How does classification learner handle missing data for logistic regression?

I am curious on how missing data is handled in the classification learner for methods that do not have approaches for missing data (like logistic regression).

[Ive J]

fitglm excludes samples with missing values (in either of predictors/response).

Answer 1

[Sahil Jain]

The documentation of generalized linear regression model (fitglm) mentions that rows with even a single missing value are removed from the fit (either predictor or response). However, this behaviour can vary depending on the type of classifier being used. For example, binary decision trees (fitctree) can use surrogate decision splits to improve accuracy as long as not all the predictor values are missing for a row. Information about how different classifiers deal with missing data is described in their respective documentation.