weightlist[datatotrain[tr]] *= beta[-1]
sum = 0
for i in weightlist:
sum += i
for i in range(len(weightlist)):
weightlist[i] /= sum
print(error)
#test
sum = 0
tp = 0
fp = 0
fn = 0
tn = 0
result = []
for lea in learner:
result.append(decision.getResult(lea, data_example, ma, 'bot', test))
rs = []
for i in range(len(test)):
total = 0
rs.append(0)
for j in range(len(learner)):
total += 1
if result[j][i] == 1:
rs[-1] += math.log(1 / beta[j], 2)
else:
rs[-1] -= math.log(1 / beta[j], 2)
if rs[-1] > 0:
rs[-1] = 1
elif rs[-1] < 0:
rs[-1] = 0
elif i % 2 == 0:
row = []
for name in ma:
row.append(data_trai[name][i])
thedatatr.append(row)
targettr.append(data_trai['bot'][i])
for i in range(len(data_test)):
row = []
for name in ma:
row.append(data_test[name][i])
thedatate.append(row)
for k in range(1):
model = mybayes.buildmodel(data_trai, ma, 'bot', trainindexlist)
result1 = model.getResult(data_test, ma, 'bot', testindexlist)
mynode = decision.makesubtree(data_trai, ma, 'bot', trainindexlist)
result2 = decision.getResult(mynode, data_test, ma, 'bot', testindexlist)
logreg = LogisticRegression(C=1000).fit(thedatatr, targettr)
thetar = logreg.predict(thedatate)
result = []
for i in range(len(result1)):
t = result1[i] + result2[i] + thetar[i]
if t >= 2:
result.append(1)
else:
result.append(0)
accuracy = 0.0
precision = 0.0
recall = 0.0
f1score = 0.0
tp = 0
fp = 0