def basic_selection(common_path,binary_path,clf="LR"):
datasets=ens.read_dataset(common_path,binary_path)
acc=np.array([ validate_acc(data_i).get_acc()
for data_i in datasets])
s_clf=dataset_selection(datasets,acc)
print(len(s_clf))
return ens.ensemble(common_path,binary_path,True,clf,s_clf)[0]
def selection_exp(common,binary,cf_path=None): s_clf=random_selection(common,binary,1000,27,clf="LR") print(s_clf) result,votes=ens.ensemble(common,binary, clf="LR",binary=False,s_clf=s_clf) result.report() if(cf_path): result.get_cf(cf_path)
def select_ens(common, binary):
best_set = selection.random_selection(common, binary, 1000, 27)
result_i = ens.ensemble(common,
binary,
s_clf=best_set,
clf="LR",
binary=False)[0]
desc_i = exp.exp_info(common, binary, result_i)
return "%d,%s" % (len(best_set), desc_i[-1])
def exp_person(common, binary, n, n_clf, clf="LR"):
fun = person.total_person_selection
# fun=basic.total_basic_selection
all_clf = fun(common, binary, n, n_clf, clf)
acc = []
for all_clf_i in all_clf:
result = ens.ensemble(common, binary, clf, False, all_clf_i)[0]
acc.append(result.get_acc())
for acc_i, clf_set_i in zip(acc, all_clf):
print(clf_set_i)
print(acc_i)
def inliner_exp(dtw, deep, binary1, binary2):
paths = [(deep, binary1), (deep + dtw, binary1), (deep, [binary1,
binary2]),
(deep + dtw, [binary1, binary2])]
results = []
for common_i, binary_i in paths:
basic_i = ens.ensemble(common_i, binary_i, clf="LR", binary=False)
results.append((False, common_i, binary_i, basic_i))
inline_i = inliner_voting(common_i, binary_i, clf="LR")
results.append((True, common_i, binary_i, inline_i))
for inline_i, common_i, binary_i, result_i in results:
desc_common, desc_binary, metrics = exp.exp_info(
common_i, binary_i, result_i)
line_i = (int(inline_i), desc_common, desc_binary, metric)
print("%d,%s,%s,%s" % line_i)
def reduced_selection(common, binary, n_feats):
def helper(common, binary, clf="LR"):
read = reduction.SepSelected(n_feats, 0)
return ens.make_votes(common, binary, "LR", read)
s_clf = selection.random_selection(common,
binary,
1000,
27,
clf="LR",
fun=helper)
read = reduction.SepSelected(n_feats, 0)
result, votes = ens.ensemble(common,
binary,
clf="LR",
binary=False,
s_clf=s_clf,
read=read)
return result, s_clf
def basic_exp(in_path):
paths = files.get_paths(in_path, name="dtw")
result = ens.ensemble(paths[0], None, clf="LR", binary=False)[0]
result.report()
def person_selection(common_path, binary_path, clf="LR"):
datasets = ens.read_dataset(common_path, binary_path)
clf_acc = np.array([person_acc(data_i).get_acc() for data_i in datasets])
s_clf = acc.dataset_selection(datasets, clf_acc)
print(len(s_clf))
return ens.ensemble(common_path, binary_path, True, clf, s_clf)[0]
def simple_ens(common, binary):
result_i = ens.ensemble(common, binary, clf="LR", binary=False)[0]
desc_i = exp.exp_info(common, binary, result_i)
return desc_i[-1]
def acc_curve(common,binary,clf,clfs): clf_selction=range(2,len(clfs)) results=[ens.ensemble(common,binary,False,clf,clfs[:k])[0] for k in clf_selction] acc=[result_i.get_acc() for result_i in results] print(acc)