7. 降維－PCA

n_components為降到多少維，用原數(shù)據(jù)fit后，再用transform轉(zhuǎn)換成降維后的數(shù)據(jù)。

from sklearn.decomposition import PCA
pca = PCA(n_components = 6)
pca.fit(good_data)
reduced_data = pca.transform(good_data)
reduced_data = pd.DataFrame(reduced_data, columns = ['Dimension 1', 'Dimension 2'])

8. 聚類－選擇類別數(shù)

用 silhouette coefficient 計(jì)算每個(gè)數(shù)據(jù)到中心點(diǎn)的距離，-1 (dissimilar) to 1 (similar) 根據(jù)這個(gè)系數(shù)來評價(jià)聚類算法的優(yōu)劣。

from sklearn.cluster import KMeans
from sklearn.metrics import silhouette_score
cluster = KMeans(n_clusters=2, random_state=0).fit(reduced_data)
preds = cluster.predict(reduced_data)
score = silhouette_score(reduced_data, preds)

選擇分?jǐn)?shù)最大的個(gè)數(shù)作為聚類的類別數(shù)。

9. 恢復(fù)維度

例如數(shù)據(jù)，先經(jīng)過 log，又經(jīng)過 PCA降維， 要恢復(fù)回去，先用 pca.inverse_transform，再用 np.exp

log_centers = pca.inverse_transform(centers)
true_centers = np.exp(log_centers)
true_centers = pd.DataFrame(np.round(true_centers), columns = data.keys())

數(shù)據(jù)變化：

10. 自定義accuracy

分類問題可以自己寫accuracy的函數(shù)

def accuracy_score(truth, pred):
    """ Return accuracy score for input truth and prediction"""
    
    if len(truth)==len(pred):
        return "Accuracy for prediction: {:.2f}%.".format((truth==pred).mean()*100)
    else:
        return "Numbers do not match!"