This documentation is for scikit-learn version 0.11-gitOther versions

Citing

If you use the software, please consider citing scikit-learn.

This page

4.3. Clustering

Clustering of unlabeled data can be performed with the module sklearn.cluster.

Each clustering algorithm comes in two variants: a class, that implements the fit method to learn the clusters on train data, and a function, that, given train data, returns an array of integer labels corresponding to the different clusters. For the class, the labels over the training data can be found in the labels_ attribute.

Input data

One important thing to note is that the algorithms implemented in this module take different kinds of matrix as input. On one hand, MeanShift and KMeans take data matrices of shape [n_samples, n_features]. These can be obtained from the classes in the sklearn.feature_extraction module. On the other hand, AffinityPropagation and SpectralClustering take similarity matrices of shape [n_samples, n_samples]. These can be obtained from the functions in the sklearn.metrics.pairwise module. In other words, MeanShift and KMeans work with points in a vector space, whereas AffinityPropagation and SpectralClustering can work with arbitrary objects, as long as a similarity measure exists for such objects.

4.3.1. K-means

The KMeans algorithm clusters data by trying to separate samples in n groups of equal variance, minimizing a criterion known as the ‘inertia’ of the groups. This algorithm requires the number of cluster to be specified. It scales well to large number of samples, however its results may be dependent on an initialisation. As a result, the computation is often done several times, with different initialisation of the centroids.

K-means is often referred to as Lloyd’s algorithm. After initialization, k-means consists of looping between two major steps. First the Voronoi diagram of the points is calculated using the current centroids. Each segment in the Voronoi diagram becomes a separate cluster. Secondly, the centroids are updated to the mean of each segment. The algorithm then repeats this until a stopping criteria is fulfilled. Usually, as in this implementation, the algorithm stops when the relative increment in the results between iterations is less than the given tolerance value.

K-means can be used for vector quantization. This is achieved using the transform method of a trained model of KMeans.

Examples:

4.3.1.1. Mini Batch K-Means

The MiniBatchKMeans is a variant of the KMeans algorithm using mini-batches, random subset of the dataset, to compute the centroids.

Althought the MiniBatchKMeans converge faster than the KMeans version, the quality of the results, measured by the inertia, the sum of the distance of each points to the nearest centroid, is not as good as the KMeans algorithm.

../_images/plot_mini_batch_kmeans_11.png

Examples:

References:

4.3.2. Affinity propagation

AffinityPropagation clusters data by diffusion in the similarity matrix. This algorithm automatically sets its numbers of cluster. It will have difficulties scaling to thousands of samples.

../_images/plot_affinity_propagation_11.png

Examples:

4.3.3. Mean Shift

MeanShift clusters data by estimating blobs in a smooth density of points matrix. This algorithm automatically sets its numbers of cluster. It will have difficulties scaling to thousands of samples.

../_images/plot_mean_shift_11.png

Examples:

4.3.4. Spectral clustering

SpectralClustering does a low-dimension embedding of the affinity matrix between samples, followed by a KMeans in the low dimensional space. It is especially efficient if the affinity matrix is sparse and the pyamg module is installed. SpectralClustering requires the number of clusters to be specified. It works well for a small number of clusters but is not advised when using many clusters.

For two clusters, it solves a convex relaxation of the normalised cuts problem on the similarity graph: cutting the graph in two so that the weight of the edges cut is small compared to the weights in of edges inside each cluster. This criteria is especially interesting when working on images: graph vertices are pixels, and edges of the similarity graph are a function of the gradient of the image.

noisy_img segmented_img

Warning

Shapeless isotropic data

When the data is really shapeless (i.e. generated from a random distribution with no clusters), the spectral-clustering problem is ill-conditioned: the different choices are almost equivalent, and the spectral clustering solver chooses an arbitrary one, putting the first sample alone in one bin.

Examples:

References:

4.3.5. Hierarchical clustering

Hierarchical clustering is a general family of clustering algorithms that build nested clusters by merging them successively. This hierarchy of clusters represented as a tree (or dendrogram). The root of the tree is the unique cluster that gathers all the samples, the leaves being the clusters with only one sample. See the Wikipedia page for more details.

