Getting Started with Mltooler

Installation using pip

Crack open your command prompt and type

pip install mltooler

That was easy wasn’t it?

Mltooler requires the latest version of multiple 3rd party packages, make sure you update to the required version if asked.

Huge thanks goes out to the creators and contributors of these!

Importing Mltooler

Importing mltooler is again easy as pie - no bells and whistles here.

from mltooler import mltools as mt

Improving your first model

The SelfImprovingEstimator is what drives Mltooler. All you need is some training data, and the default settings will work their magic. This example shows the basics with minimal lines of code.

(See full API documentation for advanced usage and examples)

from mltooler import mltools as mt
sie = mt.SelfImprovingEstimator('knn')
sie.self_improve(X,y)

Once running the SelfImprovingEstimator will iteratively improve its cross validation score through a series of tests adding data transformations to its pipeline.

Once fit you can use the SelfImprovingEstimator like any scikit-learn estimator.

sie.fit(X_train,y_train)
preds = sie.predict(X_test)