Time-Series K-means in Causal Inference and Mechanism Clustering for Financial Data

This paper examines the application of the time series k-means (TS-k-means) algorithm in the analysis of financial time series data. It employs dynamic time warping (DTW) as the metric and addresses the limitation of traditional clustering methods in capturing time dynamics. When combined with the additive noise Model Mixed Model (ANM-MM) architecture, this strategy not only enhances the accuracy of causal inference, but also enables clustering based on intrinsic generation mechanisms, thereby providing solid support for an in-depth understanding of complex financial time series data. The combination of TS-k-means and ANM-MM has been demonstrated to exhibit excellent performance in capturing fine causality and data clustering. Furthermore, it has been shown to possess significant advantages over traditional methods, while also demonstrating good stability and reliability on small- scale data sets.