This repository contains popular Machine Learning algorithms, which have been introduced in various blog posts (http://ataspinar.com). Most of the algorithms are accompanied with blog-posts in which I try to explain the mathematics behind and the interpretation of these algorithms.

Machine Learning is fun! But more importantly, Machine Learning is easy. But the academic literature or even (wikipedia-pages) is full with unnecessary complicated terminology, notation and formulae. This gives people the idea that these ML algorithms can only be understood with a full understanding of advanced math and statistics. Stripped from all of these superfluous language we are left with simple maths which can be expressed in a few lines of code.

I have also provided some notebooks, explaining the mathematics of some Machine Learning algorithms.

• Linear Regression and Logistic Regression
• Naive Bayes Classification

To install siML:

(sudo) pip install siml

or you can clone the repository and in the folder containing setup.py

python setup.py install

Once it has been installed, the logistic regression algorithm can be used as follows:

from siml import classifiers

alpha = 0.5
theta = np.ones(no_features)
theta = classifiers.gradient_descent(X, Y, theta, alpha, no_training_examples)

'theta' is a vector which now contains the parameter values, calculated with the training set X,Y. This vector can be used to classify new examples in the test-set X_test, Y_test (with the dot-product):

determined_Y_test = [np.dot(theta, X_test[ii]) for ii in range(no_test_examples)]
determined_Y_test_binary = [classifiers.to_binary(elem) for elem in determined_Y_test]

The determined Y values can be compared against the actual Y-values with the F1-score:

f1 = evaluators.f1_score(determined_Y_test_binary, Y_test)
print "\nf1-score on the test-set is %s" % (f1)