iterations = 50
k = 11
kernel = "linear"
knn_array = np.empty(0)
knn_array_2 = np.empty(0)
knnprob_array = np.empty(0)
wknn_array = np.empty(0)
wknnprob_array = np.empty(0)
lr_array = np.empty(0)
svm_array = np.empty(0)
test_var = 2
for j in range(iterations):
trainsets, testsets = CrossValidation(df, TrainTestdiv)
feat = ["radius_mean", "texture_mean", "perimeter_mean", "area_mean", "smoothness_mean",
"compactness_mean", "concavity_mean", "concave points_mean", "symmetry_mean",
"fractal_dimension_mean", "radius_se", "texture_se", "perimeter_se", "area_se",
"smoothness_se", "compactness_se", "concavity_se", "concave points_se", "symmetry_se",
"fractal_dimension_se", "radius_worst", "texture_worst", "perimeter_worst",
"area_worst", "smoothness_worst", "compactness_worst", "concavity_worst",
"concave points_worst", "symmetry_worst", "fractal_dimension_worst"]
wfeat = ["texture_worst", "perimeter_worst", "smoothness_worst", "concave points_worst", "symmetry_worst"]
#feat = wfeat
for i in range(len(testsets)):
train = trainsets[i]
test = testsets[i]
knn = score(train, test, feat, "knn", k, kernel)[test_var]
knn_array = np.append(knn_array, knn)
traintestdiv) # Dividing Dataframes into train, validation, and test set
test_class = test["diagnosis"]
test_class = np.array(test_class)
test_features = test[[
"radius_mean", "texture_mean", "perimeter_mean", "area_mean",
"smoothness_mean", "compactness_mean", "concavity_mean",
"concave points_mean", "symmetry_mean", "fractal_dimension_mean",
"radius_se", "texture_se", "perimeter_se", "area_se", "smoothness_se",
"compactness_se", "concavity_se", "concave points_se", "symmetry_se",
"fractal_dimension_se", "radius_worst", "texture_worst", "perimeter_worst",
"area_worst", "smoothness_worst", "compactness_worst", "concavity_worst",
"concave points_worst", "symmetry_worst", "fractal_dimension_worst"
]]
for j in range(iterations):
trains, validations = CrossValidation(trainset, crossvalidationdiv)
accuracy_list = []
for i in range(len(trains)):
train = trains[i]
validation = validations[i]
# Specifiying Features
train_features = np.array(train[[
"radius_mean", "texture_mean", "perimeter_mean", "area_mean",
"smoothness_mean", "compactness_mean", "concavity_mean",
"concave points_mean", "symmetry_mean", "fractal_dimension_mean",
"radius_se", "texture_se", "perimeter_se", "area_se",
"smoothness_se", "compactness_se", "concavity_se",
"concave points_se", "symmetry_se", "fractal_dimension_se",
"radius_worst", "texture_worst", "perimeter_worst", "area_worst",
correct += 1
else:
incorrect += 1
accuracy = (correct / (correct + incorrect)) * 100
print(
"The accuracy of one iteration through a Train-Test split in the Wisconsin database is: ",
accuracy)
############################################# UNCOMMENT FOR FINDING SCORE WITH CROSSFOLD TESTING
data = pd.read_csv("data.csv")
df = pd.DataFrame(data)
# Dividing Dataframes into train and test set
trains, tests = CrossValidation(df, 20)
accuracy_list = []
for i in range(
len(trains)): # Looping through each train and test set together
train = trains[i]
test = tests[i]
# Specifiying Features
train_features = np.array(train[[
"radius_mean", "texture_mean", "perimeter_mean", "area_mean",
"smoothness_mean", "compactness_mean", "concavity_mean",
"concave points_mean", "symmetry_mean", "fractal_dimension_mean",
"radius_se", "texture_se", "perimeter_se", "area_se", "smoothness_se",
"compactness_se", "concavity_se", "concave points_se", "symmetry_se",
"fractal_dimension_se", "radius_worst", "texture_worst",
def joint_score_prev_current_next(training, features, classes, labels,
crossvalidationdiv, iterations, k, w1, w2,
w3, method1, method2):
y1_accuracy = 0
y2_accuracy = 0
y3_accuracy = 0
y1_loss = 0
y2_loss = 0
y3_loss = 0
y1_specificity = 0
y2_specificity = 0
y3_specificity = 0
y1_sensitivity = 0
y2_sensitivity = 0
y3_sensitivity = 0
y1_accuracy_means_list = np.empty(0)
y2_accuracy_means_list = np.empty(0)
y3_accuracy_means_list = np.empty(0)
y1_loss_means_list = np.empty(0)
y2_loss_means_list = np.empty(0)
y3_loss_means_list = np.empty(0)
y1_specificity_means_list = np.empty(0)
y2_specificity_means_list = np.empty(0)
y3_specificity_means_list = np.empty(0)
y1_sensitivity_means_list = np.empty(0)
y2_sensitivity_means_list = np.empty(0)
y3_sensitivity_means_list = np.empty(0)
for i in range(iterations):
training_sets, validation_sets = CrossValidation(
training, crossvalidationdiv)
y1_accuracy_list = np.empty(0)
y2_accuracy_list = np.empty(0)
y3_accuracy_list = np.empty(0)
y1_loss_list = np.empty(0)
