Jeff Levesque

The unit tests may check the probability, among other logit characteristics: [Week9_SensitivityAnalysisTitanicCompleted_2_2_2_2.xlsx](https://github.com/jeff1evesque/machine-learning/files/2089895/Week9_SensitivityAnalysisTitanicCompleted_2_2_2_2.xlsx)

It seems the latter link evolved to the [mlxtend](http://rasbt.github.io/mlxtend/) package, as also demonstrated: - https://stackoverflow.com/a/43298736

Our next milestone 0.8 will be application development heavy. Therefore, for the sake of organization, we'll relocate this, and related application development issues.

We could write a custom class, with corresponding methods, to validate specific object types. This type of granular validation, could be reused across different layers of restructuring, and parsing respective...

We will create class methods for the following non-exhaustive cases: - strings - list of strings - integers - list of integers - floats - list of floats - class...

This is very good. Assume that I don't understand the meaning of the variables for this bagging classifier, and I only understand basic statistics (dependent vs independent variables). What does...

What are some of the other (supported / nonsupported) options, other than the decision tree `base_estimator`, for either the classifier, or the regressor? Can we do random forrest, or some...

Can the api readily support the following classifiers, and regressors as its base estimator: - `svm` - `svr` - `random forrest` - `knn` - `decision tree` (this their default) -...

Ah, took a little bit more time thinking of your [statement](https://github.com/jeff1evesque/machine-learning/issues/3025#issuecomment-333382073): >The base estimator can basically be any type of classifier. You just give it an object and it'll train...

That's great! My guess, is that all of them would work for the `BaggingRegressor`, except the `svm`? Basically, an inverse of your most previous statement.