anomaly_sparse.py

from sklearn.decomposition import DictionaryLearning
from sklearn.decomposition import SparseCoder
import pandas as pd

# load data from CSV 
df = pd.read_csv('/mnt/c/Users/davev/Documents/test_sparse.csv')
# get rid of the "label" column - AS Number in our case
del df['AS Number']

# change data into required format from scikit learn
t=df.as_matrix()

# create a dictionary with 2 components (to make it easier to plot later)
# the dictionary is learnt by iterating over the data a 100 times
dict=DictionaryLearning(n_components=2, max_iter=100)
dict.fit(t)

# load the dictionary we just created into a Sparse Coder
sp = SparseCoder(dict.components_)
# instruct the sparse coder to represent our data in terms of the dictionary we previously "learnt"
sp.transform(t)

# ... [results displayed] ...