def problem1():
sample = generate_sample()
kms = {}
for K in [2,3,4,5]:
km = KMeans(sample, K=K)
km.iterate(20)
kms[K] = km
return kms
def problem2_3(rec=None):
if rec is None:
rec = problem2()
U, V = rec.U, rec.V
km = KMeans(U, K=30)
km.iterate(10)
centroids = km.MU[km.MU != 0].dropna()
assert len(centroids) >= 5, 'Not enough centroids!'
print len(centroids), 'centroids'
ptypes = set()
while len(ptypes) < 5:
ptypes.add(random.choice(centroids.index))
similarities = map_ptypes(ptypes, centroids, V)
pprint(similarities)
return km
