Probabilistic mixture model driven interpretable modeling, clustering, and predicting for physical system data

doi:10.1016/j.engappai.2025.112069

Probabilistic mixture model driven interpretable modeling, clustering, and predicting for physical system data

2025-08-23

0

PRE

AI

DOI：10.1016/j.engappai.2025.112069

Original

Original request for help

Save

Abstract

En 中文

• A set of interpretable modeling, clustering, and prediction methods driven by Gaussian mixture model (GMM) is proposed. • Component numbers for the modeling of GMM can be adaptive determination. • Time series prediction driven by multiple bivariate GMMs for physical systems of multiple-input single-output is achieved. • Prediction performance of the proposed method approaches the deep neural networks while possessing interpretability. • Real measurement data from two bridges is used to verify the performance of algorithms.

Keywords:

Gaussian mixture model

interpretable modeling

time series prediction

clustering

physical systems

Probabilistic mixture model driven interpretable modeling, clustering, and predicting for physical system data

Abstract

Journal

Organization