Electronic Tongue ElectronicTongue.com
December, 2006
Second Announcement of ISOEN 2007 is available now at Official Page.
For more information please visit conference's Official Web Site.

December, 5th, 2006
A list of our partners is expanded.

December, 5th, 2006
New literature references added.

June, 26th, 2006
You can have a look at our research projects and relevant references.
Follow the link.

April, 4th, 2006
We developed a new ion-selective sensor, selective to nitrite ions (NO2-). All characteristics of a new sensor you can find in the table.

March, 31th, 2006
The date of ISOEN 2007 is defined. The conference will be held from 3 to 5 of May, 2007. For more info please, visit the Official Site.

March, 30th, 2006
Await soon (in 2-5 days) some new photos of working systems.

March, 9th, 2006
Our site is available in Pocket PC mode now, still with no pictures, but with full-text content. You can find it at http://pda.electronictongue.com

December, 1st, 2005
We will organize ISOEN Conference in May, 2007. Please, visit our Official Page!
 


May, 17th, 2005
Two NEW sensors were added! Their characteristics you can see at the bottom of the table with PVC-based sensors. All kinds of sensors are negotiable!

June, 30th, 2005
Now you can ask your questions to us in our forum! We'll respond you as soon as possible!

July, 29th, 2005
Our Ion-selective electrodes have got a Pattern Approval Certificate of Measuring Instruments which is registered in the State Register of Measuring Instruments of Russian Federation (Certificate №13763-05). A copy of this certificate will appear here soon!

December, 15th, 2005
Four of our new papers added to References page.



Disclaimer     Terms and Conditions    
All rights reserved, owned by Andrey V. Legin, Laboratory of chemical sensors, St.-Petersburg State University, 1995-2005
Pattern recognition and multivariate calibration methods

To deal with the data from a sensor array it is necessary to use different methods, which allow fitting the multidimensional output of the sensor set. Depending on the task, some of the methods shown in the table can be used. In some cases several methods are equally applicable, in the other cases the certain method gives the best results.

DATA PROCESSING TECHNIQUES
Pattern recognition methods

Principal component analysis
Cluster analysis
Fuzzy logic
Artificial neural networks (back-propagation, SOM)
Multivariate calibration methods

Multi linear regression
Principal component regression
Partial least square regression
Non-linear regression
Artificial neural networks (back-propagation)

An important principle we strictly follow - adequate data processing is the essential part of the electronic tongue approach but even advanced math methods cannot improve the results. The data produced by the sensor array must be reproducible and reliable. This is ensured by the responsible design of the sensor array and thoroughly elaborated measuring procedure.










Search