def computing_cv_accuracy_imprecise(in_path=None, ell_optimal=0.1, cv_n_fold=10):
def u65(mod_Y):
return 1.6 / mod_Y - 0.6 / mod_Y ** 2
def u80(mod_Y):
return 2.2 / mod_Y - 1.2 / mod_Y ** 2
data = export_data_set('iris.data') if in_path is None else pd.read_csv(in_path)
print("-----DATA SET TRAINING---", in_path)
X = data.iloc[:, :-1].values
y = np.array(data.iloc[:, -1].tolist())
mean_u65, mean_u80 = 0, 0
lqa = LinearDiscriminant(init_matlab=True)
kf = KFold(n_splits=cv_n_fold, random_state=None, shuffle=True)
for idx_train, idx_test in kf.split(y):
X_cv_train, y_cv_train = X[idx_train], y[idx_train]
X_cv_test, y_cv_test = X[idx_test], y[idx_test]
lqa.learn(X_cv_train, y_cv_train, ell=ell_optimal)
sum_u65, sum_u80 = 0, 0
n_test, _ = X_cv_test.shape
for i, test in enumerate(X_cv_test):
print("--TESTING-----", i, ell_optimal)
evaluate, _ = lqa.evaluate(test)
print(evaluate, "-----", y_cv_test[i])
if y_cv_test[i] in evaluate:
sum_u65 += u65(len(evaluate))
sum_u80 += u80(len(evaluate))
mean_u65 += sum_u65 / n_test
mean_u80 += sum_u80 / n_test
mean_u65 = mean_u65 / cv_n_fold
mean_u80 = mean_u80 / cv_n_fold
print("--ell-->", ell_optimal, "--->", mean_u65, mean_u80)
def computing_cv_accuracy_imprecise(in_path=None,
ell_optimal=0.1,
cv_n_fold=10):
def u65(mod_Y):
return 1.6 / mod_Y - 0.6 / mod_Y**2
def u80(mod_Y):
return 2.2 / mod_Y - 1.2 / mod_Y**2
data = export_data_set('iris.data') if in_path is None else pd.read_csv(
in_path)
print("-----DATA SET TRAINING---", in_path)
X = data.iloc[:, :-1].values
y = np.array(data.iloc[:, -1].tolist())
mean_u65, mean_u80 = 0, 0
lqa = LinearDiscriminant(init_matlab=True)
kf = KFold(n_splits=cv_n_fold, random_state=None, shuffle=True)
for idx_train, idx_test in kf.split(y):
X_cv_train, y_cv_train = X[idx_train], y[idx_train]
X_cv_test, y_cv_test = X[idx_test], y[idx_test]
lqa.learn(X_cv_train, y_cv_train, ell=ell_optimal)
sum_u65, sum_u80 = 0, 0
n_test, _ = X_cv_test.shape
for i, test in enumerate(X_cv_test):
print("--TESTING-----", i, ell_optimal)
evaluate, _ = lqa.evaluate(test)
print(evaluate, "-----", y_cv_test[i])
if y_cv_test[i] in evaluate:
sum_u65 += u65(len(evaluate))
sum_u80 += u80(len(evaluate))
mean_u65 += sum_u65 / n_test
mean_u80 += sum_u80 / n_test
mean_u65 = mean_u65 / cv_n_fold
mean_u80 = mean_u80 / cv_n_fold
print("--ell-->", ell_optimal, "--->", mean_u65, mean_u80)
def computing_best_imprecise_mean(in_path=None,
seed=0,
cv_n_fold=10,
from_ell=0.1,
to_ell=1.0,
by_ell=0.1):
def u65(mod_Y):
return 1.6 / mod_Y - 0.6 / mod_Y**2
def u80(mod_Y):
return 2.2 / mod_Y - 1.2 / mod_Y**2
data = export_data_set('iris.data') if in_path is None else pd.read_csv(
in_path)
print("-----DATA SET TRAINING---", in_path)
X = data.iloc[:, :-1].values
y = np.array(data.iloc[:, -1].tolist())
ell_u65, ell_u80 = dict(), dict()
lqa = LinearDiscriminant(init_matlab=True)
X_train, X_test, y_train, y_test = \
train_test_split(X, y, test_size=0.4, random_state=seed)
kf = KFold(n_splits=cv_n_fold, random_state=None, shuffle=True)
splits = list([])
for idx_train, idx_test in kf.split(y_train):