The Ward object performs a hierarchical clustering based on the Ward algorithm, that is a variance-minimizing approach. At each step, it minimizes the sum of squared differences within all clusters (inertia criterion).

This algorithm can scale to large number of samples when it is used jointly with an connectivity matrix, but can be computationally expensive when no connectivity constraints are added between samples: it considers at each step all the possible merges.

4.3.5.1. Adding connectivity constraints

An interesting aspect of the Ward object is that connectivity constraints can be added to this algorithm (only adjacent clusters can be merged together), through an connectivity matrix that defines for each sample the neighboring samples following a given structure of the data. For instance, in the swiss-roll example below, the connectivity constraints forbid the merging of points that are not adjacent on the swiss roll, and thus avoid forming clusters that extend across overlapping folds of the roll.

unstructured structured

The connectivity constraints are imposed via an connectivity matrix: a scipy sparse matrix that has elements only at the intersection of a row and a column with indices of the dataset that should be connected. This matrix can be constructed from apriori information, for instance if you whish to cluster web pages, but only merging pages with a link pointing from one to another. It can also be learned from the data, for instance using sklearn.neighbors.kneighbors_graph to restrict merging to nearest neighbors as in the swiss roll example, or using sklearn.feature_extraction.image.grid_to_graph to enable only merging of neighboring pixels on an image, as in the Lena example.

Examples:

4.3.6. DBSCAN

The DBSCAN algorithm clusters data by finding core points which have many neighbours within a given radius. After a core point is found, the cluster is expanded by adding its neighbours to the current cluster and recusively checking if any are core points. Formally, a point is considered a core point if it has more than min_points points which are of a similarity greater than the given threshold eps. This is shown in the figure below, where the color indicates cluster membership and large circles indicate core points found by the algorithm. Moreover, the algorithm can detect outliers, indicated by black points below.

dbscan_results

Examples:

References:

  • “A Density-Based Algorithm for Discovering Clusters in Large Spatial Databases with Noise” Ester, M., H. P. Kriegel, J. Sander, and X. Xu, In Proceedings of the 2nd International Conference on Knowledge Discovery and Data Mining, Portland, OR, AAAI Press, pp. 226–231. 1996

4.3.7. Clustering performance evaluation

Evaluating the performance of a clustering algorithm is not as trivial as counting the number of errors or the precision and recall of a supervised classification algorithm. In particular any evaluation metric should not take the absolute values of the cluster labels into account but rather if this clustering define separations of the data similar to some ground truth set of classes or satisfying some assumption such that members belong to the same class are more similar that members of different classes according to some similarity metric.

4.3.7.1. Inertia

4.3.7.1.1. Presentation and usage

TODO: factorize inertia computation out of kmeans and then write me!

4.3.7.1.2. Advantages

  • No need for the ground truth knowledge of the “real” classes.

4.3.7.1.3. Drawbacks

  • Inertia makes the assumption that clusters are convex and isotropic which is not always the case especially of the clusters are manifolds with weird shapes: for instance inertia is a useless metrics to evaluate clustering algorithm that tries to identify nested circles on a 2D plane.
  • Inertia is not a normalized metrics: we just know that lower values are better and bounded by zero. One potential solution would be to adjust inertia for random clustering (assuming the number of ground truth classes is known).

4.3.7.2. Ajusted Rand index

4.3.7.2.1. Presentation and usage

Given the knowledge of the ground truth class assignments labels_true and our clustering algorithm assignments of the same samples labels_pred, the adjusted Rand index is a function that measures the similarity of the two assignements, ignoring permutations and with chance normalization:

>>> from sklearn import metrics
>>> labels_true = [0, 0, 0, 1, 1, 1]
>>> labels_pred = [0, 0, 1, 1, 2, 2]

>>> metrics.adjusted_rand_score(labels_true, labels_pred)  
0.24...

One can permute 0 and 1 in the predicted labels and rename 2 by 3 and get the same score:

>>> labels_pred = [1, 1, 0, 0, 3, 3]
>>> metrics.adjusted_rand_score(labels_true, labels_pred)  
0.24...

Furthermore, adjusted_rand_score is symmetric: swapping the argument does not change the score. It can thus be used as a consensus measure:

>>> metrics.adjusted_rand_score(labels_pred, labels_true)  
0.24...

