KI-Glossar
Das vollständige Wörterbuch der Künstlichen Intelligenz
Agglomerative Clustering
Bottom-up approach starting from individual points and progressively merging the most similar clusters until a single cluster is obtained.
Hierarchical Clustering - Divisive
Top-down approach starting with a single cluster and recursively dividing it into more homogeneous sub-clusters.
K-Means and K-Medoids
Iterative partitioning algorithms assigning points to the k nearest centers and recalculating these centers to minimize intra-cluster variance.
DBSCAN and Density-Based Clustering
Techniques that identify clusters as dense regions separated by low-density areas, capable of detecting arbitrary shapes.
Spectral Clustering
Method using the eigenvalues and eigenvectors of the similarity matrix to project data into a space where clustering becomes simpler.
Hierarchical Linkage Methods
Distance measurement criteria between clusters including single linkage, complete linkage, average linkage and Ward's method for hierarchical clustering.
Clustering par Modèle
Approche probabiliste supposant que les données proviennent d'un mélange de distributions statistiques, typiquement des mélanges gaussiens.
Grid Clustering
Techniques discretizing the data space into a grid of cells and performing clustering on this structure for increased efficiency.
Stream Clustering
Algorithms adapted to processing continuous data arriving in real-time, requiring incremental cluster updates.
Fuzzy Clustering
Methods allowing a point to belong to multiple clusters with varying degrees of membership rather than binary assignment.
Dendrograms and Visualization
Graphical tools for representing hierarchical clustering structures showing successive merges/splits and similarity levels.
Cluster Validation
Set of metrics and techniques for evaluating the quality of obtained partitions, including silhouette indices, Davies-Bouldin, and internal/external validation.
Constraint-Based Clustering
Semi-supervised approaches integrating must-link and cannot-link constraints to guide the clustering process according to prior knowledge.
Multi-scale Clustering
Techniques identifying structures at different levels of granularity, suited for data with natural hierarchical patterns.