Scikit-learn:

  • sklearn is built on top of NumPy and Pandas, and is designed to work with them.
  • Includes estimator objects and algorithms to support many machine learning operations.
  • It supports both supervised learning methods such as classification and regression and unsupervised methods such as clustering and dimensionality reduction. In addition, it provides much of the software infrastructure to enable the construction of machine learning pipelines, supporting activities such as train-test splitting, cross validation, and hyperparameter optimization.

The sklearn Package:

  • The sklearn Python module supports “classical machine learning algorithms”
  • Machine-learning algorithms have been developed and used for decades, although much of the recent excitement around the field revolves around newer “deep-learning” algorithms that use neural networks for their implementation. While sklearn does provide some support for machine learning using neural networks, that is not its primary focus, hence the emphasis on “classical” algorithms.
  • A strength of sklearn is that it supports the use and comparison of many different algorithms within one consistent and integrated framework.