Multiway clustering with time-varying parameters

Area of Science:

• Statistics
• Data Science
• Environmental Science

Background:

• Clustering of time series is well-established for static distributions.
• Methods for time-varying parameters are emerging but scarce for multivariate data.

Purpose of the Study:

• To propose a novel clustering approach for multivariate time series with time-varying parameters.
• To address the gap in existing methods for this specific data type.

Main Methods:

• Developed a multiway framework for distribution-based clustering.
• Incorporated time-varying parameter estimation within the clustering procedure.

Main Results:

• Demonstrated the effectiveness of the proposed clustering method.
• Validated the approach through a simulation study.

Conclusions:

• The proposed multiway clustering approach is effective for multivariate time series with time-varying parameters.
• The method shows practical applicability with real air quality data.