Perfect labeling is scored 1.0:

>>> labels_pred = labels_true[:]
>>> metrics.adjusted_rand_score(labels_true, labels_pred)
1.0

Bad (e.g. independent labelings) have negative or close to 0.0 scores:

>>> labels_true = [0, 1, 2, 0, 3, 4, 5, 1]
>>> labels_pred = [1, 1, 0, 0, 2, 2, 2, 2]
>>> metrics.adjusted_rand_score(labels_true, labels_pred)  
-0.12...

4.3.7.2.2. Advantages

  • Random (uniform) label assignements have a ARI score close to 0.0 for any value of n_clusters and n_samples (which is not the case for raw Rand index or the V-measure for instance).
  • Bounded range [-1, 1]: negative values are bad (independent labelings), similar clusterings have a positve ARI, 1.0 is the perfect match score.
  • No assumption is made on the cluster structure: can be used to compare clustering algorithms such as k-means which assumes isotropic blob shapes with results of spectral clustering algorithms which can find cluster with “folded” shapes.

4.3.7.2.3. Drawbacks

  • Contrary to inertia, ARI requires the knowlege of the ground truth classes while almost never available in practice or requires manual assignment by human annotators (as in the supervised learning setting).

    However ARI can also be useful in purely unsupervised setting as a building block for a Consensus Index that can be used for clustering model selection (TODO).

Examples:

4.3.7.2.4. Mathematical formulation

If C is a ground truth class assignement and K the clustering, let us define a and b as:

  • a, the number of pairs of elements that are in the same set in C and in the same set in K
  • b, the number of pairs of elements that are in different sets in C and in different sets in K

The raw (unadjusted) Rand index is then given by:

RI = \frac{a + b}{C_2^{n_{samples}}}

Where C_2^{n_{samples}} is the total number of possible pairs in the dataset (without ordering).

However the RI score does not guarantee that random label assignements will get a value close to zero (esp. if the number of clusters is in the same order of magnitude as the number of samples).

To counter this effect we can discount the expected RI of random labelings by defining the adjusted Rand index as follows:

ARI = \frac{RI - Expected\_RI}{max(RI) - Expected\_RI}

References

4.3.7.3. Ajusted Mutual Information

4.3.7.3.1. Presentation and usage

Given the knowledge of the ground truth class assignments labels_true and our clustering algorithm assignments of the same samples labels_pred, the Adjusted Mutual Information is a function that measures the agreement of the two assignements, ignoring permutations and with chance normalization:

>>> from sklearn import metrics
>>> labels_true = [0, 0, 0, 1, 1, 1]
>>> labels_pred = [0, 0, 1, 1, 2, 2]

>>> metrics.adjusted_mutual_info_score(labels_true, labels_pred)  
0.22504...

One can permute 0 and 1 in the predicted labels and rename 2 by 3 and get the same score:

>>> labels_pred = [1, 1, 0, 0, 3, 3]
>>> metrics.adjusted_mutual_info_score(labels_true, labels_pred)  
0.22504...

Furthermore, adjusted_mutual_info_score is symmetric: swapping the argument does not change the score. It can thus be used as a consensus measure:

>>> metrics.adjusted_mutual_info_score(labels_pred, labels_true)  
0.22504...

Perfect labeling is scored 1.0:

>>> labels_pred = labels_true[:]
>>> metrics.adjusted_mutual_info_score(labels_true, labels_pred)
1.0

Bad (e.g. independent labelings) have non-positive scores:

>>> labels_true = [0, 1, 2, 0, 3, 4, 5, 1]
>>> labels_pred = [1, 1, 0, 0, 2, 2, 2, 2]
>>> metrics.adjusted_mutual_info_score(labels_true, labels_pred)  
-0.10526...

4.3.7.3.2. Advantages

  • Random (uniform) label assignements have a AMI score close to 0.0 for any value of n_clusters and n_samples (which is not the case for raw Mutual Information or the V-measure for instance).
  • Bounded range [0, 1]: Values close to zero indicate two label assignments that are largely independent, while values close to one indicate significant agreement. Further, values of exactly 0 indicate purely independent label assignments and a AMI of exactly 1 indicates that the two label assignments are equal (with or without permutation).
  • No assumption is made on the cluster structure: can be used to compare clustering algorithms such as k-means which assumes isotropic blob shapes with results of spectral clustering algorithms which can find cluster with “folded” shapes.

