## command-line program reference

mlpack provides command-line executables for many of the algorithms it implements. These may be used to perform many machine learning tasks without the overhead of writing C++, or may be used as part of a larger machine learning solution.

Below is a list of the command-line executables mlpack 1.0.9 provides, with links to the documentation for each executable. This documentation may also be accessed with the --help parameter or through the man pages provided with your distribution of mlpack.

- allknn: all
*k*-nearest neighbor search with trees - allkfn: all
*k*-furthest neighbor search with trees - allkrann: rank-approximate
*k*-nearest neighbor search with trees - cf: generate recommendations via collaborative filtering
- decision_stump: classify with a decision stump
- det: density estimation trees
- emst: calculate a Euclidean minimum spanning tree
- fastmks: perform fast max-kernel search with trees
- gmm: train or classify with a Gaussian mixture model
- hmm_generate: generate observations from a hidden Markov model (HMM)
- hmm_loglik: calculate the log-likelihood of some observations from an HMM
- hmm_train: train a hidden Markov model (HMM)
- hmm_viterbi: find the most probable hidden states in an HMM for some observations
- kernel_pca: perform kernel principal components analysis
- kmeans: perform
*k*-means clustering - lars: least-angle regression
- linear_regression: simple least-squares linear regression
- local_coordinate_coding: local coordinate coding
- logistic_regression: train or classify with logistic regression
- lsh: approximate
*k*-nearest neighbor search with locality-sensitive hashing - nbc: train or classify with the naive Bayes classifier
- nca: neighborhood components analysis
- nmf: non-negative matrix factorization
- pca: principal components analysis
- perceptron: train or classify with a perceptron
- radical: RADICAL (independent components analysis)
- range_search: range search with trees
- sparse_coding: sparse coding with dictionary learning