Re: DM: ML in environment

[Debashish_Chakravarty]

Elias:

The kind of study that you are interested in involves the consideration of the saptio-temporal distribution of the data, i mean that the sampling process starting from the collection of the data to the prediction of the pollutants concentration should all be carried out keeping the particular distribution in mind. I hope, you would be looking for this while studying, and you can try out connecting to the site: http://curie.ei.jrc.it/software/index.htm and have got to download the GSLIB or GEOEAS or other packages and have a look ...

You can get the relevant documentations :
http://curie.ei.jrc.it/ai-geostats.htm
I hope this helps ...


Regards:
Debashish.




ekalap@ee.upatras.gr on 04/19/99 12:55:07 PM

To:   datamine-l@nautilus-sys.com
cc:    (bcc: Debashish Chakravarty/PWA)
Subject:  DM: ML in environment



Hello,

I am a post-grad. student who tries to write an overview about the applications of Machine Learning in air-quality and air-pollution control problems.

I would be grateful if anyone of you could provide me with useful links to similar work done in the past. I know so far of efforts made with Neural Networks, and I wander if someone else has ever tried to do the same using Decision Trees, Bayesian modeling, evolutionary algorithms or some other algorithm.

[Usually what one wants to predict is the levels of some dangerous pollutant for the next 8 or 12 or 24 or even 48 hours]

------------------------------
===== Elias Kalapanidas ======
==   University of Patras   ==
=  Electric Engineering Dept =

ekalap@ee.upatras.gr