
Load anaconda

http://scikit-learn.org/stable/tutorial/basic/tutorial.html

scikit learn: 
https://www.youtube.com/watch?v=80fZrVMurPM
Starting at 1 hour see the general framework of fitting on training data
and then predicting on test data and seeing how good the model is.

from sklearn import datasets
iris=datasets.load_iris()
digits=datasets.load_digits()

b = numpy.array([range(5), [6, 1, 3, 4, 2], [1, 9, 6, 3, 8, 4])
t = numpy.array(range(5))

 digits.target = t
digits.data = b

from sklearn import svm
clf = svm.SVC(gamma=0.001, C=100.)
clf.fit(digits.data[:-1], digits.target[:-1]) 
clf.predict(digits.data[-1:])

clf.socre(X_test, y_test)

matplotlib -- graphics package

=====

In parallel approach, Bagging is based on Bootstraps. Each classifier is ′′
trained on a set of n training instances (n < n), drawn randomly with re- 120 placement from the original training set of size n. Such a training set is called a Bootstrap replicate of the original set. Each Bootstrap replicate contains, on average, 63.2% of the original training set, with many instances appearing sev- eral times. Predictions on the new instances are made by taking the majority vote of the ensemble. The particularity of these training sets is to reduce the
125 impact of hard instances to learn (called outliers and misleaders) [20].
