pb2Bia

"""
Use the digits dataset to perform a classification task using Linear Support Vector Machine(SVM).
Set the parameter C to 5.
Split the data such the test split is 25% of the initial dataset.
Compare the image classification task using the original number of features (64) to the reduced features (20) using the PCA algorithm.
Compare:
- the training time for both models
- the difference in accuracy score
"""
import time
from sklearn import datasets
from sklearn.model_selection import train_test_split
from sklearn.svm import LinearSVC
from sklearn.decomposition import PCA

digits = datasets.load_digits()
X = digits.data
y = digits.target # 0-9 asta e unidisimensional


# Split the data such the test split is 25% of the initial dataset
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=42)

# Train the model using the original number of features (64)
# daca loss nu e definit, pot sa il las gol
svm_clf1 = LinearSVC(C=5, random_state=42)

# start the timer
start_time = time.time()

# Train the model using the original number of features (64)
svm_clf1.fit(X_train, y_train)

# end the timer
end_time = time.time()

# Print the accuracy score and the time elapsed
print("1 - Accuracy score:",svm_clf1.score(X_test, y_test))
print("1 - Time elapsed: ", end_time - start_time)

# reduce the number of features to 20 using PCA
pca = PCA(n_components=20)

# Train the model using the reduced number of features (20)
X_train_reduced = pca.fit_transform(X_train)
X_test_reduced = pca.fit_transform(X_test)

svm_clf2 = LinearSVC(C=5, random_state=42)

# start the timer
start_time = time.time()

# Train the model using the reduced number of features (20)
svm_clf2.fit(X_train_reduced, y_train)
# end the timer
end_time = time.time()

print("Shape: ",y_test.shape)
# Print the accuracy score and the time elapsed
print("2 - Accuracy score:", svm_clf2.score(X_test_reduced, y_test))
print("2 - Time elapsed: ", end_time - start_time)