fauc = open('%s-auc.dat' % bname, 'w+')
auc = m.auc(aa_learning, pt_learning)
print("%.4f" % auc, end='', file=fauc)
fauc.close()
fmisclassified = open('%s-misclassified.dat' % bname, 'w+')
print("{Alternative} ", file=fmisclassified, end='')
print("{Original assignment} ", file=fmisclassified, end='')
print("{Model assignment}", file=fmisclassified, end='')
for c in criteria:
print(" {%s}" % criteria_names[c], file=fmisclassified, end='')
print("\n", file=fmisclassified, end='')
misclassified_aids = []
for aid in aa_learning.keys():
aa1 = aa_learning[aid].category_id
aa2 = aa_learned[aid].category_id
if aa1 == aa2:
continue
misclassified_aids.append(aid)
misclassified_aids.sort(key=lambda item: (len(item), item))
for aid in misclassified_aids:
ap = pt_learning[aid]
aa1 = aa_learning[aid].category_id
aa2 = aa_learned[aid].category_id
print("%s " % aid, file=fmisclassified, end='')
print("%s %s" % (aa1, aa2), file=fmisclassified, end='')
ap1 = AlternativePerformances('a1', {'c1': 1, 'c2': 1, 'c3': 0, 'c4': 0})
ap2 = AlternativePerformances('a2', {'c1': 0, 'c2': 0, 'c3': 1, 'c4': 1})
ap3 = AlternativePerformances('a3', {'c1': 1, 'c2': 0, 'c3': 1, 'c4': 0})
ap4 = AlternativePerformances('a4', {'c1': 1, 'c2': 0, 'c3': 0, 'c4': 1})
ap5 = AlternativePerformances('a5', {'c1': 0, 'c2': 1, 'c3': 1, 'c4': 0})
ap6 = AlternativePerformances('a6', {'c1': 0, 'c2': 1, 'c3': 0, 'c4': 1})
pt = PerformanceTable([ap1, ap2, ap3, ap4, ap5, ap6])
aa1 = AlternativeAssignment('a1', 'cat1')
aa2 = AlternativeAssignment('a2', 'cat1')
aa3 = AlternativeAssignment('a3', 'cat2')
aa4 = AlternativeAssignment('a4', 'cat2')
aa5 = AlternativeAssignment('a5', 'cat2')
aa6 = AlternativeAssignment('a6', 'cat2')
aa = AlternativesAssignments([aa1, aa2, aa3, aa4, aa5, aa6])
print_pt_and_assignments(aa.keys(), c.keys(), [aa], pt)
model = MRSort(c, None, None, None, cps)
worst = pt.get_worst(model.criteria)
best = pt.get_best(model.criteria)
# Run the MIP
mip = MipCMRSort(model, pt, aa)
mip.solve()
# Display learned model parameters
print('Learned model')
print('=============')
model.bpt.display()
model.cv.display()
table_ca_learning.append(ca_learning)
table_ca_test.append(ca_test)
# Compute area under the curve
auc_learning = m.auc(aa_learning, pt_learning)
auc_test = m.auc(aa_test, pt_test)
table_auc_learning.append(auc_learning)
table_auc_test.append(auc_test)
if m.veto_lbda is not None:
nveto += 1
# Compute confusion matrices
for a in aa_learning.keys():
key = (aa_learning[a].category_id, aa_learning_m2[a].category_id)
if key in cmatrix_learning:
cmatrix_learning[key] += 1
else:
cmatrix_learning[key] = 1
for a in aa_test.keys():
key = (aa_test[a].category_id, aa_test_m2[a].category_id)
if key in cmatrix_test:
cmatrix_test[key] += 1
else:
cmatrix_test[key] = 1
print("nveto: %d" % nveto)
avg_ca_learning = sum(table_ca_learning) / float(len(table_ca_learning))