splits.append((idx_train, idx_test))
print("Splits --->", splits)
for ell_current in np.arange(from_ell, to_ell, by_ell):
ell_u65[ell_current], ell_u80[ell_current] = 0, 0
for idx_train, idx_test in splits:
print("---k-FOLD-new-executing--")
X_cv_train, y_cv_train = X_train[idx_train], y_train[idx_train]
X_cv_test, y_cv_test = X_train[idx_test], y_train[idx_test]
lqa.learn(X_cv_train, y_cv_train, ell=ell_current)
sum_u65, sum_u80 = 0, 0
n_test = len(idx_test)
for i, test in enumerate(X_cv_test):
evaluate, _ = lqa.evaluate(test)
print("----TESTING-----", i, ell_current, "|---|", evaluate,
"-----", y_cv_test[i])
if y_cv_test[i] in evaluate:
sum_u65 += u65(len(evaluate))
sum_u80 += u80(len(evaluate))
ell_u65[ell_current] += sum_u65 / n_test
ell_u80[ell_current] += sum_u80 / n_test
print("-------ELL_CURRENT-----", ell_current)
ell_u65[ell_current] = ell_u65[ell_current] / cv_n_fold
ell_u80[ell_current] = ell_u80[ell_current] / cv_n_fold
print("u65-->", ell_u65[ell_current])
print("u80-->", ell_u80[ell_current])
print("--->", ell_u65, ell_u80)
def output_paper_result(in_path=None):
data = export_data_set(
'bin_normal_rnd.data') if in_path is None else pd.read_csv(in_path)
X = data.loc[:, ['x1', 'x2']].values
y = data.y.tolist()
lqa = LinearDiscriminant(init_matlab=True)
lqa.learn(X, y, ell=2)
lqa.plot2D_classification(np.array([2, 2]), colors={0: 'red', 1: 'blue'})
lqa.plot2D_decision_boundary()
def computing_best_imprecise_mean(in_path=None, seed=0, cv_n_fold=10,
from_ell=0.1, to_ell=1.0, by_ell=0.1):
def u65(mod_Y):
return 1.6 / mod_Y - 0.6 / mod_Y ** 2
def u80(mod_Y):
return 2.2 / mod_Y - 1.2 / mod_Y ** 2
data = export_data_set('iris.data') if in_path is None else pd.read_csv(in_path)
print("-----DATA SET TRAINING---", in_path)
X = data.iloc[:, :-1].values
y = np.array(data.iloc[:, -1].tolist())
ell_u65, ell_u80 = dict(), dict()
lqa = LinearDiscriminant(init_matlab=True)
X_train, X_test, y_train, y_test = \
train_test_split(X, y, test_size=0.4, random_state=seed)
kf = KFold(n_splits=cv_n_fold, random_state=None, shuffle=True)
splits = list([])
for idx_train, idx_test in kf.split(y_train):
splits.append((idx_train, idx_test))
print("Splits --->", splits)
for ell_current in np.arange(from_ell, to_ell, by_ell):
ell_u65[ell_current], ell_u80[ell_current] = 0, 0
for idx_train, idx_test in splits:
print("---k-FOLD-new-executing--")
X_cv_train, y_cv_train = X_train[idx_train], y_train[idx_train]
X_cv_test, y_cv_test = X_train[idx_test], y_train[idx_test]
lqa.learn(X_cv_train, y_cv_train, ell=ell_current)
sum_u65, sum_u80 = 0, 0
n_test = len(idx_test)
for i, test in enumerate(X_cv_test):
evaluate, _ = lqa.evaluate(test)
print("----TESTING-----", i, ell_current, "|---|", evaluate, "-----", y_cv_test[i])
if y_cv_test[i] in evaluate:
sum_u65 += u65(len(evaluate))
sum_u80 += u80(len(evaluate))
ell_u65[ell_current] += sum_u65 / n_test
ell_u80[ell_current] += sum_u80 / n_test
print("-------ELL_CURRENT-----", ell_current)
ell_u65[ell_current] = ell_u65[ell_current] / cv_n_fold
