NMF impara le parti delle immagini

Ora metti in pratica ciò che hai imparato su NMF per scomporre il dataset delle cifre. Ti vengono fornite di nuovo le immagini delle cifre come array 2D samples. Questa volta hai anche a disposizione una funzione show_as_image() che visualizza l'immagine codificata da qualsiasi array 1D:

def show_as_image(sample):
    bitmap = sample.reshape((13, 8))
    plt.figure()
    plt.imshow(bitmap, cmap='gray', interpolation='nearest')
    plt.colorbar()
    plt.show()

Quando hai finito, prenditi un momento per guardare i grafici e nota come NMF ha rappresentato la cifra come somma dei componenti!

Questo esercizio fa parte del corso

Apprendimento non supervisionato in Python

Visualizza il corso

Istruzioni dell'esercizio

Importa NMF da sklearn.decomposition.
Crea un'istanza di NMF chiamata model con 7 componenti. (7 è il numero di celle in un display a LED).
Applica il metodo .fit_transform() di model a samples. Assegna il risultato a features.
Per ciascun componente del modello (accessibile tramite model.components_), applica la funzione show_as_image() a quel componente all'interno del ciclo.
Assegna la riga 0 di features a digit_features.
Stampa digit_features.

Esercizio pratico interattivo

Prova a risolvere questo esercizio completando il codice di esempio.

# Import NMF
____

# Create an NMF model: model
model = ____

# Apply fit_transform to samples: features
features = ____

# Call show_as_image on each component
for component in model.components_:
    ____

# Select the 0th row of features: digit_features
digit_features = ____

# Print digit_features
print(digit_features)

Modifica ed esegui il codice

Questo esercizio fa parte del corso

Apprendimento non supervisionato in Python

IntermediárioNível de habilidade

4.8+

Inizia il corso gratis

Learn how to discover the underlying groups (or "clusters") in a dataset. By the end of this chapter, you'll be clustering companies using their stock market prices, and distinguishing different species by clustering their measurements.

Exercise 1: Unsupervised Learning Exercise 2: How many clusters?Exercise 3: Clustering 2D points Exercise 4: Inspect your clustering Exercise 5: Evaluating a clustering Exercise 6: How many clusters of grain?Exercise 7: Evaluating the grain clustering Exercise 8: Transforming features for better clusterings Exercise 9: Scaling fish data for clustering Exercise 10: Clustering the fish data Exercise 11: Clustering stocks using KMeans Exercise 12: Which stocks move together?

In this chapter, you'll learn about two unsupervised learning techniques for data visualization, hierarchical clustering and t-SNE. Hierarchical clustering merges the data samples into ever-coarser clusters, yielding a tree visualization of the resulting cluster hierarchy. t-SNE maps the data samples into 2d space so that the proximity of the samples to one another can be visualized.

Exercise 1: Visualizing hierarchies Exercise 2: How many merges?Exercise 3: Hierarchical clustering of the grain data Exercise 4: Hierarchies of stocks Exercise 5: Cluster labels in hierarchical clustering Exercise 6: Which clusters are closest?Exercise 7: Different linkage, different hierarchical clustering!Exercise 8: Intermediate clusterings Exercise 9: Extracting the cluster labels Exercise 10: t-SNE for 2-dimensional maps Exercise 11: t-SNE visualization of grain dataset Exercise 12: A t-SNE map of the stock market

Dimension reduction summarizes a dataset using its common occuring patterns. In this chapter, you'll learn about the most fundamental of dimension reduction techniques, "Principal Component Analysis" ("PCA"). PCA is often used before supervised learning to improve model performance and generalization. It can also be useful for unsupervised learning. For example, you'll employ a variant of PCA will allow you to cluster Wikipedia articles by their content!

Exercise 1: Visualizing the PCA transformation Exercise 2: Correlated data in nature Exercise 3: Decorrelating the grain measurements with PCA Exercise 4: Principal components Exercise 5: Intrinsic dimension Exercise 6: The first principal component Exercise 7: Variance of the PCA features Exercise 8: Intrinsic dimension of the fish data Exercise 9: Dimension reduction with PCA Exercise 10: Dimension reduction of the fish measurements Exercise 11: A tf-idf word-frequency array Exercise 12: Clustering Wikipedia part I Exercise 13: Clustering Wikipedia part II

In this chapter, you'll learn about a dimension reduction technique called "Non-negative matrix factorization" ("NMF") that expresses samples as combinations of interpretable parts. For example, it expresses documents as combinations of topics, and images in terms of commonly occurring visual patterns. You'll also learn to use NMF to build recommender systems that can find you similar articles to read, or musical artists that match your listening history!

Exercise 1: Fattorizzazione di matrici non negative (NMF)Exercise 2: Dati non negativi Exercise 3: NMF applicata agli articoli di Wikipedia Exercise 4: Feature NMF degli articoli di Wikipedia Exercise 5: NMF ricostruisce i campioni Exercise 6: NMF apprende parti interpretabili Exercise 7: NMF impara gli argomenti dei documenti Exercise 8: Esplora il dataset dei numeri LED Exercise 9: NMF impara le parti delle immagini

Esercizio in corso

Exercise 10: PCA non apprende parti Exercise 11: Creare sistemi di raccomandazione con NMF Exercise 12: Quali articoli sono simili a 'Cristiano Ronaldo'?Exercise 13: Consiglia artisti musicali parte I Exercise 14: Consiglia artisti musicali parte II Exercise 15: Considerazioni finali