/ml

A set of machine learning experiments in Clojure

Primary LanguageClojureMIT LicenseMIT

ML

ML is a machine-learning library/framework for Clojure. It's currently a pet project, but I am trying to keep algorithms as general and composable as possible. Flexibility, and not efficiency, is the first concern. (Efficiency will come later.) Feel free to fork, experiment, and contribute!

To implement

  • k-nearest neighbors
  • Linear regression
  • Logistic regression
  • Neural network for logistic regression feature discovery
    • Predict values based on pre-generated weights
    • Calculate cost of values
    • Use backpropagation to adjust weights
  • Regularization for linear regression
  • Regularization for logistic regression
  • Regularization for neural network
  • Decision tree builder
  • Support Vector Machine
  • Naive Bayes classifier
  • K-means clustering