ell_u80[ell_current] = ell_u80[ell_current] / cv_n_fold
print("u65-->", ell_u65[ell_current])
print("u80-->", ell_u80[ell_current])
print("--->", ell_u65, ell_u80)
def computing_time_prediction(in_path=None):
import time
data = export_data_set('iris.data') if in_path is None else pd.read_csv(in_path)
print("-----DATA SET TRAINING---", in_path)
X = data.iloc[:, :-1].values
y = data.iloc[:, -1].tolist()
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.4, random_state=0)
lqa = LinearDiscriminant(init_matlab=True)
lqa.learn(X_train, y_train, ell=0.01)
sum_time = 0
n, _ = X_test.shape
for i, test in enumerate(X_test):
start = time.time()
evaluate, _ = lqa.evaluate(test)
end = time.time()
print(evaluate, "-----", y_test[i], '--time--', (end - start))
sum_time += (end - start)
print("--->", sum_time, '---n---', n)
def computing_precise_vs_imprecise(in_path=None, ell_optimal=0.1, seeds=0):
def u65(mod_Y):
return 1.6 / mod_Y - 0.6 / mod_Y**2
def u80(mod_Y):
return 2.2 / mod_Y - 1.2 / mod_Y**2
data = export_data_set('iris.data') if in_path is None else pd.read_csv(
in_path)
print("-----DATA SET TRAINING---", in_path)
X = data.iloc[:, :-1].values
y = data.iloc[:, -1].tolist()
n_time = len(seeds)
lda_imp = LinearDiscriminant(init_matlab=True)
lda = LinearDiscriminantAnalysis(solver="svd", store_covariance=True)
mean_u65_imp, mean_u80_imp, u_mean = 0, 0, 0
for k in range(0, n_time):
X_train, X_test, y_train, y_test = \
train_test_split(X, y, test_size=0.4, random_state=seeds[k])
lda_imp.learn(X_train, y_train, ell=ell_optimal)
lda.fit(X_train, y_train)
sum_u65, sum_u80 = 0, 0
u_precise, n_real_test = 0, 0
n_test, _ = X_test.shape
for i, test in enumerate(X_test):
print("--TESTING-----", i)
evaluate_imp, _ = lda_imp.evaluate(test)
if len(evaluate_imp) > 1:
n_real_test += 1
if y_test[i] in evaluate_imp:
sum_u65 += u65(len(evaluate_imp))
sum_u80 += u80(len(evaluate_imp))
evaluate = lda.predict([test])
if y_test[i] in evaluate:
u_precise += u80(len(evaluate))
mean_u65_imp += sum_u65 / n_real_test
mean_u80_imp += sum_u80 / n_real_test
u_mean += u_precise / n_real_test
print("--time_k--u65-->", k, sum_u65 / n_real_test)
print("--time_k--u80-->", k, sum_u80 / n_real_test)
print("--time_k--precise-->", k, u_precise / n_real_test)
print("--global--u65-->", mean_u65_imp / n_time)
print("--global--u80-->", mean_u80_imp / n_time)
print("--global--precise-->", u_mean / n_time)
def output_paper_result(in_path=None):
data = export_data_set('bin_normal_rnd.data') if in_path is None else pd.read_csv(in_path)
X = data.loc[:, ['x1', 'x2']].values
y = data.y.tolist()
lqa = LinearDiscriminant(init_matlab=True)
lqa.learn(X, y, ell=2)
lqa.plot2D_classification(np.array([2, 2]), colors={0: 'red', 1: 'blue'})
lqa.plot2D_decision_boundary()
def computing_precise_vs_imprecise(in_path=None, ell_optimal=0.1, seeds=0):
def u65(mod_Y):
return 1.6 / mod_Y - 0.6 / mod_Y ** 2
def u80(mod_Y):
return 2.2 / mod_Y - 1.2 / mod_Y ** 2
data = export_data_set('iris.data') if in_path is None else pd.read_csv(in_path)
print("-----DATA SET TRAINING---", in_path)
