topic Re: Boundary equation, a la phase boundary in Discussions
https://community.jmp.com/t5/Discussions/Boundary-equation-a-la-phase-boundary/m-p/222634#M44423
<P>There might be a better way......Hopefully another community member will speak up.....but what you might want to try, is to run the regression, and save the predicteds, and then using the Distribution Platform, Fit a Normal 2 Mixture distribution. It will give you the Mean and Sigma of the 2 distributions, and from there you should be able to estimate a division point</P>
<P><span class="lia-inline-image-display-wrapper lia-image-align-inline" image-alt="twodist.PNG" style="width: 416px;"><img src="https://kvoqx44227.i.lithium.com/t5/image/serverpage/image-id/18945i3C500AF176F1CA95/image-size/large?v=1.0&px=999" title="twodist.PNG" alt="twodist.PNG" /></span></P>Thu, 22 Aug 2019 08:55:21 GMTtxnelson2019-08-22T08:55:21ZBoundary equation, a la phase boundary
https://community.jmp.com/t5/Discussions/Boundary-equation-a-la-phase-boundary/m-p/222578#M44413
<P>Hi. </P><P>I have some data where each set appears to have a linear boundary between "pass" and "no pass" results from 2 factor tests. Is there a way to calculate the boundary equation between the two populations of results? </P><P> </P><P>Thank you for advise. Example of a data set is attached. </P><P> </P>Thu, 22 Aug 2019 02:16:37 GMThttps://community.jmp.com/t5/Discussions/Boundary-equation-a-la-phase-boundary/m-p/222578#M44413VAW2019-08-22T02:16:37ZRe: Boundary equation, a la phase boundary
https://community.jmp.com/t5/Discussions/Boundary-equation-a-la-phase-boundary/m-p/222634#M44423
<P>There might be a better way......Hopefully another community member will speak up.....but what you might want to try, is to run the regression, and save the predicteds, and then using the Distribution Platform, Fit a Normal 2 Mixture distribution. It will give you the Mean and Sigma of the 2 distributions, and from there you should be able to estimate a division point</P>
<P><span class="lia-inline-image-display-wrapper lia-image-align-inline" image-alt="twodist.PNG" style="width: 416px;"><img src="https://kvoqx44227.i.lithium.com/t5/image/serverpage/image-id/18945i3C500AF176F1CA95/image-size/large?v=1.0&px=999" title="twodist.PNG" alt="twodist.PNG" /></span></P>Thu, 22 Aug 2019 08:55:21 GMThttps://community.jmp.com/t5/Discussions/Boundary-equation-a-la-phase-boundary/m-p/222634#M44423txnelson2019-08-22T08:55:21ZRe: Boundary equation, a la phase boundary
https://community.jmp.com/t5/Discussions/Boundary-equation-a-la-phase-boundary/m-p/222659#M44429
<P>This example is a case of classification. There are many techniques for this goal. One, in particular, that might satisfy your need for the 'boundary' is the <EM>linear discriminant function</EM>. Here is the result applied to your data:</P>
<P> </P>
<P><span class="lia-inline-image-display-wrapper lia-image-align-inline" image-alt="Screen Shot 2019-08-22 at 6.38.44 AM.png" style="width: 756px;"><img src="https://kvoqx44227.i.lithium.com/t5/image/serverpage/image-id/18946iB1155372039FC751/image-size/large?v=1.0&px=999" title="Screen Shot 2019-08-22 at 6.38.44 AM.png" alt="Screen Shot 2019-08-22 at 6.38.44 AM.png" /></span></P>
<P> </P>
<P>This classification is quite good for the binary response with one 1 in 16 misclassification.</P>
<P> </P>
<P>Select <STRONG>Analyze</STRONG> > <STRONG>Multivariate Methods</STRONG> > <STRONG>Discriminant</STRONG>. Select the <STRONG>predictors</STRONG> and click <STRONG>Y, Covariates</STRONG>. Select the <STRONG>response</STRONG> and click <STRONG>X, Categories</STRONG>. (Yes, this way seems the opposite of the usual meaning of the X and Y analysis roles.) Click <STRONG>OK</STRONG>.</P>
<P> </P>
