 8:27 am: | |
I did try a bit to use PLS for NIR. I did have problems as the PLS package doesn't include a calculation of the Q2 measure (available in SIMCA). Another useful feature would be allowing classificiation of a test sample based on the model from a training sample |
Christian Mora (cmora)
Advanced Member Username: cmora
Post Number: 23 Registered: 2-2007
| Posted on Friday, May 09, 2008 - 8:14 am: | |
for example: pre-processing algorithms... |
Scott Parsons (aussiecologist)
New member Username: aussiecologist
Post Number: 3 Registered: 4-2007
| Posted on Friday, May 09, 2008 - 1:55 am: | |
Thank you for your replies! I will update here, in time, my experiences with R for NIR applications. Christian, can you give examples of what applications you have had to write code for that are not included in the curtrently available packages?
|
Christian Mora (cmora)
Advanced Member Username: cmora
Post Number: 22 Registered: 2-2007
| Posted on Tuesday, May 06, 2008 - 10:04 am: | |
I guess the best way to start learning R is by going to ttp://www.r-project.org. There you'll find the R source, manual, books that have been published, etc. R can be linked to C, C++, Fortran, Matlab among others. |
venkatarman (venkynir)
Senior Member Username: venkynir
Post Number: 65 Registered: 3-2004
| Posted on Tuesday, May 06, 2008 - 12:21 am: | |
Dear Christian Mora; Thanks .But in the publication , I have not seen much about "R". I have seen PLS ,Unscrambler and Infromatrix are refered frequently. Seven Years in quite long period Is it.? Have you implemeneted it in on-line measurment ? Can we link with C++ as front end . Please answer for the query. Since you have good expeeince in "R" can you suggest the web address(right resource ) and good book to read and work . |
Christian Mora (cmora)
Advanced Member Username: cmora
Post Number: 21 Registered: 2-2007
| Posted on Monday, May 05, 2008 - 8:23 am: | |
I use R everyday for all kind of analysis, especially for NIR calibration models. The main source for fitting PLS models is the package pls. There are other options such as the packages plsgenomics and chemometrics. PCA can be done by SVD although the chemometrics package and the pcaMethods package from a related project (Bioconductor) can perform other analysis like NIPALS, bayesian, etc. Of course the main advantage of using R is the no cost associate to it and because it is open source you can download the packages and make any changes in the code as you like (you need to understand them first ). However, to take full advantage of R, I believe, you must have some intermediate to advance knowledge in R programming which can take some time (I'm still trying to learn how to master it after 7 years of using R). In summary, you can get the same results, and sometimes more efficiently, by using R compared to other traditional chemometrics softwares but requires that you write you own code for several analysis not implemented in packages yet. Regards CM |
venkatarman (venkynir)
Senior Member Username: venkynir
Post Number: 64 Registered: 3-2004
| Posted on Monday, May 05, 2008 - 3:08 am: | |
Dear Thomas; It is beta version only. Kindly refer "ParLeS :software for chemometric analysis of spctroscopic data " published "chemometrics &Intellegent Laborratory Systems 2007 page 7-12 Technical note by Raphel A. |
Dintis Thomas (dintis)
New member Username: dintis
Post Number: 2 Registered: 11-2007
| Posted on Monday, May 05, 2008 - 2:30 am: | |
Here is the link to the page were Principal Component Regression (PCR) and Partial least squares (PLS) modeling are discussed for Scilab. http://www.scilab.org/contrib/index_contrib.php?page=displayContribution&fileID=231 |
Dintis Thomas (dintis)
New member Username: dintis
Post Number: 1 Registered: 11-2007
| Posted on Monday, May 05, 2008 - 2:18 am: | |
Dear Venky Sir, Here is another open source program which can be used for http://www.scilab.org/ the best part is the data can be taken to Labview, Matlab and it is open source license. |
Scott Parsons (aussiecologist)
New member Username: aussiecologist
Post Number: 2 Registered: 4-2007
| Posted on Monday, May 05, 2008 - 12:05 am: | |
It is discussed with some short examples in the August 2006 R newsletter (Vol 6/3), which can be downloaded from: http://cran.r-project.org/doc/Rnews/ The main "R" package is "pls", but I have also noted a genetics package ("genalg") that mentions NIR analysis as an extended application
|
venkatarman (venkynir)
Senior Member Username: venkynir
Post Number: 63 Registered: 3-2004
| Posted on Sunday, May 04, 2008 - 11:23 pm: | |
Dear Scott; Can you give the web page and more about "R" for NIR analysis. I have worked on PLS Tool box and Unscrambler. I eager to know more about "R" . I will try learn "R" in short time. |
Scott Parsons (aussiecologist)
New member Username: aussiecologist
Post Number: 1 Registered: 4-2007
| Posted on Friday, May 02, 2008 - 9:46 pm: | |
Hello, I am curious if anyone uses "R" for NIR spectral analysis? As this software is free, it holds many benefits to students over expensive packages such as The Unscrambler, however there are obvious drawbacks due to the complexity of the "R" language. Has anyone had success in using this software and the available relevant packages, such as the many pca and pls functions, and could provide some feedback of their success and the pros and cons they have found in R's use in any aspect of NIRS? Any feedback is greatly appreciated!
|