Classification Learner App :: Logistic Regression - Decision Making

For Logistic Regression using the Classification Learner App, the classifier models the class probabilities as a function of the linear combination of predictors, using the 'fitglm' function (as specified in the documentation).

The predicted response of this model to a new data set is the predicted probabilities for each class. I would like to know how does the classification learner app classifies this predicted data based on probabilities (the value for the probability at which Yes/No decisions are made).

Answer 1

[Shashank Gupta]

Hi Tarunima,

Classification Learner App is just an UI, interior the same Deep learning algorithm has been implemented. If you are using logistic regression the default setting uses “fitglm” function as you mentioned, this function just tries to model the probability of certain class or events existing such as pass/fail. The class with more probability assigned as the predicted class. If you are interested more in finding the algorithm behind it, you can check out following link.

https://www.mathworks.com/help/stats/fitglm.html#mw_0fd53dcd-2470-4661-8735-434f6ef41591

I hope this helps you.