The skippy python package

Skip the boilerplate of scikit-learn machine learning examples.

Installation

pip install skippy

Usage

In a shell environment, you can run skippy with no arguments to perform a Logistic Regression on the digits dataset.

This will produce a 10 x 10 confusion matrix with the Accuracy Score at the top.

You can also pass arguments to skippy at the command line.

For example,

skippy -data diabetes -type linear_model -name Lasso
# Or
skippy -d diabetes -t linear_model -n Lasso

will run a linear regression with lasso regularization (L1) on the diabetes dataset.

The data argument can be any of the following built-in scikit-learn datasets:

• Regression
  • boston
  • diabetes
• Classification
  • digits
  • iris
  • wine
  • breast_cancer

The type and name arguments are referring to the model type and name from scikit-learn. The type is the submodule, e.g.

• linear_model
• naive_bayes
• ensemble
• svm

while the name is the what is actually imported, e.g.

• LinearRegression
• GaussianNB
• RandomForestRegressor
• SVC

Simplify code to a single function call per step:

from sklearn.metrics import confusion_matrix, accuracy_score
import skippy as skp

data = skp.get_data('digits')
x_train, x_test, y_train, y_test = skp.split_data(data)

model = skp.get_model(model_type='ensemble',
                      model_name='RandomForestClassifier')

fit = model.fit(x_train, y_train)
skp.pickle_model(filename='digits_rf.pickle', model=fit)
predictions = fit.predict(x_test)

confmat = confusion_matrix(y_true=y_test, y_pred=predictions)
accuracy = accuracy_score(y_true=y_test, y_pred=predictions)

skp.confusion_matrix_plot(cm=confmat,
                          acc=accuracy,
                          filename='digits_rf.png')

Or run a whole pipeline with one function:

import skippy as skp

skp.classification(dataset='digits',
                   model_type='ensemble',
                   model_name='RandomForestClassifier',
                   pickle_name='digits_rf.pickle',
                   plot_name='digits_rf.png')

For inspiration, look at the example pipelines in the pipelines folder of the skippy repo.