tSNE
tSNE is the t-distributed stochastic neighborhood embedding
- PCA works by preserving the variability of the data; tSENE works by non-linear probility
- PCA focusses on global structure rather than local, tSNE focuses on local
- tSNE is computationally more expensive than PCA
- PCA intuition is simpler and has fewer parameters than tSNE, as a result, has wider applicability
-
the result would not be unique and you’ll get different results on every run
from sklearn.manifold import TSNE tsne = TSNE(n_components=2,perplexity=40, random_state=42) X_reduced_tsne = tsne.fit_transform(X) plt.scatter(X_reduced_tsne[:,0], X_reduced_tsne[:,1],c=vectorizer(y))
visualize feature vectors
# Create a two dimensional t-SNE projection of the embeddings tsne = TSNE(2, verbose=1) tsne_proj = tsne.fit_transform(test_embeddings) # Plot those points as a scatter plot and label them based on the pred labels cmap = cm.get_cmap(‘tab20’) fig, ax = plt.subplots(figsize=(8,8)) num_categories = 10 for lab in range(num_categories): indices = test_predictions==lab ax.scatter(tsne_proj[indices,0],tsne_proj[indices,1], c=np.array(cmap(lab)).reshape(1,4), label = lab ,alpha=0.5) ax.legend(fontsize=’large’, markerscale=2) plt.show()
# three dimensional t-sne projection of the embeddings tsne = TSNE(3, verbose=1) tsne_proj = tsne.fit_transform(test_embeddings) cmap = cm.get_cmap(‘tab20’) num_categories = 10 for lab in range(num_categories): indices = test_predictions == lab ax.scatter(tsne_proj[indices, 0], tsne_proj[indices, 1], tsne_proj[indices, 2], c=np.array(cmap(lab)).reshape(1, 4), label=lab, alpha=0.5) ax.legend(fontsize=’large’, markerscale=2) plt.show()
PCA
used to simplify a large and complex dataset into a smaller, more easily understandable dataset.
pca = PCA(n_components=2)
pca.fit(test_embeddings)
pca_proj = pca.transform(test_embeddings)
pca.explained_variance_ratio_
# explained_variance_ratio_ indicate the amount of variance from your data that the principal components capture, the higher the better
# array([0.16178058, 0.08613876], dtype=float32)
Related Issues not found
Please contact @neverset123 to initialize the comment
