sklearn 聚类：计算 TF-IDF 加权数据的轮廓系数答案

【问题标题】：sklearn clustering: calculate silhouette coefficient on TF-IDF-weigthed datasklearn 聚类：计算 TF-IDF 加权数据的轮廓系数
【发布时间】：2017-04-21 00:46:18
【问题描述】：

我想像 scikit-learn 示例 silhouette_analysis 一样计算 silhouette_score。

from sklearn.feature_extraction.text import TfidfVectorizer

tfidf_vectorizer = TfidfVectorizer(use_idf=True)
sampleText = []
sampleText.append("Some text for document clustering")
tfidf_matrix = tfidf_vectorizer.fit_transform(sampleText)

我必须如何转换我的 tfidf_matrix 来做这样的事情：

import matplotlib.cm as cm
from sklearn.metrics import silhouette_samples, silhouette_score
import matplotlib.pyplot as plt


for num_clusters in range(2,6):
    # Create a subplot with 1 row and 2 columns
    fig, (ax1, ax2) = plt.subplots(1, 2)
    fig.set_size_inches(18, 7)

    # The 1st subplot is the silhouette plot
    # The silhouette coefficient can range from -1, 1 but in this example all
    # lie within [-0.1, 1]
    ax1.set_xlim([-0.1, 1])
    # The (n_clusters+1)*10 is for inserting blank space between silhouette
    # plots of individual clusters, to demarcate them clearly.
    ax1.set_ylim([0, len(tfidf_matrix) + (num_clusters + 1) * 10])

    km = KMeans(n_clusters=num_clusters,
                n_init=10,                        # number of iterations with different seeds
                random_state=1                    # fixes the seed 
               )

    cluster_labels = km.fit_predict(tfidf_matrix)

    # The silhouette_score gives the average value for all the samples.
    # This gives a perspective into the density and separation of the formed
    # clusters
    silhouette_avg = silhouette_score(tfidf_matrix, cluster_labels)

【问题讨论】：

标签： python scikit-learn cluster-analysis analysis silhouette

【解决方案1】：

tf-idf 是多维的，必须简化为二维。这可以通过将 tf-idf 减少到具有最高方差的两个特征来完成。我使用 PCA 来减少 tf-idf。完整的例子：

from sklearn.feature_extraction.text import TfidfVectorizer

tfidf_vectorizer = TfidfVectorizer(use_idf=True)
sampleText = []
sampleText.append("Some text for document clustering")
tfidf_matrix = tfidf_vectorizer.fit_transform(sampleText)
X = tfidf_vectorizer.fit_transform(jobDescriptions).todense()

from sklearn.decomposition import PCA
pca = PCA(n_components=2).fit(X)
data2D = pca.transform(X)  

import matplotlib.cm as cm
from sklearn.metrics import silhouette_samples, silhouette_score
import matplotlib.pyplot as plt


for num_clusters in range(2,6):
# Create a subplot with 1 row and 2 columns
fig, (ax1, ax2) = plt.subplots(1, 2)
fig.set_size_inches(18, 7)

# The 1st subplot is the silhouette plot
# The silhouette coefficient can range from -1, 1 but in this example all
# lie within [-0.1, 1]
ax1.set_xlim([-0.1, 1])
# The (n_clusters+1)*10 is for inserting blank space between silhouette
# plots of individual clusters, to demarcate them clearly.
ax1.set_ylim([0, len(data2D) + (num_clusters + 1) * 10])

km = KMeans(n_clusters=num_clusters,
            n_init=10,                        # number of iterations with different seeds
            random_state=1                    # fixes the seed 
           )

cluster_labels = km.fit_predict(data2D)

# The silhouette_score gives the average value for all the samples.
# This gives a perspective into the density and separation of the formed
# clusters
 silhouette_avg = silhouette_score(data2D, cluster_labels)

【讨论】：