Skip to content

Latest commit

 

History

History
37 lines (28 loc) · 1.54 KB

README.md

File metadata and controls

37 lines (28 loc) · 1.54 KB

SpLearn: A CS6350 Machine Learning Library

SpLearn is a machine learning library developed by Spencer Elkington for use during the CS6350: Machine Learning course at the University of Utah.

SpLearn is built to run off of the same general interface as the scikit-learn package commmonly used for statistical analysis and modelling.

Getting Started

You can run run.sh in order to install dependencies and begin a new Jupyter Lab session. Python files and notebooks can use import splearn to begin using SpLearn models.

SpLearn is dependent on:

Documentation

All machine learning models and methods in SpLearn use PyDocs to document method signatures, along with weak typing provided. In general, training and test data will accept Pandas DataFrame or Series objects.

The available SpLearn packages are:

  • SpLearn.Metrics: A collection of metric methods used for model testing and additional utility
  • SpLearn.DecisionTree: A collection of DecisionTree models
  • SpLearn.EnsembleLearning: A collection of Ensemble learning classifiers that make use of weaker modesl to boost accuracy.
  • SpLearn.Linear: A collection of regression algorithms to generate models with continuous outputs.
  • SpLearn.Perceptron: A Perceptron learning model used to create boundaries for linear separability.