4.3.7.3.3. Drawbacks

  • Contrary to inertia, AMI requires the knowlege of the ground truth classes while almost never available in practice or requires manual assignment by human annotators (as in the supervised learning setting).

    However AMI can also be useful in purely unsupervised setting as a building block for a Consensus Index that can be used for clustering model selection.

Examples:

4.3.7.3.4. Mathematical formulation

Assume two label assignments (of the same data), U with R classes and V with C classes. The entropy of either is the amount of uncertaintly for an array, and can be calculated as:

H(U) = \sum_{i=1}^{|R|}P(i)log(P(i))

Where P(i) is the number of instances in U that are in class R_i. Likewise, for V:

H(V) = \sum_{j=1}^{|C|}P'(j)log(P'(j))

Where P’(j) is the number of instances in V that are in class C_j.

The (non-adjusted) mutual information between U and V is calculated by:

MI(U, V) = \sum_{i=1}^{|R|}\sum_{j=1}^{|C|}P(i, j)log(\frac{P(i,j)}{P(i)P'(j)})

Where P(i, j) is the number of instances with label R_i and also with label C_j.

This value of the mutual information is not adjusted cfor chance and will tend to increase as the number of different labels (clusters) increases, regardless of the actual amount of “mutual information” between the label assignments.

The expected value for the mutual information can be calculated using the following equation, from Vinh, Epps, and Bailey, (2009). In this equation, a_i is the number of instances with label U_i and b_j is the number of instances with label V_j.

E\{MI(U,V)\}=\sum_{i=1}^R \sum_{j=1}^C \sum_{n_{ij}=(a_i+b_j-N)^+
}^{\min(a_i, b_j)} \frac{n_{ij}}{N}\log ( \frac{ N.n_{ij}}{a_i b_j})
\frac{a_i!b_j!(N-a_i)!(N-b_j)!}{N!n_{ij}!(a_i-n_{ij})!(b_j-n_{ij})!
(N-a_i-b_j+n_{ij})!}

Using the expected value, the adjusted mutual information can then be calculated using a similar form to that of the adjusted Rand index:

AMI = \frac{MI - Expected\_MI}{max(H(U), H(V)) - Expected\_MI}

References

4.3.7.4. Homogeneity, completeness and V-measure

4.3.7.4.1. Presentation and usage

Given the knowledge of the ground truth class assignments of the samples, it is possible to define some intuitive metric using conditional entropy analysis.

In particular Rosenberg and Hirschberg (2007) define the following two desirable objectives for any cluster assignment:

  • homogeneity: each cluster contains only members of a single class.
  • completeness: all members of a given class are assigned to the same cluster.

We can turn those concept as scores homogeneity_score and completeness_score. Both are bounded below by 0.0 and above by 1.0 (higher is better):

>>> from sklearn import metrics
>>> labels_true = [0, 0, 0, 1, 1, 1]
>>> labels_pred = [0, 0, 1, 1, 2, 2]

>>> metrics.homogeneity_score(labels_true, labels_pred)  
0.66...

>>> metrics.completeness_score(labels_true, labels_pred) 
0.42...

Their harmonic mean called V-measure is computed by v_measure_score:

>>> metrics.v_measure_score(labels_true, labels_pred)    
0.51...

All three metrics can be computed at once using homogeneity_completeness_v_measure as follows:

>>> metrics.homogeneity_completeness_v_measure(labels_true, labels_pred)
...                                                      
(0.66..., 0.42..., 0.51...)

The following clustering assignment is slighlty better, since it is homogeneous but not complete:

>>> labels_pred = [0, 0, 0, 1, 2, 2]
>>> metrics.homogeneity_completeness_v_measure(labels_true, labels_pred)
...                                                      
(1.0, 0.68..., 0.81...)

Note

v_measure_score is symmetric: it can be used to evaluate the agreement of two independent assignements on the same dataset.

