Search All of the Math Forum:

Views expressed in these public forums are not endorsed by NCTM or The Math Forum.

Notice: We are no longer accepting new posts, but the forums will continue to be readable.

Topic: fuzzy logic value
Replies: 3   Last Post: Oct 2, 2009 11:11 AM

 Messages: [ Previous | Next ]
 cpp.matlab@gmail.com Posts: 47 Registered: 10/24/08
Re: fuzzy logic value
Posted: Oct 2, 2009 11:11 AM

On Sep 30, 9:47 pm, noor hazlini borhan <nhazlinimohd...@gmail.com>
wrote:
> On Sep 30, 8:28 pm, "Sprinceana " <mihai...@yahoo.com> wrote:
>

> > Hi!
>
> > You gave little detalis but I hope this link can help you:
>
> >http://www.mathworks.com/access/helpdesk/help/toolbox/fuzzy/index.htm...
>
> hi,
>
> thanks for sharing with me.Here are the little details regarding my
> project.Firstly, i got the responses from the coders.For the first
> set, i took 2 coders.Then, they will give their responses on the
> students' discussion sentences.Actually , my project is coding data
> reliability usingfuzzylogicfor students' discussion forum.But , for
> the normal calculation of reliability withoutfuzzylogicvalue, if
> one of the coder disagree with that sentence, the reliability value
> will be low.So to overcome this problem, i want to usefuzzylogic
> approach.So, usingfuzzylogic, if one of the coder did not agree with
> the sentence, it's still got value right?So, the problem is, how to
> calculate thefuzzyvalue for this.Please help me on this.If u have
> any suggestion please tell me asap...thank you very much

From what i understand...

A training data set must be formed with 3 columns
Coder is the number assigned to the coder,
Value is the value he gave to the discussion
and desired output is the value it should have.

Coder | Value | Desired output
001 1.5 3
002 0.5 4
:
:

Open anfis edit...throw training data.
generate FIS

similar to training data you should have some testing data without
desired ouput
throw that...and model will predict desired output

Date Subject Author
9/30/09 noor
9/30/09 Sprinceana
9/30/09 noor hazlini borhan
10/2/09 cpp.matlab@gmail.com