• Given: N samples of data
• Goal: predict properties of unseen data

• Simple and efficient tools for data mining and data analysis
• Built on NumPy, SciPy, and matplotlib
• Open source, BSD license
• Under active development

from sklearn import svm

train_X = ...  # load sample inputs
train_y = ...  # load sample outputs
to_predict = ...  # we want to predict the output of this value

clf = svm.SVC()  # create a classifier

clf.fit(train_X, train_y)  # train the classifier

prediction = clf.predict(to_predict)  # predict our unknown output

scikit-learn.org/stable/documentation.html

from sklearn import tree

train_X = ...  # load sample inputs
train_y = ...  # load sample outputs
to_predict = ...  # we want to predict the output of this value

clf = tree.DecisionTreeClassifier()  # create a classifier

clf.fit(train_X, train_y)  # train the classifier

prediction = clf.predict(to_predict)  # predict our unknown output

scikit-learn.org/stable/documentation.html

• Optimization of Facebook campaigns
• Regression to predict certain outputs of a campaign