Skip to content

Simple machine-learning algorithms for classification and clustering, as well as model evaluation tools such as performance measures, cross-validation and tests.

cukinhou/my-ML

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

MyML

MyML will implement simple machine-learning algorithms for classification and clustering, as well as model evaluation tools such as performance measures, cross-validation and tests. The goal of this repo is to keep track of my findings and personal skills with machine learning in python. In order to allow scalability, each algorithm implements a Classifier, Clustering or DimensionalityReduction object depending on the type of task it performs. Theory can be found in the Wiki.

Currently (21/05/2017) the library includes the following methods:

  • Perceptron or Single-layer perceptron: is an algorithm for supervised learning of binary classifiers.

Usage example

from myML.classification.perceptron import Perceptron

x_train =  [
            [2.7810836, 2.550537003],
            [1.465489372, 2.362125076],
            [3.396561688, 4.400293529],
            [1.38807019, 1.850220317],
            [3.06407232, 3.005305973],
            [7.627531214, 2.759262235],
            [5.332441248, 2.088626775],
            [6.922596716, 1.77106367],
            [8.675418651, -0.242068655],
            [7.673756466,   3.508563011]
        ]
y_train =  [0, 0, 0, 0, 0, 1, 1, 1, 1, 1]
l_rate = 0.1
iter = 5
perceptron = Perceptron()
perceptron.fit(x_train, y_train, l_rate, iter)
perceptron.predict([2.7810836,2.550537003])

Install

If you do not have pip installed in your machine, run the following command:

  • easy_install pip

For installing MyML run (you might have to use sudo for super user privileges):

  • make install

For running tests:

  • make test

About

Simple machine-learning algorithms for classification and clustering, as well as model evaluation tools such as performance measures, cross-validation and tests.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published