머신 러닝 - 비지도 학습(주성분 분석) (11)

fruits_2d = fruits.reshape(-1, 100*100)

from sklearn.decomposition import PCA

pca = PCA(n_components=50)
pca.fit(fruits_2d)

print(pca.components_) # (50, 10000)

def draw_fruits(arr, ratio=1):
  n = len(arr)
  rows = int(np.ceil(n/10))
  cols = n if rows < 2 else 10
  
  fig, axs = plt.subplots(rows, cols, figsize = (cols*ratio, rows*ratio), squeeze=False)
  
  for i in range(rows):
    for j in range(cols):
      if i * 10 + j < n:
        axs[i, j].imshow(arr[i * 10 + j], cmap='gray_r')
      axs[i, j].axis('off')
  plt.show()

draw_fruits(pca.components_.reshape(-1, 100, 100))

fruits_pca = pca.transform(fruits_2d) # (300, 50)

fruits_inverse = pca.inverse_transform(fruits_pca) # (300, 10000)

fruits_reconstruct = fruits_inverse.reshape(-1, 100, 100)
 
for start in [0, 100, 200]:
  draw_fruits(fruits_reconstruct[start:start + 100])
  print('\n')

pca.explained_variance_ratio_
# array([0.42357017, 0.09941755, 0.06577863, 0.04031172, 0.03416875,
#      0.03281329, 0.02573267, 0.02054963, 0.01372276, 0.01342773,
#      0.01152146, 0.00944596, 0.00878232, 0.00846697, 0.00693049,
#      0.00645188, 0.00578896, 0.00511202, 0.00486383, 0.00480346,
#      0.00447836, 0.00437318, 0.0040804 , 0.00389474, 0.00372443,
#      0.00359285, 0.00331465, 0.00317839, 0.00304314, 0.00303756,
#      0.0028891 , 0.00275745, 0.002648  , 0.00255949, 0.00252005,
#      0.00247363, 0.00239597, 0.00230914, 0.00221862, 0.00216902,
#      0.0021394 , 0.00194782, 0.00193262, 0.00191025, 0.00184647,
#      0.00182962, 0.00173503, 0.00168848, 0.00163037, 0.00160394])

print(np.sum(pca.explained_variance_ratio_)) # 0.9215824394667924

plt.plot(pca.explained_variance_ratio_)
plt.show()

from sklearn.linear_model import LogisticRegression
 
lr = LogisticRegression()

target = np.array([0]*100 + [1]*100 + [2]*100)

from sklearn.model_selection import cross_validate
 
scores = cross_validate(lr, fruits_2d, target)
print(np.mean(scores['test_score'])) # 0.99...
print(np.mean(scores['fit_time'])) # 0.25...

scores = cross_validate(lr, fruits_pca, target)
print(np.mean(scores['test_score'])) # 0.99...
print(np.mean(scores['fit_time'])) # 0.02...