Sklearn-PCA dimensionality reduction and model preservation

Use the fit_transform function to learn on the training set

# 已知训练集train、验证集dev和测试集test
from sklearn.decomposition import PCA 
pca = PCA(n_components=2)	# 特征个数
# 对训练集使用fit_transform函数进行学习
X_train = pca.fit_transform(X_train)

Use the transform function for dev and test to test

# 对dev、test使用transform函数，进行测试
X_dev = pca.transform(X_dev)

Use joblib to save and load models

# 保存模型--python 3.7
import joblib
joblib.dump(pca_model, pca_model_dir)

# 加载模型
pca_model = joblib.load(pca_model_dir)
X_test = pca_model.transform(X_test)
...