def case_1(c1_test, c2_test, c3_test, p_c1, p_c2, p_c3, cov_matrix1,
cov_matrix2, cov_matrix3, mean1, mean2, mean3, x1, x2, x3, y1, y2,
y3, which_plot, c_min_max):
cov_matrix = np.zeros((2, 2))
covv_matrix = np.zeros((2, 2))
inv_cov_matrix = np.zeros((2, 2))
inv_cov_matrix1 = np.zeros((2, 2))
inv_cov_matrix2 = np.zeros((2, 2))
inv_cov_matrix3 = np.zeros((2, 2))
confusion_matrix = np.zeros((3, 3))
for i in range(2):
for j in range(2):
if (i + j) % 2 == 0:
covv_matrix[i][j] = (cov_matrix1[i][j] + cov_matrix2[i][j] +
cov_matrix3[i][j]) / 3
else:
cov_matrix[i][j] = 0
cov_matrix[0][0] = (covv_matrix[0][0] + covv_matrix[1][1]) / 2.0000
cov_matrix[1][1] = (covv_matrix[0][0] + covv_matrix[1][1]) / 2.0000
inv_cov_matrix = np.linalg.inv(cov_matrix)
inv_cov_matrix1 = inv_cov_matrix2 = inv_cov_matrix3 = inv_cov_matrix
#confusion_matrix=test.test_1(c1_test,c2_test,c3_test,inv_cov_matrix1,inv_cov_matrix2,inv_cov_matrix3,mean1,mean2,mean3)
if which_plot == 0:
confusion_matrix = test.test_1(c1_test, c2_test, c3_test, p_c1, p_c2,
p_c3, inv_cov_matrix1, inv_cov_matrix2,
inv_cov_matrix3, mean1, mean2, mean3)
plot.plot_1(p_c1, p_c2, p_c3, inv_cov_matrix1, inv_cov_matrix2,
inv_cov_matrix3, mean1, mean2, mean3, x1, x2, x3, y1, y2,
y3, 0, c_min_max)
return (confusion_matrix)
elif which_plot == 10:
plot.plot_1(p_c1, p_c2, p_c3, inv_cov_matrix1, inv_cov_matrix2,
inv_cov_matrix3, mean1, mean2, mean3, x1, x2, x3, y1, y2,
y3, 10, c_min_max)
elif which_plot == -1:
plot.plot_1_1(p_c1, p_c2, inv_cov_matrix1, inv_cov_matrix2, mean1,
mean2, x1, x2, y1, y2, -1, c_min_max)
elif which_plot == 1:
plot.plot_1_1(p_c1, p_c2, inv_cov_matrix1, inv_cov_matrix2, mean1,
mean2, x1, x2, y1, y2, 1, c_min_max)
elif which_plot == -2:
plot.plot_1_1(p_c2, p_c3, inv_cov_matrix2, inv_cov_matrix3, mean2,
mean3, x2, x3, y2, y3, -2, c_min_max)
elif which_plot == 2:
plot.plot_1_1(p_c2, p_c3, inv_cov_matrix2, inv_cov_matrix3, mean2,
mean3, x2, x3, y2, y3, 2, c_min_max)
elif which_plot == -3:
plot.plot_1_1(p_c1, p_c3, inv_cov_matrix1, inv_cov_matrix3, mean1,
mean3, x1, x3, y1, y3, -3, c_min_max)
elif which_plot == 3:
plot.plot_1_1(p_c1, p_c3, inv_cov_matrix1, inv_cov_matrix3, mean1,
mean3, x1, x3, y1, y3, 3, c_min_max)
def case_4(c1_test, c2_test, c3_test, p_c1, p_c2, p_c3, cov_matrix1,
cov_matrix2, cov_matrix3, mean1, mean2, mean3, x1, x2, x3, y1, y2,
y3, which_plot, c_min_max):
cov_matrix = np.zeros((2, 2))
inv_cov_matrix1 = np.zeros((2, 2))
inv_cov_matrix2 = np.zeros((2, 2))
inv_cov_matrix3 = np.zeros((2, 2))
confusion_matrix = np.zeros((3, 3))
inv_cov_matrix1 = np.linalg.inv(cov_matrix1)
inv_cov_matrix2 = np.linalg.inv(cov_matrix2)
inv_cov_matrix3 = np.linalg.inv(cov_matrix3)
#confusion_matrix=test.test_1(c1_test,c2_test,c3_test,inv_cov_matrix1,inv_cov_matrix2,inv_cov_matrix3,mean1,mean2,mean3)
if which_plot == 0:
confusion_matrix = test.test_1(c1_test, c2_test, c3_test, p_c1, p_c2,
