topic Set an interval for a prediction in Discussions
https://community.jmp.com/t5/Discussions/Set-an-interval-for-a-prediction/m-p/208391#M41878
<P>I want to have a prediction formula for a variable, so with Fit model I do my regression and I select <EM>save the prediction formula</EM> to have my prediction. But I have an issue because the variable I want predict is a probability so has to be between 0 and 1 (included) and the prediction formula produces results greater than 1. I would like to know if there exist way(s) to set an interval for a prediction ? (briefly, if we can add constraints on the prediction).</P><P>Two more information which may be useful :</P><P>To perform my regression I do a <EM>Generalized Regression </EM>with a <EM>Normal </EM>distribution.</P><P>I am an user of JMP Pro 14.</P><P> </P><P>Thanks</P>Wed, 15 May 2019 07:09:21 GMTnico102019-05-15T07:09:21ZSet an interval for a prediction
https://community.jmp.com/t5/Discussions/Set-an-interval-for-a-prediction/m-p/208391#M41878
<P>I want to have a prediction formula for a variable, so with Fit model I do my regression and I select <EM>save the prediction formula</EM> to have my prediction. But I have an issue because the variable I want predict is a probability so has to be between 0 and 1 (included) and the prediction formula produces results greater than 1. I would like to know if there exist way(s) to set an interval for a prediction ? (briefly, if we can add constraints on the prediction).</P><P>Two more information which may be useful :</P><P>To perform my regression I do a <EM>Generalized Regression </EM>with a <EM>Normal </EM>distribution.</P><P>I am an user of JMP Pro 14.</P><P> </P><P>Thanks</P>Wed, 15 May 2019 07:09:21 GMThttps://community.jmp.com/t5/Discussions/Set-an-interval-for-a-prediction/m-p/208391#M41878nico102019-05-15T07:09:21ZRe: Set an interval for a prediction
https://community.jmp.com/t5/Discussions/Set-an-interval-for-a-prediction/m-p/208421#M41884
<P>So, your response is a probability? That is, you are not modeling a quantitative response with a logistic model?</P>
<P> </P>
<P>If so, you might try a logit transform of the response.</P>Wed, 15 May 2019 11:39:17 GMThttps://community.jmp.com/t5/Discussions/Set-an-interval-for-a-prediction/m-p/208421#M41884markbailey2019-05-15T11:39:17Z