Newbie Help JASP Machine learning

cramsden

Hi,

I am very new to this world and have been working on making predictive models using linear regression. I've gone through excel and have been using JASP more recently to make these linear regression models.

I just came across the machine learning section in JASP and was excited to use that as opposed to my manual attempts at choosing predictors going through training sets, validation sets and test sets etc.

However I am very confused as when I attempt to use the machine learning in JASP I am not seeing what the coefficients would be for each of the predictors in the model or what the actual model looks like. I am under the assumption that using machine learning would create the optimal regression model, performing all the training and validation testing on its own. Is this correct or am I completely misunderstanding it?

Thank you

EJ

I'll forward this to the team, but the purpose of ML is usually not to estimate parameters; the focus is squarely on visualization and prediction.

Cheers,

E.J.

vandenman

Which analysis are you using?

However I am very confused as when I attempt to use the machine learning in JASP I am not seeing what the coefficients would be for each of the predictors in the model or what the actual model looks like.

Well some analyses, say Random forest, may have 1000+ parameters so we don't show those. For others, such as regularized linear regression, there is a checkbox in the input options that will show the regression coefficients.

I am under the assumption that using machine learning would create the optimal regression model, performing all the training and validation testing on its own.

That is correct, for all analyses, the data are by default split into a training, validation, and test set. If you want you can change this in the section "Data Split Preferences".

Let me know if you have any more questions/ if anything is unclear!

cramsden

Thank you both for the help.

Yes this is exactly what I am looking for in the regularized linear regression section. Is this available in the other machine learning options?

Basically I am looking for the coefficients so I can see the model in a y = mx1 + mx2 + mx3 +b form that way I can use the model the regression gave me going forward to make predictions. Again I am totally new to this and just self teaching at the moment so if I am heading in the wrong direction with expecting to see that please let me know!