m=m,
n=n)
print(eigenfaceModel)
y_predict = zeros(len(y_test))
for i, X_i in enumerate(X_test):
y_predict[i] = eigenfaceModel.noFitTest(X_i, 1)
from sklearn.metrics import classification_report
print(
classification_report(y_test,
y_predict,
target_names=faces95_own_classes))
if doPlot:
plotConfusionMatrix(y_test, y_predict, labels=faces95_own_classes)
# ---------------------- OTHER ALGORITHMS -----------------------------
from sklearn.neural_network import MLPClassifier
X_train_pca = eigenfaceModel.exportTrainPca()
X_test_pca = []
for X_i in X_test:
X_test_pca.append(eigenfaceModel.projectOntoFacespace(X_i))
X_test_pca = array(X_test_pca)
# train a neural network
print("Fitting the classifier to the training set")
clf = MLPClassifier(hidden_layer_sizes=(
1024,
w_predict = zeros(len(y_test)) # Array to fill with predictions
epsilons = zeros(len(y_test))
for i, X_i in enumerate(X_test):
y_pred_i, min_err_i = eigenfaceModel._noFitTest(X_i)
epsilons[i] = min_err_i
# Method 1: Fixed threshold = 2000 (number deducted from the plots in EDA submodule)
w_predict[epsilons <= 250] = 1
print(
classification_report(w_true,
w_predict,
target_names=['Known', 'Unknown']))
plotConfusionMatrix(w_true, w_predict, labels=['Known', 'Unknown'])
# Method 2: Threshold calculated as mean + 2sd from the test1 split taken from the known dataset
sub_eps = epsilons[:len(y_test1)]
threshold = np.mean(sub_eps) + 2 * np.std(
sub_eps, ddof=1, dtype=np.float64)
w_predict2 = zeros(len(y_test))
w_predict2[epsilons <= threshold] = 1
print(
classification_report(w_true,
w_predict2,
target_names=['Known', 'Unknown']))
plotConfusionMatrix(w_true, w_predict2, labels=['Known', 'Unknown'])
# Method 3: Threshold calculated as mean + 3*sd from the test1 split taken from the known dataset
sub_eps = epsilons[:len(y_test1)]
#X_test = X_test[y_test < output100_all_classes.index('io')]
#y_test = y_test[y_test < output100_all_classes.index('io')]
eigenfaceModel = EigenfaceModel(X_train, y_train, nmin=2, nmax=150)
print(eigenfaceModel)
y_predict2 = zeros(len(y_test))
for i, X_i in enumerate(X_test):
y_predict2[i] = eigenfaceModel.noFitTest(X_i, 1)
from sklearn.metrics import classification_report
import matplotlib.pyplot as plt
print(classification_report(y_test, y_predict2, target_names=labels))
plotConfusionMatrix(y_test, y_predict2, labels=labels)
# ---------------------- ANGLE ----------------------------------------
y_predict = zeros(len(y_test))
for i, X_i in enumerate(X_test):
y_predict[i] = eigenfaceModel.angleTest(X_i, 1)
from sklearn.metrics import classification_report
import matplotlib.pyplot as plt
print(classification_report(y_test, y_predict, target_names=labels))
plotConfusionMatrix(y_test, y_predict, labels=labels)
# ---------------------- OTHER ALGORITHMS -----------------------------
eigenfaceModel = EigenfaceModel(X_train, y_train, nmin=2, nmax=150)
print(eigenfaceModel)
y_predict = zeros(len(y_test))
for i, X_i in enumerate(X_test):
y_predict[i] = eigenfaceModel.noFitTest(X_i, 1)
from sklearn.metrics import classification_report
import matplotlib.pyplot as plt
print(
classification_report(y_test,
y_predict,
target_names=output100strict_classes))
plotConfusionMatrix(y_test, y_predict, labels=output100strict_classes)
# ---------------------- ANGLE ----------------------------------------
y_predict = zeros(len(y_test))
for i, X_i in enumerate(X_test):
y_predict[i] = eigenfaceModel.angleTest(X_i, 1)
from sklearn.metrics import classification_report
import matplotlib.pyplot as plt
print(
classification_report(y_test,
y_predict,
target_names=output100strict_classes))