Mapping of Regional-scale Multi-element Geochemical Anomalies Using Hierarchical Clustering Algorithms

CPSO: Chaotic Particle Swarm Optimization for Cluster Analysis

“…Partition-based methods K-means [37] Using the center of a cluster to represent a cluster K-means++ [38] Improvements have been made in the way K-means performs the initialization of cluster centers CLARA [39] Sampling techniques are used to enable the processing of large-scale data Hierarchical methods BIRCH [40] Using a tree structure to process the dataset CURE [41] Partitioning the dataset by randomly selected samples and global clustering after local clustering of each partition ROCK [42] Taking the influence of surrounding objects into account when calculating the similarity of two objects Density-based methods DBSCAN [43] Using spatial indexing techniques to search the neighborhood of an object, introducing the concept of density reachability OPTICS [44] Different parameters can be set for different clusters FDC [45] Dividing the entire data space into rectangular spaces Grid-based methods STING [46] Using grid cells to preserve data statistics, thus enabling multi-resolution clustering Wave Cluster [47] Introducing the wavelet transform, mainly used in signal processing CLIQUE [48] Combining grid and density clustering algorithms Others Quantum Clustering [49] Introducing the theory of quantum potential energy from quantum mechanics Kernel Clustering [50] Mapping of kernel functions to highlight features not originally shown Spectral Clustering [51] Clustering on arbitrarily shaped sample spaces and convergence to a globally optimal solution Optimization methods FHHO-NDPC algorithm [27] Significant improvements in convergence speed and solution accuracy when solving highdimensional problems A-SSC Algorithm [28] Better than the other four algorithms in identifying compact and well-separated clusters HCO algorithm [29] Abstracted from the hydrological cycling phenomenon WOATS [30] Significant differences can be observed in different outlier criteria solution it found and learns the optimum solution found by the whole group. The optimal solution discovered by a particle alone is referred to as the individual extremum, while the optimal solution currently found by the entire group is known as the global extreme.…”