p_c3, inv_cov_matrix1, inv_cov_matrix2,
inv_cov_matrix3, mean1, mean2, mean3)
plot.plot_1(p_c1, p_c2, p_c3, inv_cov_matrix1, inv_cov_matrix2,
inv_cov_matrix3, mean1, mean2, mean3, x1, x2, x3, y1, y2,
y3, 0, c_min_max)
return (confusion_matrix)
elif which_plot == 10:
plot.plot_1(p_c1, p_c2, p_c3, inv_cov_matrix1, inv_cov_matrix2,
inv_cov_matrix3, mean1, mean2, mean3, x1, x2, x3, y1, y2,
y3, 10, c_min_max)
elif which_plot == -1:
plot.plot_1_1(p_c1, p_c2, inv_cov_matrix1, inv_cov_matrix2, mean1,
mean2, x1, x2, y1, y2, -1, c_min_max)
elif which_plot == 1:
plot.plot_1_1(p_c1, p_c2, inv_cov_matrix1, inv_cov_matrix2, mean1,
mean2, x1, x2, y1, y2, 1, c_min_max)
elif which_plot == -2:
plot.plot_1_1(p_c2, p_c3, inv_cov_matrix2, inv_cov_matrix3, mean2,
mean3, x2, x3, y2, y3, -2, c_min_max)
elif which_plot == 2:
plot.plot_1_1(p_c2, p_c3, inv_cov_matrix2, inv_cov_matrix3, mean2,
mean3, x2, x3, y2, y3, 2, c_min_max)
elif which_plot == -3:
plot.plot_1_1(p_c1, p_c3, inv_cov_matrix1, inv_cov_matrix3, mean1,
mean3, x1, x3, y1, y3, -3, c_min_max)
elif which_plot == 3:
plot.plot_1_1(p_c1, p_c3, inv_cov_matrix1, inv_cov_matrix3, mean1,
mean3, x1, x3, y1, y3, 3, c_min_max)
def case_1(c1, c2, c3, cov_matrix1, cov_matrix2, cov_matrix3, mean1, mean2,
mean3, x1, x2, x3, y1, y2, y3):
cov_matrix = np.zeros((2, 2))
inv_cov_matrix = np.zeros((2, 2))
confusion_matrix = np.zeros((3, 3))
for i in range(2):
for j in range(2):
if (i + j) % 2 == 0:
cov_matrix[i][j] = (cov_matrix1[i][j] + cov_matrix2[i][j] +
cov_matrix3[i][j]) / 3
else:
cov_matrix[i][j] = 0
inv_cov_matrix = np.linalg.inv(cov_matrix)
confusion_matrix = test.test_1(c1, c2, c3, inv_cov_matrix, mean1, mean2,
mean3)
plot.plot_1(inv_cov_matrix, mean1, mean2, mean3, x1, x2, x3, y1, y2, y3)
return (confusion_matrix)
def case_3(c1_test, c2_test, c3_test, p_c1, p_c2, p_c3, cov_matrix1,
cov_matrix2, cov_matrix3, mean1, mean2, mean3, x1, x2, x3, y1, y2,
y3, which_plot, c_min_max):
cov_matrix = np.zeros((2, 2))
cov_matrix1_1 = np.zeros((2, 2))
cov_matrix2_2 = np.zeros((2, 2))
cov_matrix3_3 = np.zeros((2, 2))
cov_matrix1_1[0][0] = cov_matrix1[0][0]
cov_matrix2_2[0][0] = cov_matrix2[0][0]
cov_matrix3_3[0][0] = cov_matrix3[0][0]
cov_matrix1_1[1][1] = cov_matrix1[1][1]
cov_matrix2_2[1][1] = cov_matrix2[1][1]
cov_matrix3_3[1][1] = cov_matrix3[1][1]
inv_cov_matrix1 = np.zeros((2, 2))
inv_cov_matrix2 = np.zeros((2, 2))
inv_cov_matrix3 = np.zeros((2, 2))
confusion_matrix = np.zeros((3, 3))
'''
for i in range(2):
for j in range(2):
if(i+j)%2!=0:
cov_matrix1_1[i][j]=0;
cov_matrix2_2[i][j]=0;
cov_matrix3_3[i][j]=0;
'''
inv_cov_matrix1 = np.linalg.inv(cov_matrix1_1)
inv_cov_matrix2 = np.linalg.inv(cov_matrix2_2)
inv_cov_matrix3 = np.linalg.inv(cov_matrix3_3)
#confusion_matrix=test.test_1(c1_test,c2_test,c3_test,inv_cov_matrix1,inv_cov_matrix2,inv_cov_matrix3,mean1,mean2,mean3)
if which_plot == 0:
