WikiGalaxy
Clustering Algorithms
Understanding Clustering:
Clustering is a type of unsupervised learning technique used to group similar data points together. It helps in identifying patterns and structures within the data without any prior labels.
- Used for market segmentation, social network analysis, and image compression.
- Helps in understanding data distribution and identifying anomalies.
- Commonly used clustering algorithms include K-Means, Hierarchical Clustering, and DBSCAN.
K-Means Clustering
How K-Means Works:
K-Means is a popular clustering algorithm that partitions data into K clusters, where each data point belongs to the cluster with the nearest mean.
- Initialize K centroids randomly.
- Assign each data point to the nearest centroid.
- Recalculate the centroids as the mean of all points in the cluster.
- Repeat until centroids do not change significantly.
import org.apache.commons.math3.ml.clustering.KMeansPlusPlusClusterer;
import org.apache.commons.math3.ml.clustering.DoublePoint;
import java.util.List;
import java.util.ArrayList;
class KMeansExample {
public static void main(String[] args) {
List points = new ArrayList<>();
points.add(new DoublePoint(new double[]{1.0, 2.0}));
points.add(new DoublePoint(new double[]{2.0, 1.0}));
points.add(new DoublePoint(new double[]{4.0, 5.0}));
KMeansPlusPlusClusterer clusterer = new KMeansPlusPlusClusterer<>(2);
List> clusters = clusterer.cluster(points);
System.out.println("Number of clusters formed: " + clusters.size());
}
}
Benefits of K-Means:
K-Means is efficient and simple to implement. However, it requires the number of clusters to be defined beforehand and can be sensitive to outliers.
Console Output:
Number of clusters formed: 2
Hierarchical Clustering
Understanding Hierarchical Clustering:
Hierarchical clustering creates a tree of clusters called a dendrogram. It can be agglomerative (bottom-up) or divisive (top-down).
- Agglomerative starts with each data point as a single cluster and merges them iteratively.
- Divisive starts with all data points in one cluster and splits them iteratively.
- Useful for visualizing data and understanding its structure.
import org.apache.commons.math3.ml.clustering.hierarchical.*;
import org.apache.commons.math3.ml.clustering.DoublePoint;
import java.util.List;
import java.util.ArrayList;
class HierarchicalExample {
public static void main(String[] args) {
List points = new ArrayList<>();
points.add(new DoublePoint(new double[]{1.0, 2.0}));
points.add(new DoublePoint(new double[]{2.0, 1.0}));
points.add(new DoublePoint(new double[]{4.0, 5.0}));
AgglomerativeClusterer clusterer = new AgglomerativeClusterer<>(new EuclideanDistance(), 2);
List> clusters = clusterer.cluster(points);
System.out.println("Number of clusters formed: " + clusters.size());
}
}
Advantages of Hierarchical Clustering:
It does not require specifying the number of clusters in advance and provides a clear visualization through dendrograms. However, it can be computationally intensive for large datasets.
Console Output:
Number of clusters formed: 2
DBSCAN Clustering
What is DBSCAN?
DBSCAN (Density-Based Spatial Clustering of Applications with Noise) is a density-based clustering algorithm that groups together points that are closely packed and marks points in low-density regions as outliers.
- Does not require specifying the number of clusters.
- Can find arbitrarily shaped clusters.
- Robust to noise and outliers.
import org.apache.commons.math3.ml.clustering.DBSCANClusterer;
import org.apache.commons.math3.ml.clustering.DoublePoint;
import java.util.List;
import java.util.ArrayList;
class DBSCANExample {
public static void main(String[] args) {
List points = new ArrayList<>();
points.add(new DoublePoint(new double[]{1.0, 2.0}));
points.add(new DoublePoint(new double[]{2.0, 1.0}));
points.add(new DoublePoint(new double[]{4.0, 5.0}));
DBSCANClusterer clusterer = new DBSCANClusterer<>(1.0, 2);
List> clusters = clusterer.cluster(points);
System.out.println("Number of clusters formed: " + clusters.size());
}
}
Why Use DBSCAN?
DBSCAN is effective in handling noise and discovering clusters of varying shapes and sizes, making it suitable for spatial data analysis.
Console Output:
Number of clusters formed: 2
Mean Shift Clustering
Understanding Mean Shift:
Mean Shift is a non-parametric clustering technique that does not require specifying the number of clusters. It works by updating candidate points to the average of points within a given region.
- Tracks the mode of the data distribution.
- Effective for finding the center of clusters.
- Can handle clusters of varying shapes and sizes.
import smile.clustering.MeanShift;
import smile.data.DataFrame;
import smile.data.vector.DoubleVector;
class MeanShiftExample {
public static void main(String[] args) {
double[][] data = {
{1.0, 2.0},
{2.0, 1.0},
{4.0, 5.0}
};
DataFrame df = DataFrame.of(DoubleVector.of("x", data[0]), DoubleVector.of("y", data[1]));
MeanShift clustering = new MeanShift(df, 2.0);
System.out.println("Number of clusters formed: " + clustering.k());
}
}
Advantages of Mean Shift:
Mean Shift is versatile and can adapt to the data's inherent structure. However, it can be computationally expensive and sensitive to bandwidth selection.
Console Output:
Number of clusters formed: 2
Gaussian Mixture Models (GMM)
What is GMM?
Gaussian Mixture Models (GMM) use a probabilistic model to represent normally distributed subpopulations within an overall population. It is a soft clustering technique.
- Models the data as a mixture of several Gaussian distributions.
- Each Gaussian represents a cluster.
- Can handle complex cluster shapes and overlap.
import smile.clustering.GaussianMixture;
import smile.data.DataFrame;
import smile.data.vector.DoubleVector;
class GMMExample {
public static void main(String[] args) {
double[][] data = {
{1.0, 2.0},
{2.0, 1.0},
{4.0, 5.0}
};
DataFrame df = DataFrame.of(DoubleVector.of("x", data[0]), DoubleVector.of("y", data[1]));
GaussianMixture gmm = new GaussianMixture(df, 2);
System.out.println("Number of clusters formed: " + gmm.k());
}
}
Benefits of GMM:
GMM is flexible in modeling complex data distributions and provides a probabilistic clustering approach, allowing for soft assignments of data points to clusters.
Console Output:
Number of clusters formed: 2
