An open-source software package for multivariate modeling and clustering: applications to air quality management

Area of Science:

• Environmental Science
• Statistical Modeling
• Data Analysis

Background:

• Modeling complex relationships between multiple dependent and independent variables is challenging.
• Existing statistical tools may struggle with nonlinear relationships or mixed variable types.

Purpose of the Study:

• To introduce rSCA, an open-source software package for statistical modeling.
• To provide a tool for analyzing complex relationships between continuous and discrete variables, including nonlinear ones.
• To demonstrate the application of rSCA in real-world environmental management scenarios.

Main Methods:

• Developed an open-source software package named rSCA.
• Utilized a stepwise cluster analysis method based on multivariate analysis of variance (MANOVA).
• Implemented cutting and merging operations to divide sample sets into subclusters, visualized as a cluster tree.

Main Results:

• rSCA efficiently handles continuous and discrete variables, as well as nonlinear relationships.
• The software generates a cluster tree illustrating variable relationships and analysis pathways.
• Demonstrated effectiveness in air quality management, showcasing its utility in complex real-world problems.

Conclusions:

• rSCA is a versatile and user-friendly statistical tool for modeling intricate variable interactions.
• The software provides valuable insights for environmental management and other fields dealing with complex data.
• rSCA is freely available, promoting wider adoption and application in scientific research.