X = data.iloc[:, :-1].values
y = data.iloc[:, -1].tolist()
n_time = len(seeds)
lda_imp = LinearDiscriminant(init_matlab=True)
lda = LinearDiscriminantAnalysis(solver="svd", store_covariance=True)
mean_u65_imp, mean_u80_imp, u_mean = 0, 0, 0
for k in range(0, n_time):
X_train, X_test, y_train, y_test = \
train_test_split(X, y, test_size=0.4, random_state=seeds[k])
lda_imp.learn(X_train, y_train, ell=ell_optimal)
lda.fit(X_train, y_train)
sum_u65, sum_u80 = 0, 0
u_precise, n_real_test = 0, 0
n_test, _ = X_test.shape
for i, test in enumerate(X_test):
print("--TESTING-----", i)
evaluate_imp, _ = lda_imp.evaluate(test)
if len(evaluate_imp) > 1:
n_real_test += 1
if y_test[i] in evaluate_imp:
sum_u65 += u65(len(evaluate_imp))
sum_u80 += u80(len(evaluate_imp))
evaluate = lda.predict([test])
if y_test[i] in evaluate:
u_precise += u80(len(evaluate))
mean_u65_imp += sum_u65 / n_real_test
mean_u80_imp += sum_u80 / n_real_test
u_mean += u_precise / n_real_test
print("--time_k--u65-->", k, sum_u65 / n_real_test)
print("--time_k--u80-->", k, sum_u80 / n_real_test)
print("--time_k--precise-->", k, u_precise / n_real_test)
print("--global--u65-->", mean_u65_imp / n_time)
print("--global--u80-->", mean_u80_imp / n_time)
print("--global--precise-->", u_mean / n_time)
def output_paper_zone_imprecise():
data = export_data_set('iris.data')
X = data.iloc[:, 0:2].values
y = data.iloc[:, -1].tolist()
lqa = LinearDiscriminant(init_matlab=True)
lqa.learn(X, y, ell=5)
query = np.array([5.0, 2])
answer, _ = lqa.evaluate(query)
# lqa.plot2D_classification(query)
lqa.plot2D_decision_boundary()
def computing_time_prediction(in_path=None):
import time
data = export_data_set('iris.data') if in_path is None else pd.read_csv(
in_path)
print("-----DATA SET TRAINING---", in_path)
X = data.iloc[:, :-1].values
y = data.iloc[:, -1].tolist()
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.4,
random_state=0)
lqa = LinearDiscriminant(init_matlab=True)
lqa.learn(X_train, y_train, ell=0.01)
sum_time = 0
n, _ = X_test.shape
for i, test in enumerate(X_test):
start = time.time()
evaluate, _ = lqa.evaluate(test)
end = time.time()
print(evaluate, "-----", y_test[i], '--time--', (end - start))
sum_time += (end - start)
print("--->", sum_time, '---n---', n)
def computing_accuracy_imprecise(in_path=None,
seeds=list([0]),
ell_optimal=0.1):
def u65(mod_Y):
return 1.6 / mod_Y - 0.6 / mod_Y**2
def u80(mod_Y):
return 2.2 / mod_Y - 1.2 / mod_Y**2
data = export_data_set('iris.data') if in_path is None else pd.read_csv(
in_path)
print("-----DATA SET TRAINING---", in_path)
X = data.iloc[:, :-1].values
y = data.iloc[:, -1].tolist()
n_time = len(seeds)
mean_u65 = 0
mean_u80 = 0
lqa = LinearDiscriminant(init_matlab=True)
for k in range(0, n_time):
X_train, X_test, y_train, y_test = \
train_test_split(X, y, test_size=0.4, random_state=seeds[k])
lqa.learn(X_train, y_train, ell=ell_optimal)
sum_u65 = 0
sum_u80 = 0
n_test, _ = X_test.shape
for i, test in enumerate(X_test):
print("--TESTING-----", i, ell_optimal)
evaluate, _ = lqa.evaluate(test)