ap1 = AlternativePerformances('a1', {'c1': 1, 'c2': 1, 'c3': 0, 'c4': 0})
ap2 = AlternativePerformances('a2', {'c1': 0, 'c2': 0, 'c3': 1, 'c4': 1})
ap3 = AlternativePerformances('a3', {'c1': 1, 'c2': 0, 'c3': 1, 'c4': 0})
ap4 = AlternativePerformances('a4', {'c1': 1, 'c2': 0, 'c3': 0, 'c4': 1})
ap5 = AlternativePerformances('a5', {'c1': 0, 'c2': 1, 'c3': 1, 'c4': 0})
ap6 = AlternativePerformances('a6', {'c1': 0, 'c2': 1, 'c3': 0, 'c4': 1})
pt = PerformanceTable([ap1, ap2, ap3, ap4, ap5, ap6])
aa1 = AlternativeAssignment('a1', 'cat1')
aa2 = AlternativeAssignment('a2', 'cat1')
aa3 = AlternativeAssignment('a3', 'cat2')
aa4 = AlternativeAssignment('a4', 'cat2')
aa5 = AlternativeAssignment('a5', 'cat2')
aa6 = AlternativeAssignment('a6', 'cat2')
aa = AlternativesAssignments([aa1, aa2, aa3, aa4, aa5, aa6])
print_pt_and_assignments(aa.keys(), c.keys(), [aa], pt)
model = MRSort(c, None, None, None, cps)
worst = pt.get_worst(model.criteria)
best = pt.get_best(model.criteria)
# Run the MIP
mip = MipCMRSort(model, pt, aa)
mip.solve()
# Display learned model parameters
print('Learned model')
print('=============')
model.bpt.display()
model.cv.display()
table_ca_learning.append(ca_learning)
table_ca_test.append(ca_test)
# Compute area under the curve
auc_learning = m.auc(aa_learning, pt_learning)
auc_test = m.auc(aa_test, pt_test)
table_auc_learning.append(auc_learning)
table_auc_test.append(auc_test)
if m.veto_lbda is not None:
nveto += 1
# Compute confusion matrices
for a in aa_learning.keys():
key = (aa_learning[a].category_id, aa_learning_m2[a].category_id)
if key in cmatrix_learning:
cmatrix_learning[key] += 1
else:
cmatrix_learning[key] = 1
for a in aa_test.keys():
key = (aa_test[a].category_id, aa_test_m2[a].category_id)
if key in cmatrix_test:
cmatrix_test[key] += 1
else:
cmatrix_test[key] = 1
print("nveto: %d" % nveto)
avg_ca_learning = sum(table_ca_learning) / float(len(table_ca_learning))
matrix = compute_confusion_matrix(aa_test_m1, aa_test_m2, m2.categories)
print_confusion_matrix(matrix, m2.categories)
aids = [a.id for a in aa_test_m1 \
if aa_test_m1[a.id].category_id != aa_test_m2[a.id].category_id]
if len(aids) > 0:
print("List of alternatives wrongly assigned:")
print_pt_and_assignments(aids, None, [aa_test_m1, aa_test_m2],
pt_test)
if type(m2) == MRSort:
worst = AlternativePerformances('worst', {c.id: 0 for c in m2.criteria})
best = AlternativePerformances('best', {c.id: 1 for c in m2.criteria})
categories = m2.categories
a_learning = aa_learning_m1.keys()
pt_learning_ok = []
pt_learning_too_low = []
pt_learning_too_high = []
for a in a_learning:
i1 = categories.index(aa_learning_m1[a].category_id)
i2 = categories.index(aa_learning_m2[a].category_id)
if i1 == i2:
pt_learning_ok.append(pt_learning[a])
elif i1 < i2:
pt_learning_too_high.append(pt_learning[a])
elif i1 > i2:
pt_learning_too_low.append(pt_learning[a])
a_test = aa_test_m1.keys()
pt_test_ok = []