Forums before death by AOL, social media and spammers... "We can't have nice things"
|    comp.ai.fuzzy    |    Fuzzy logic... all warm and fuzzy-like    |    1,275 messages    |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
|    Message 614 of 1,275    |
|    ghirardelli |
|    Re: Finding classifiers & extracting fuz    |
|    07 Mar 06 22:46:07    |
      XPost: comp.ai.neural-nets, comp.ai.genetic, comp.soft-sys.math.scilab              csoren a écrit :       > Hello,       >       > I've undertaken a complex data-mining task and I need advice, please!       > We are trying to develop a model to measure and predict the complexity       > of a traffic situation within a region of airspace. We're very lucky       > to have access to a database from a previous research effort that       > includes 40 relevant variables, all continuous, and 1 discrete response       > variable: a subject matter expert complexity rating. The previous       > researchers tried linear regression and a limited neural net       > application to find the best model, but they had only moderate success,       > and in a particular situation which is recognized as 'highly complex'       > their model was not unsuccessful. So we want to extend the approach by       > identifying a set of classification indicators, firstly, and then       > pursuing nonlinear models and fuzzy logic. (It is indeed a very fuzzy       > problem.)       >       > And of course, we do not have Matlab and do not have permission to buy       > it nor other expensive commercial COTS. (I didn't expect to get access       > to this database, but I am convinced it can provide a great baseline of       > rules for us, if we can squeeze them out.... so this is unfortunately a       > bit of a 'spare-time' research effort.) I'll be happy to buy something       > cheap or invest a bit of time extending an algorithm (C++), if someone       > has a tool that fits 'almost' ...       >       > So, .... any recommendations for approach, tools, open source       > solution(s)? At a first glance the ANN and fuzzy addons to SciLab do       > not include classification nor rule extraction, but I will look into       > them further. I've heard about CANFIS but cannot find it (other than 1       > expensive commercial implementation). There was a NEFCLASS application       > also but it is quite old, crashes on my PC and I don't find any source       > code.       >       > Advice & tips much appreciated, thanks!       > Carolyn       >       Have a look at www.gepsoft.com       I have used APS3 extensively, do have a try              --- SoupGate-Win32 v1.05        * Origin: you cannot sedate... all the things you hate (1:229/2)    |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
(c) 1994, bbs@darkrealms.ca