Density based spatial clustering of applications with noise and fuzzy C-means algorithms for unsupervised mineral prospectivity mapping

Our research focuses on examining two clustering methods, namely Density-Based Spatial Clustering of Applications with Noise (DBSCAN) and fuzzy c-means (FCM) algorithms, to create prospectivity models for Cu mineralization in the Kerman belt, SE Iran. Traditional clustering techniques struggle with...

Full description

Saved in:
Bibliographic Details
Published in:Earth science informatics Vol. 18; no. 2; p. 217
Main Authors: Ghezelbash, Reza, Daviran, Mehrdad, Maghsoudi, Abbas, Hajihosseinlou, Mahsa
Format: Journal Article
Language:English
Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.02.2025
Springer Nature B.V
Subjects:
Algorithms
Bioinformatics
Cluster analysis
Clustering
Copper
Density
Earth and Environmental Science
Earth Sciences
Earth System Sciences
Factor analysis
Fuzzy control
Geochemistry
Geographic information systems
Geological mapping
Image processing
Information Systems Applications (incl.Internet)
Interpolation
Mineral deposits
Mineralization
Network analysis
Ontology
Parameter identification
Remote sensing
Signatures
Simulation and Modeling
Social networks
Space Exploration and Astronautics
Space Sciences (including Extraterrestrial Physics
Cu-porphyry deposits
DBSCAN
Unsupervised MPM
Kerman
Fuzzy c-means
ISSN:1865-0473, 1865-0481
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Our research focuses on examining two clustering methods, namely Density-Based Spatial Clustering of Applications with Noise (DBSCAN) and fuzzy c-means (FCM) algorithms, to create prospectivity models for Cu mineralization in the Kerman belt, SE Iran. Traditional clustering techniques struggle with computational complexity and adaptability to large datasets, leading to a growing interest in developing improved algorithms. While numerous clustering algorithms have been utilized with promising outcomes, their performance heavily relies on user-specified parameters. The DBSCAN and FCM clustering methods were implemented to reduce the dimensions of nine attribute vectors derived from different exploration criteria. DBSCAN stands out for its ability to detect clusters with diverse shapes, finding applications in image processing, bioinformatics, and social network analysis. In contrast, traditional partitional clustering techniques face challenges in implementing non-convex clustering and may converge to a locally optimal solution. To identify indicators and potential controls on mineralization, we employed various evidence layers such as geochemical signatures, geological-structural clues, geophysical and remote sensing data. To enhance multi-element geochemical signatures in areas with Cu mineralization, we employed multifractal inverse distance weighting interpolation in conjunction with factor analysis. We transformed values of various evidence layers, including geological-structural controls and alterations, using the GIS-based fuzzy membership function MSLarge to fit within a range of 0 to 1. The Xie and Beni (VXB) index has been used in determining the optimal number of clusters for FCM-based MPM. We also utilized normalized density indices for a quantitative evaluation of the DBSCAN and FCM prospectivity maps. The evaluation results confirm the superior reliability of higher favorability classes derived from DBSCAN over those obtained from FCM in the identification of existing mineral deposits and the identification of new potential Cu mineralization zones within the study area.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1865-0473
1865-0481
DOI:10.1007/s12145-025-01708-0