<P>I suggest that you see <STRONG>Help</STRONG> > <STRONG>Books</STRONG> > <STRONG>Multivariate Methods</STRONG> and the chapter about the Discriminant platform for more information.</P>Thu, 22 Aug 2019 10:48:19 GMThttps://community.jmp.com/t5/Discussions/Boundary-equation-a-la-phase-boundary/m-p/222659#M44429markbailey2019-08-22T10:48:19ZRe: Boundary equation, a la phase boundary
https://community.jmp.com/t5/Discussions/Boundary-equation-a-la-phase-boundary/m-p/222670#M44430
<P>Another popular classification method is <I>binary logistic regression</I>. Here is the result of such an analysis of your data:</P>
<P> </P>
<P><span class="lia-inline-image-display-wrapper lia-image-align-inline" image-alt="Screen Shot 2019-08-22 at 6.54.49 AM.png" style="width: 744px;"><img src="https://kvoqx44227.i.lithium.com/t5/image/serverpage/image-id/18947iA73639467924AF1F/image-size/large?v=1.0&px=999" title="Screen Shot 2019-08-22 at 6.54.49 AM.png" alt="Screen Shot 2019-08-22 at 6.54.49 AM.png" /></span></P>
<P> </P>
<P>The response is the logit( Result ) versus the linear predictor (linear combination), so the interpretation is more familiar perhaps than the discriminant function.</P>
<P> </P>
<P>See the chapter about the <STRONG>Nominal Logistic</STRONG> platform in <STRONG>Help</STRONG> > <STRONG>Books</STRONG> > <STRONG>Fitting Linear Models</STRONG>.</P>
<P> </P>
<P><I>Recursive partitioning</I> also provides classification but because of the strong linear relationships in this case, it would require very many splits. Such a huge tree would be more difficult to interpret.</P>Thu, 22 Aug 2019 10:59:51 GMThttps://community.jmp.com/t5/Discussions/Boundary-equation-a-la-phase-boundary/m-p/222670#M44430markbailey2019-08-22T10:59:51ZRe: Boundary equation, a la phase boundary
https://community.jmp.com/t5/Discussions/Boundary-equation-a-la-phase-boundary/m-p/222698#M44434
<P>Thank you all! Building upon the logistic regression solution, I have noticed that Lin[Fail] function is positive for [Fail] category and negative for [Pass] category. So, it is intuitive to suggest that Lin[Fail] is zero on the boundary between the predicted [Pass] and [Fail] categories. Since the formula for Lin[Fail] is an output from the model, equating Lin[Fail] to zero in this formula gives the boundary equation I am after. </P>Thu, 22 Aug 2019 17:29:58 GMThttps://community.jmp.com/t5/Discussions/Boundary-equation-a-la-phase-boundary/m-p/222698#M44434VAW2019-08-22T17:29:58ZRe: Boundary equation, a la phase boundary
https://community.jmp.com/t5/Discussions/Boundary-equation-a-la-phase-boundary/m-p/222700#M44436
<P>That is correct. The Lin[Fail] stores the linear predictor from the fitted model. It represents the Logit. When the logit is zero, the odds are 1, so the probability of Fail and not Fail are equal.</P>Thu, 22 Aug 2019 17:49:19 GMThttps://community.jmp.com/t5/Discussions/Boundary-equation-a-la-phase-boundary/m-p/222700#M44436markbailey2019-08-22T17:49:19ZRe: Boundary equation, a la phase boundary
https://community.jmp.com/t5/Discussions/Boundary-equation-a-la-phase-boundary/m-p/222711#M44437
<P>There are visualizations for the logistic model to help with the interpretation and exploitation of the model. I used the Big Class data set in the Sample Data folder to fit a model Logit( sex ) F( height, weight). Here are two of the plots:</P>
<P> </P>
<P><span class="lia-inline-image-display-wrapper lia-image-align-inline" image-alt="Capture.PNG" style="width: 386px;"><img src="https://kvoqx44227.i.lithium.com/t5/image/serverpage/image-id/18950i46A0C18842C61152/image-size/large?v=1.0&px=999" title="Capture.PNG" alt="Capture.PNG" /></span></P>
<P> </P>
<P>I set the height and weight values to (nearly) achieve zero output. You can see that the probability of the outcome is essentially 0.5 either way.</P>Thu, 22 Aug 2019 17:55:02 GMThttps://community.jmp.com/t5/Discussions/Boundary-equation-a-la-phase-boundary/m-p/222711#M44437markbailey2019-08-22T17:55:02Z