km.fit(rules[k])
nrules[k] = km.cluster_centers_
nvotemdl = SimpleVoting(nrules)
y = nvotemdl.fit(data[test_idx, :])
correct.append(
sum([1 for i, j in zip(y, labels[test_idx]) if i == j]) /
len(labels[test_idx]))
igbincorrect = []
for i in range(2, n_clusters + 1):
nrules = {}
for k in rulesbin.keys():
km = NRules(i=k, n_clusters=i)
km.fit(rulesbin[k])
km.restore(data[train_idx, :], labels[train_idx],
RulesStats.infogain)
nrules[k] = km.cluster_centers_
binvotemdl = SimpleVoting(nrules)
y = binvotemdl.fit(data[test_idx, :])
igbincorrect.append(
sum([1 for i, j in zip(y, labels[test_idx]) if i == j]) /
len(labels[test_idx]))
stbincorrect = []
for i in range(2, n_clusters + 1):
nrules = {}
for k in rulesbin.keys():
km = NRules(i=k, n_clusters=i)
km.fit(rulesbin[k])
km.restore(data[train_idx, :], labels[train_idx],
nvotemdl = SimpleVoting(nrules)
y = nvotemdl.fit(data[test_idx, :])
correct.append(
sum(
[1 for i, j in zip(y, labels[test_idx]) if i == j]
) / len(labels[test_idx])
)
igbincorrect = []
for i in range(2, n_clusters + 1):
nrules = {}
for k in rulesbin.keys():
km = NRules(i=k, n_clusters=i)
km.fit(rulesbin[k])
km.restore(
data[train_idx, :], labels[train_idx],
RulesStats.infogain
)
nrules[k] = km.cluster_centers_
binvotemdl = SimpleVoting(nrules)
y = binvotemdl.fit(data[test_idx, :])
igbincorrect.append(
sum(
[1 for i, j in zip(y, labels[test_idx]) if i == j]
) / len(labels[test_idx])
)
stbincorrect = []
for i in range(2, n_clusters + 1):
nrules = {}
for k in rulesbin.keys():