y2_loss_list = np.empty(0)
y3_loss_list = np.empty(0)
y1_specificity_list = np.empty(0)
y2_specificity_list = np.empty(0)
y3_specificity_list = np.empty(0)
y1_sensitivity_list = np.empty(0)
y2_sensitivity_list = np.empty(0)
y3_sensitivity_list = np.empty(0)
for j in range(len(training_sets)):
training_set = training_sets[j]
validation_set = validation_sets[j]
if (w1 != 0).all():
y1 = joint_score(training_set, validation_set, features,
classes, labels, k, w1, method1, method2)
#print(y1)
y1_accuracy_list = np.append(y1_accuracy_list, y1[1])
y1_loss_list = np.append(y1_loss_list, y1[0])
y1_sensitivity_list = np.append(y1_sensitivity_list, y1[2])
y1_specificity_list = np.append(y1_specificity_list, y1[3])
if (w2 != 0).all():
y2 = joint_score(training_set, validation_set, features,
classes, labels, k, w2, method1, method2)
#print(y2)
y2_accuracy_list = np.append(y2_accuracy_list, y2[1])
y2_loss_list = np.append(y2_loss_list, y2[0])
y2_sensitivity_list = np.append(y2_sensitivity_list, y2[2])
y2_specificity_list = np.append(y2_specificity_list, y2[3])
if (w3 != 0).all():
y3 = joint_score(training_set, validation_set, features,
classes, labels, k, w3, method1, method2)
#print(y3)
y3_accuracy_list = np.append(y3_accuracy_list, y3[1])
y3_loss_list = np.append(y3_loss_list, y3[0])
y3_sensitivity_list = np.append(y3_sensitivity_list, y3[2])
y3_specificity_list = np.append(y3_specificity_list, y3[3])
if (w1 != 0).all():
y1_accuracy_means_list = np.append(y1_accuracy_means_list,
mean(y1_accuracy_list))
y1_loss_means_list = np.append(y1_loss_means_list,
mean(y1_loss_list))
y1_sensitivity_means_list = np.append(y1_sensitivity_means_list,
mean(y1_sensitivity_list))
y1_specificity_means_list = np.append(y1_specificity_means_list,
mean(y1_specificity_list))
if (w2 != 0).all():
y2_accuracy_means_list = np.append(y2_accuracy_means_list,
mean(y2_accuracy_list))
y2_loss_means_list = np.append(y2_loss_means_list,
mean(y2_loss_list))
y2_sensitivity_means_list = np.append(y2_sensitivity_means_list,
mean(y2_sensitivity_list))
y2_specificity_means_list = np.append(y2_specificity_means_list,
mean(y2_specificity_list))
if (w3 != 0).all():
y3_accuracy_means_list = np.append(y3_accuracy_means_list,
mean(y3_accuracy_list))
y3_loss_means_list = np.append(y3_loss_means_list,
mean(y3_loss_list))
y3_sensitivity_means_list = np.append(y3_sensitivity_means_list,
mean(y3_sensitivity_list))
y3_specificity_means_list = np.append(y3_specificity_means_list,
mean(y3_specificity_list))
if (w1 != 0).all():
y1_accuracy = mean(y1_accuracy_means_list)
y1_loss = mean(y1_loss_means_list)
y1_sensitivity = mean(y1_sensitivity_means_list)
y1_specificity = mean(y1_specificity_means_list)
if (w2 != 0).all():
y2_accuracy = mean(y2_accuracy_means_list)
y2_loss = mean(y2_loss_means_list)
y2_sensitivity = mean(y2_sensitivity_means_list)
y2_specificity = mean(y2_specificity_means_list)
if (w3 != 0).all():
y3_accuracy = mean(y3_accuracy_means_list)
y3_loss = mean(y3_loss_means_list)
y3_sensitivity = mean(y3_sensitivity_means_list)
y3_specificity = mean(y3_specificity_means_list)
if (w1 == 0).all():
y1_accuracy = 0
y1_loss = 0
y1_sensitivity = 0
y1_specificity = 0
if (w2 == 0).all():
y2_accuracy = 0
y2_loss = 0
y2_sensitivity = 0
y2_specificity = 0
if (w3 == 0).all():
y3_accuracy = 0
y3_loss = 0
y3_sensitivity = 0
y3_specificity = 0
return [[y1_loss, y2_loss, y3_loss],
[y1_accuracy, y2_accuracy, y3_accuracy],
[y1_sensitivity, y2_sensitivity, y3_sensitivity],
[y1_specificity, y2_specificity, y3_specificity]]
incorrect += 1
accuracy = (correct / (correct + incorrect)) * 100
accuracy_list.append(accuracy)
accuracy_average = mean(accuracy_list)
print("The accuracy of", iterations,
"iterations through a Train-Test split in the Wisconsin database is: ",
accuracy_average)
########################################### Uncomment for finding average score with Crossfold Testing
data = pd.read_csv("data.csv")
df = pd.DataFrame(data)
trains, tests = CrossValidation(
df, 20) # Dividing Dataframes into train and test set
k = 20
accuracy_list = []
# Looping through each train and test set together
for i in range(len(trains)):
train = trains[i]
test = tests[i]
# Specifiying Features
train_features = np.array(train[[
"radius_mean", "texture_mean", "perimeter_mean", "area_mean",
"smoothness_mean", "compactness_mean", "concavity_mean",
"concave points_mean", "symmetry_mean", "fractal_dimension_mean",
"radius_se", "texture_se", "perimeter_se", "area_se", "smoothness_se",
"compactness_se", "concavity_se", "concave points_se", "symmetry_se",