dimension reduction techniques

dimension_reduction/dimension_red.py

@@ -0,0 +1,50 @@
+#%%
+from sklearn.datasets import fetch_mldata
+import numpy as np
+mnist = fetch_mldata('MNIST original')
+print(mnist)
+
+#%%
+X, Y = mnist["data"], mnist["target"]
+print(X)
+print(Y)
+print(X.shape)
+print(Y.shape)
+
+#%%
+
+from sklearn.decomposition import PCA
+
+pca = PCA(n_components=2)
+X_2d = pca.fit(X)
+print(X_2d)
+
+#%%
+'''choosing n_components effieciently'''
+pca = PCA()
+pca.fit(X)
+cumsum = np.cumsum(pca.explained_variance_ratio_)
+d = np.argmax(cumsum >= 0.95) + 1
+print(d)
+
+#%%
+pca = PCA(n_components=0.95)
+X_reduced = pca.fit_transform(X)
+print(X_reduced.shape)
+
+#%%
+pca = PCA(n_components = 154)
+X_mnist_reduced = pca.fit_transform(X)
+X_mnist_recovered = pca.inverse_transform(X_mnist_reduced)
+
+#%%
+from sklearn.decomposition import IncrementalPCA
+n_batches = 100
+inc_pca = IncrementalPCA(n_components=154)
+for X_batch in np.array_split(X, n_batches):
+    inc_pca.partial_fit(X_batch)
+X_mnist_reduced = inc_pca.transform(X)
+
+#%%
+rnd_pca = PCA(n_components=154, svd_solver="randomized")
+X_reduced = rnd_pca.fit_transform(X)