How do I create Non-Linear Fit Modeling for multiple unknown dependent variables? If I use the fit model button on the task bar, it automatically assumes I want a linear fit for my PCA and creates an expression around that. But what if I want a curved model or different ordered fit. There are ways of fitting individual parameters of a model to a log or power, but not the whole model. Attempted to draw a crude model fit in purple to show the type of model I’d like to be able to produce from Jump on curved data. I’d like it to spit out the non-linear equation as well. So instead of my predicted expression being Y = Intercept +a1X1 + a2X2... etc I need to know if there is a better fit on a Y = Intercept + a1X1^b1 + a2X2^b2 ... etc. where b2 is not necessarilry an integer. Any ideas on how to create a non-linear predicted equation for principle components analysis in Jmp?
Thanks
Jared
Re: How do I create Non-Linear Fit Modeling for multiple unknown dependent variables?This does get me closer, unfortunately, it assumes a single coefficient on every parameter in the model. I could have param A be best fit by 0.6 exponent , and param B best fit by 1.4 exponent. By using bivariate fit of Actual by predicted I am forced to assume all coefficients are equal. I also don't know if other parameters that should have been included in my original model as a curved parameters are left out, now that I have fit the model linearly and curved those pieces.
Re: How do I create Non-Linear Fit Modeling for multiple unknown dependent variables?You can't access a model fit using this tool, but you can save the predicted data to a new column and then do a bivariate plot between your predicted and actual data.
Re: How do I create Non-Linear Fit Modeling for multiple unknown dependent variables?Unfortunately your response is from the Fit Y by X button, I am looking into the Fit Model Function. Under Analyze, then Fit Model. There is no polynomial fit in the drop down on the responseRe: How do I create Non-Linear Fit Modeling for multiple unknown dependent variables?Hi Jared,
You can do a number of additional data fits from the red triangle dropdown. For a polynomial fit specifically choos 'Fit Polynomial' and choose the type of fit you want for your data. The equation for the fit should appear below the plot. You can also choose spline fits from the 'Flexible' section.
