🖼️Praktikum 2

Klasifikasi SVM Pada Data Citra

dataset yang digunakan adalah Labeled Faces in the Wild dataset (sebuah dataset yang berisi ribuan wajah publik figur).

Link dataset: https://www.kaggle.com/datasets/jessicali9530/lfw-dataset

from sklearn.datasets import fetch_lfw_people
faces = fetch_lfw_people(min_faces_per_person=60)
print(faces.target_names)
print(len(faces.target_names))
print(faces.images.shape)

['Ariel Sharon' 'Colin Powell' 'Donald Rumsfeld' 'George W Bush'
 'Gerhard Schroeder' 'Hugo Chavez' 'Junichiro Koizumi' 'Tony Blair']
8
(1348, 62, 47)

# contoh wajah yang digunakan

fig, ax = plt.subplots(3, 5)
for i, axi in enumerate(ax.flat):
    axi.imshow(faces.images[i], cmap='bone')
    axi.set(xticks=[], yticks=[],
            xlabel=faces.target_names[faces.target[i]])

from sklearn.svm import SVC
from sklearn.decomposition import PCA as RandomizedPCA
from sklearn.pipeline import make_pipeline

pca = RandomizedPCA(n_components=150, whiten=True, random_state=42)
svc = SVC(kernel='rbf', class_weight='balanced')
model = make_pipeline(pca, svc)

# pemisahan data training dan data testing

from sklearn.model_selection import train_test_split
Xtrain, Xtest, ytrain, ytest = train_test_split(faces.data, faces.target,
                                                random_state=42)

from sklearn.model_selection import GridSearchCV
param_grid = {'svc__C': [1, 5, 10, 50],
              'svc__gamma': [0.0001, 0.0005, 0.001, 0.005]}
grid = GridSearchCV(model, param_grid)

%time grid.fit(Xtrain, ytrain)
print(grid.best_params_)
print(grid.best_score_)

CPU times: user 1min 6s, sys: 1.29 s, total: 1min 7s
Wall time: 38.4 s
{'svc__C': 50, 'svc__gamma': 0.005}
0.7448080768668

prediksi label untuk data testing

model = grid.best_estimator_
yfit = model.predict(Xtest)

# hasil label pada data testing

fig, ax = plt.subplots(4, 6)
for i, axi in enumerate(ax.flat):
    axi.imshow(Xtest[i].reshape(62, 47), cmap='bone')
    axi.set(xticks=[], yticks=[])
    axi.set_ylabel(faces.target_names[yfit[i]].split()[-1],
                   color='black' if yfit[i] == ytest[i] else 'red')
fig.suptitle('Predicted Names; Incorrect Labels in Red', size=14)

Text(0.5, 0.98, 'Predicted Names; Incorrect Labels in Red')

contoh di atas hanya menunjukkan satu data dengan label salah. Secara pengukuran, performa klasifikasi dapat diketahui sebagai berikut

from sklearn.metrics import classification_report
print(classification_report(ytest, yfit,
                            target_names=faces.target_names))

                   precision    recall  f1-score   support

     Ariel Sharon       0.91      0.67      0.77        15
     Colin Powell       0.85      0.85      0.85        68
  Donald Rumsfeld       0.72      0.58      0.64        31
    George W Bush       0.76      0.92      0.83       126
Gerhard Schroeder       0.75      0.65      0.70        23
      Hugo Chavez       1.00      0.45      0.62        20
Junichiro Koizumi       0.91      0.83      0.87        12
       Tony Blair       0.80      0.76      0.78        42

         accuracy                           0.80       337
        macro avg       0.84      0.71      0.76       337
     weighted avg       0.80      0.80      0.79       337

# bentuk confusion matrix

from sklearn.metrics import confusion_matrix
mat = confusion_matrix(ytest, yfit)
sns.heatmap(mat.T, square=True, annot=True, fmt='d', cbar=False,
            xticklabels=faces.target_names,
            yticklabels=faces.target_names)
plt.xlabel('true label')
plt.ylabel('predicted label')

Text(91.68, 0.5, 'predicted label')