print(evaluate, "-----", y_test[i])
if y_test[i] in evaluate:
sum_u65 += u65(len(evaluate))
sum_u80 += u80(len(evaluate))
print("--ell_65_k_time---", k, sum_u65 / n_test)
print("--ell_u80_k_time---", k, sum_u80 / n_test)
mean_u65 += sum_u65 / n_test
mean_u80 += sum_u80 / n_test
mean_u65 = mean_u65 / n_time
mean_u80 = mean_u80 / n_time
print("--ell-->", ell_optimal, "--->", mean_u65, mean_u80)
def computing_accuracy_imprecise(in_path=None, seeds=list([0]), ell_optimal=0.1):
def u65(mod_Y):
return 1.6 / mod_Y - 0.6 / mod_Y ** 2
def u80(mod_Y):
return 2.2 / mod_Y - 1.2 / mod_Y ** 2
data = export_data_set('iris.data') if in_path is None else pd.read_csv(in_path)
print("-----DATA SET TRAINING---", in_path)
X = data.iloc[:, :-1].values
y = data.iloc[:, -1].tolist()
n_time = len(seeds)
mean_u65 = 0
mean_u80 = 0
lqa = LinearDiscriminant(init_matlab=True)
for k in range(0, n_time):
X_train, X_test, y_train, y_test = \
train_test_split(X, y, test_size=0.4, random_state=seeds[k])
lqa.learn(X_train, y_train, ell=ell_optimal)
sum_u65 = 0
sum_u80 = 0
n_test, _ = X_test.shape
for i, test in enumerate(X_test):
print("--TESTING-----", i, ell_optimal)
evaluate, _ = lqa.evaluate(test)
print(evaluate, "-----", y_test[i])
if y_test[i] in evaluate:
sum_u65 += u65(len(evaluate))
sum_u80 += u80(len(evaluate))
print("--ell_65_k_time---", k, sum_u65 / n_test)
print("--ell_u80_k_time---", k, sum_u80 / n_test)
mean_u65 += sum_u65 / n_test
mean_u80 += sum_u80 / n_test
mean_u65 = mean_u65 / n_time
mean_u80 = mean_u80 / n_time
print("--ell-->", ell_optimal, "--->", mean_u65, mean_u80)
def output_paper_zone_imprecise():
data = export_data_set('iris.data')
X = data.iloc[:, 0:2].values
y = data.iloc[:, -1].tolist()
lqa = LinearDiscriminant(init_matlab=True)
lqa.learn(X, y, ell=5)
query = np.array([5.0, 2])
answer, _ = lqa.evaluate(query)
# lqa.plot2D_classification(query)
lqa.plot2D_decision_boundary()
def _test_ILDA(in_train=None, features=None):
ilda = LinearDiscriminant(DEBUG=True)
data = export_data_set('iris.data') if in_train is None else pd.read_csv(
in_train)
__test_imprecise_model(ilda, data, features, hgrid=0.1)
@author: Yonatan-Carlos Carranza-Alarcon
Imprecise Gaussian Discriminant
multi-classes classification
'''
from classifip.models.qda import LinearDiscriminant
from classifip.dataset.uci_data_set import export_data_set
# We start by creating an instance of the base classifier we want to use
print(
"Example of Imprecise Linear Discriminant Analyse for multi-classes - Data set IRIS \n"
)
model = LinearDiscriminant(solver_matlab=False, DEBUG=False)
data = export_data_set('iris.data')
# Learning
X = data.iloc[:, :-1].values
y = data.iloc[:, -1].tolist()
model.learn(X=X, y=y, ell=5)
# Evaluation : we can set the method for minimize convex problem with quadratic
test = model.evaluate(query=X[2], method="quadratic")
# The output is a list of probability intervals, we can print each instance :
print(
"\nPrediction using interval dominance criterion with 0/1 costs + quadratic method\n"
)
print(test)