Applying logistic model coefficients

[Database Administrator]

Hi guys,

I'd like to apply the logistic model manually, by the coefficients generated
by neural network (hidden node 0) using the sigmoid function. But the
probality calculated with coefficients doesn't match with the probability
calculated by Mining.
Following is the process that I've been using:
1.Run logistic regression
2.Get the coefficients from node_type 23
3.Calculate a = c1x1 + c2x2 + c3x3 + c0 (c = coefficients / x = attributes)
4.Finally apply "a" in the sigmoid fuction to get the probability of each
case in a dataset
5.Check if the probalities matches with probabilities generated by Mining.

So does someone have any idea how to fix this problem?

best

--
Paulo Carvalho
Statistics Master
PhD Markovs Chain
University of S.Paulo

[Database Administrator]

Does this help?
http://www.bogdancrivat.net/dm/archives/36

--
--
--
This posting is provided "AS IS" with no warranties, and confers no rights.
Please do not send email directly to this alias. It is for newsgroup
purposes only.

thanks,
bogdan

"Paulo" wrote in message
news:9B6BD367-FEB0-4A1E-866D-F0F635993FD8@microsoft.com...
> Hi guys,
>
> I'd like to apply the logistic model manually, by the coefficients
> generated
> by neural network (hidden node 0) using the sigmoid function. But the
> probality calculated with coefficients doesn't match with the probability
> calculated by Mining.
> Following is the process that I've been using:
> 1.Run logistic regression
> 2.Get the coefficients from node_type 23
> 3.Calculate a = c1x1 + c2x2 + c3x3 + c0 (c = coefficients / x =
> attributes)
> 4.Finally apply "a" in the sigmoid fuction to get the probability of each
> case in a dataset
> 5.Check if the probalities matches with probabilities generated by Mining.
>
> So does someone have any idea how to fix this problem?
>
> best
>
> --
> Paulo Carvalho
> Statistics Master
> PhD Markovs Chain
> University of S.Paulo