confusion_matrix = test.test_1(c1_test, c2_test, c3_test, p_c1, p_c2,
p_c3, inv_cov_matrix1, inv_cov_matrix2,
inv_cov_matrix3, mean1, mean2, mean3)
plot.plot_1(p_c1, p_c2, p_c3, inv_cov_matrix1, inv_cov_matrix2,
inv_cov_matrix3, mean1, mean2, mean3, x1, x2, x3, y1, y2,
y3, 0, c_min_max)
return (confusion_matrix)
elif which_plot == 10:
plot.plot_1(p_c1, p_c2, p_c3, inv_cov_matrix1, inv_cov_matrix2,
inv_cov_matrix3, mean1, mean2, mean3, x1, x2, x3, y1, y2,
y3, 10, c_min_max)
elif which_plot == -1:
plot.plot_1_1(p_c1, p_c2, inv_cov_matrix1, inv_cov_matrix2, mean1,
mean2, x1, x2, y1, y2, -1, c_min_max)
elif which_plot == 1:
plot.plot_1_1(p_c1, p_c2, inv_cov_matrix1, inv_cov_matrix2, mean1,
mean2, x1, x2, y1, y2, 1, c_min_max)
elif which_plot == -2:
plot.plot_1_1(p_c2, p_c3, inv_cov_matrix2, inv_cov_matrix3, mean2,
mean3, x2, x3, y2, y3, -2, c_min_max)
elif which_plot == 2:
plot.plot_1_1(p_c2, p_c3, inv_cov_matrix2, inv_cov_matrix3, mean2,
mean3, x2, x3, y2, y3, 2, c_min_max)
elif which_plot == -3:
plot.plot_1_1(p_c1, p_c3, inv_cov_matrix1, inv_cov_matrix3, mean1,
mean3, x1, x3, y1, y3, -3, c_min_max)
elif which_plot == 3:
plot.plot_1_1(p_c1, p_c3, inv_cov_matrix1, inv_cov_matrix3, mean1,
mean3, x1, x3, y1, y3, 3, c_min_max)
one_to_one=OnetoOne(0, 0)
full_search=FullSearch(0, 0)
piles =[]
read_data(piles)
one_to_one.setAB(piles[0], piles[1])
full_search.setAB(piles[0], piles[1])
start_one = timer()
cost_o = one_to_one.simulation()
t_one = timer() - start_one
start_one = timer()
cost_f = full_search.simulation()
t_full = timer() - start_one
if cost_f == cost_o :
one_result=one_to_one.get_result()
full_result=full_search.get_result()
print("\nCosts for both solutions are the same and equal: {}".format(cost_f) )
print('\nSolution one_to_one A:{0} B:{1} \nExecution time: {2:0.3g} seconds'.format(one_result[0],one_result[1], t_one))
print('\nSolution full_search A:{0} B:{1} \nExecution time: {2:0.3g} seconds'.format(full_result[0],full_result[1], t_full))
else: print('The cost are not the same.')
test_1(OnetoOne(0, 0), FullSearch(0, 0))
test_2(OnetoOne(0, 0), FullSearch(0, 0))
"""
Create a Test object and run tests from it.
@author: Fang
"""
import test
if __name__ == "__main__":
test = test.Test()
test.test_1()
test.test_2()
epochs = 25
batch_size = 70
validation_split = 0.1
X, y, X_final = dataread()
X_train, X_test, y_train, y_test, prop_HF = datatreat_A1(
X,
y,
train_size=0.8,
Shuffle=True,
preprocess="None",
ratio="50/50",
balancing_method="SMOTEENN")
model = cnn_1()
history = model.fit(X_train,
y_train,
epochs=epochs,
batch_size=batch_size,
validation_split=validation_split)
save_model(model, id)
accuracy, roc, f1_macro, f1_wei = test_1(model, X_test, y_test, id)
save_results(id, 'datatreat_A1', 'cnn_1', epochs, batch_size, accuracy, roc,
f1_macro, f1_wei, validation_split)
plot_loss_acc_history(history, id, validation_split)