This is not the case for completeness_score and homogeneity_score: both are bound by the relationship:

homogeneity_score(a, b) == completeness_score(b, a)

4.3.7.4.2. Advantages

  • Bounded scores: 0.0 is as bad as it can be, 1.0 is a perfect score
  • Intuitive interpretation: clustering with bad V-measure can be qualitatively analyzed in terms of homogeneity and completeness to better feel what ‘kind’ of mistakes is done by the assigmenent.
  • No assumption is made on the cluster structure: can be used to compare clustering algorithms such as k-means which assumes isotropic blob shapes with results of spectral clustering algorithms which can find cluster with “folded” shapes.

4.3.7.4.3. Drawbacks

  • The previously introduced metrics are not normalized w.r.t. random labeling: this means that depending on the number of samples, clusters and ground truth classes, a completely random labeling will not always yield the same values for homogeneity, completeness and hence v-measure. In particular random labeling won’t yield zero scores especially when the number of clusters is large.

    This problem can safely be ignored when the number of samples is more than a thousand and the number of clusters is less than 10. For smaller sample sizes or larger number of clusters it is safer to use an adjusted index such as the Adjusted Rand Index (ARI).

../_images/plot_adjusted_for_chance_measures_11.png
  • These metrics require the knowlege of the ground truth classes while almost never available in practice or requires manual assignment by human annotators (as in the supervised learning setting).

Examples:

4.3.7.4.4. Mathematical formulation

Homogeneity and completeness scores are formally given by:

h = 1 - \frac{H(C|K)}{H(C)}

c = 1 - \frac{H(K|C)}{H(K)}

where H(C|K) is the conditional entropy of the classes given the cluster assignments and is given by:

H(C|K) = - \sum_{c=1}^{|C|} \sum_{k=1}^{|K|} \frac{n_{c,k}}{n}
\cdot log(\frac{n_{c,k}}{n_k})

and H(C) is the entropy of the classes and is given by:

H(C) = - \sum_{c=1}^{|C|} \frac{n_c}{n} \cdot log(\frac{n_c}{n})

with n the total number of samples, n_c and n_k the number of samples respectively belonging to class c and cluster k, and finally n_{c,k} the number of samples from class c assigned to cluster k.

The conditional entropy of clusters given class H(K|C) and the entropy of clusters H(K) are defined in a symmetric manner.

Rosenberg and Hirschberg further define V-measure as the harmonic mean of homogeneity and completeness:

v = 2 \cdot \frac{h \cdot c}{h + c}

References

4.3.7.5. Silhouette Coefficient

4.3.7.5.1. Presentation and usage

If the ground truth labels are not known, evaluation must be performed using the model itself. The Silhouette Coefficient (sklearn.metrics.silhouette_score) is an example of such an evaluation, where a higher Silhouette Coefficient score relates to a model with better defined clusters. The Silhouette Coefficient is defined for each sample and is composed of two scores:

  • a: The mean distance between a sample and all other points in the same class.
  • b: The mean distance between a sample and all other points in the next nearest cluster.

The Silhoeutte Coefficient s for a single sample is then given as:

s = \frac{b - a}{max(a, b)}

The Silhouette Coefficient for a set of samples is given as the mean of the Silhouette Coefficient for each sample.

>>> from sklearn import metrics
>>> from sklearn.metrics import pairwise_distances
>>> from sklearn import datasets
>>> dataset = datasets.load_iris()
>>> X = dataset.data
>>> y = dataset.target

In normal usage, the Silhouette Coefficient is applied to the results of a cluster analysis.

>>> import numpy as np
>>> from sklearn.cluster import KMeans
>>> kmeans_model = KMeans(k=3, random_state=1).fit(X)
>>> labels = kmeans_model.labels_
>>> metrics.silhouette_score(X, labels, metric='euclidean')
...                                                      
0.5525...

References

  • Peter J. Rousseeuw (1987). “Silhouettes: a Graphical Aid to the Interpretation and Validation of Cluster Analysis”. Computational and Applied Mathematics 20: 53–65. doi:10.1016/0377-0427(87)90125-7.

4.3.7.5.2. Advantages

  • The score is bounded between -1 for incorrect clustering and +1 for highly dense clustering. Scores around zero indicate overlapping clusters.
  • The score is higher when clusters are dense and well separated, which relates to a standard concept of a cluster.

4.3.7.5.3. Drawbacks

  • The Silhouette Coefficient is generally higher for convex clusters than other concepts of clusters, such as density based clusters like those obtained through DBSCAN.