Clustering Disjoint HJ-Biplot: A new tool for identifying pollution patterns in geochemical studies

Area of Science:

• Environmental Geochemistry
• Data Mining
• Multivariate Data Analysis

Background:

• Understanding complex environmental data requires advanced analytical methods.
• Existing methods may not effectively separate groups based on specific geochemical origins.
• Identifying pollution sources in river sediments is crucial for environmental management.

Purpose of the Study:

• To introduce and validate the Clustering Disjoint HJ-Biplot (CDBiplot) algorithm.
• To demonstrate CDBiplot's capability in classifying object groups in a reduced space.
• To apply CDBiplot to an environmental geochemistry case study for pollution source identification.

Main Methods:

• Development of the Clustering Disjoint HJ-Biplot (CDBiplot) mathematical algorithm.
• Generation of disjoint factorial axes where variables uniquely contribute.
• Application of HJ-Biplot for graphical representation of individuals and variables.
• Implementation of CDBiplot using a function in the R programming language.

Main Results:

• CDBiplot successfully classified river sediment samples based on geochemical composition.
• The algorithm achieved excellent separation of samples, correlating groups with geological origin and anthropogenic inputs.
• Clear identification of pollution sources and delimitation of polluted zones were achieved.

Conclusions:

• The proposed CDBiplot algorithm is a powerful tool for analyzing environmental geochemistry data.
• CDBiplot enables detailed study of geochemical interactions and effective sample classification.
• The methodology shows potential for application in diverse research fields beyond environmental geochemistry.